SWP-564 Basic Education and Agricultural Extension Costs, Effects, and Alternatives Hilary Perraton Dean T. Jamison Janet Jenkins Franqois Orivel Laurence Wolff WORLD BANK STAFF WORKING PAPERS Number 564 FILE COPY WORLD BANK STAFF WORKING PAPERS Number 564 Basic Education and Agricultural Extension Costs, Effects, and Alternatives Hilary Perraton Dean T. Jamison Janet Jenkins Francois Orivel Laurence Wolff The World Bank Washington, D.C., U.S.A. Copyright ® 1983 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. First printing April 1983 All rights reserved Manufactured in the United States of America This is a working document published informally by the World Bank. To present the results of research with the least possible delay, the typescript has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The publication is supplied at a token charge to defray part of the cost of manufacture and distribution. The views and interpretations in this document are those of the author(s) and should not be attributed to the World Bank, to its affiliated organizations, or to any individual acting on their behalf. Any maps used have been prepared solely for the convenience of the readers; the denominations used and the boundaries shown do not imply, on the part of the World Bank and its affiliates, any judgment on the legal status of any teritory or any endorsement or acceptance of such boundaries. The full range of World Bank publications is described in the Catalog of World Bank Publications; the continuing research program of the Bank is outlined in World Bank Research Program: Abstracts of Current Studies. Both booklets are updated annually; the most recent edition of each is available without charge from the Publications Distribution Unit of the Bank in Washington or from the European Office of the Bank, 66, avenue d1I6na, 75116 Paris, France. Hilary Perraton is codirector and Janet Jenkins is director of the International Extension College, Cambridge, England; Frangois Orivel is with the Institut de Recherche sur l'Economie de l'Education in the University of Dijon, France; Dean T. Jamison is senior projects officer in the Population, Health, and Nutrition Department and Laurence Wolff an education planner in the Eastern Africa Regional Office, both at the World Bank. Library of Congress Cataloging in Publication Data Main entry under title: Basic education and agricultural extension. (World Bank staff working papers ; no. 564) Bibliography: p. 1. Agricultural extension work--Addresses, essays, lectures. 2. Mass media in agricultural exten- sion work--Addresses, essays, lectures. 3. Basic educa- tion--Addresses. essays, lectures. 4. Farmers--Educa- tion--Addresses, essays, lectures. 5. Mass media in agricultural extension work--Africa--Case studies-- Addresses, essays, lectures. I. Perraton, H. D. II. Series. S544.B27 1983 630'.7'15 83-5903 ISBN 0-8213-0176-4 ABSTRACT The papers in this volume examine the evidence about the methods, costs and effects of agricultural extension and basic education and about the use of mass media for extension and education. The first two papers review the literature on the effectiveness of agricultural extension and on the use of mass media. Three case studies follow. All are of institutions which have used mass media, but in widely different ways. They are the Agricultural Information Service of the Ministry of Agriculture in Malawi, the non-government organisation INADES- Formation in West Africa, and the Lesotho Distance Teaching Centre. CONTENTS OVERVIEW ORTHODOX AND UNCONVENTIONAL APPROACHES TO EXTENSION ... iii Hilary Perraton I THE IMPACT OF AGRICULTURAL EXTENSION SERVICES: A REVIEW OF THE LITERATURE. 1 Francois Orivel II MASS MEDIA, BASIC EDUCATION AND AGRICULTURAL EXTENSION ............................................. 59 Hilary Perraton III MASS MEDIA FOR AGRICULTURAL EXTENSION IN MALAWI ...... 147 Hilary Perraton, Dean Jamison and Francois Orivel IV TRAINING FARMERS BY CORRESPONDENCE IN CAMEROON ....... 203 Janet Jenkins and Hilary Perraton V THE LESOTHO DISTANCE TEACHING CENTRE .................. 249 Laurence Wolff - iii - OVERVIEW ORTHODOX AND UNCONVENTIONAL APPROACHES TO EXTENSION This set of papers addresses a single theme: how effective are the traditional methods of agricultural extension and basic education and of what value are alternative approaches using the mass media? The papers have two points of origin within the Bank. First, they derive from the studies, reported by Jamison and Lau (1982) on the effect which the background education of farmers has on their farming practice. These showed a positive, significant, and consistent relationship between farmers' educational level and the management of their farms, but left more open questions about the relations between extension and farm management or productivity. The second line of research derives from studies of the use of radio for education. These showed that radio could be a cost-effective way of providing or supporting formal education (Spain, Jamison and McAnany 1977). That finding in turn provoked questions about the value of radio and other mass media for nonformal education and extension. These issues are examined in two ways. The first two papers review the literature on the effectiveness of agricultural extension and on the comparative value of mass media and traditional approaches. The remaining three papers are case studies on the use of mass media in Malawi, Cameroon, and Lesotho. What the papers say Orivel's review of the effectiveness of agricultural extension begins by summarising the development of extension work in the third world. The evaluation of this is complicated by problems of timing, where some effects of extension may be long delayed, and problems of objectives. Many research workers have wanted to look both at the impact of extension services on agricultural efficiency or production and on the effects of extension on equity. Studies of the internal efficiency of agricultural extension services have generally produced critical findings. Extension agents have often contacted only a minority of farmers and have tended to contact those who are richer and better educated. A number of reasons for these weaknesses have been identified. They include the, often unfavourable, ratio of agents to farmers; the inadequate training of extension agents; their location, which may make it difficult to visit farmers; their pay; and uncertainty about their role and the variety of functions they may be asked to undertake. Orivel concludes that, while such internal evaluations are generally pessimistic, they often leave out of account the external impact of extension work and have rarely attempted to measure its effect on yields or income. Next the paper examines the studies of the external effects of extension. Four kinds of studies are identified, using measures of internal rate of return, correlations at regional level, comparisons of yields and micro- iv _ economic production functions. Orivel looks at all four but stresses the methodological difficulties in the first two where results in terms of increased production cannot easily be ascribed to causes. Similarly, where yields are compared before or after the introduction of an extension service, or between regions with and without an extension service, it is not possible to attribute yields specifically to the extension service. Taking these weaknesses into account, it is still striking that most external studies of these three kinds do show positive results for extension. Finally Orivel examines twenty-one studies focusing on individual farms as the unit of observation which do make this kind of analysis possible. Nine of these studies produced significant positive correlations between contact with extension and productivity, eight found positive but statistically insignificant correlations, three found insignificant negative correlations and one found a significant negative correlation. Orivel argues that the first three methods tend to overemphasise the effectiveness of extension, by attributing all increases in production to it rather than to other factors, while the fourth tends to underemphasise its effectiveness, by failing to allow for the effects of diffusion. He concludes by arguing for further research on the relations between the educational level of farmers and their use of extension services and on the best vehicles for distributing information. The main emphasis of the next paper is not on traditional methods of agricultural extension but on the ways in which mass media have been used for similar purposes and for basic education. In order to place that discussion in context, it first examines the different philosophies of agricultural extension and of basic education projects, and looks at the evidence on the effectiveness of agricultural extension, parallel to the discussion in Orivel's paper, and of adult literacy programmes as an example of basic education. Perraton argues that traditional methods of basic education, as exemplified by literacy teaching, have been neither particularly successful nor particularly cheap. The main part of the paper discusses the use which has been made of mass media for rural education. It examines in detail ways in which mass media have been linked with group study through institutions like the radio schools of Latin America, farm forums in Africa and Asia, and short radio campaigns in Botswana and Tanzania. It also examines the use of mass media, using both radio and print, aimed at individuals rather than groups. Next the paper brings together evidence from existing studies on the costs of mass media as compared with alternatives. Radio schools in Latin America, which teach the equivalent of the primary school curriculum but to adults, have done so at a cost per head of between (1978) US$30 and 65 a year. It seems that radio campaigns in Tanzania, which reached audiences of between two and three million participants, cost between $1 and $5 per head for a campaign of two to three months. A significant part of the cost of either kind of scheme is the arrangement for group discussion and for feedback. The cost of radio programmes, without arrangements for group learning or organised feedback from listeners, compares favourably with most other methods of distributing information. Evidence from Malawi suggests that the cost of a one hour contact between a radio programme and a farmer is only 1/3000 of the cost of a face-to- face contact with an extension agent. The remaining three papers discuss, in much more detail, three different approaches to using mass media for rural education. The first is of a government department - The Extension Aids Service of the Ministry of Agriculture of Malawi. The Service has four major activities: producing and distributing publications, running a fleet of mobile vans which show films and puppet plays, broadcasting regular radio programmes, and running an evaluation unit which monitors the Branch's work. The Extension Aids Branch costs about 10% of the total national expenditure on agricultural extension. The paper discusses the organisation of the Extension Aids Branch, as it was at the time of the study, and its relation to the rest of the extension service. It then reviews the evidence on the effects of different sources of agricultural information, which are consistent in reporting that farmers learn both from film and from radio, although they do not establish a causal link between exposure to mass media and increased agricultural production. The authors then discuss in detail the cost of using mass media and compare these with the costs of other ways of reaching farmers. Costs are calculated for various kinds of contact between farmers and extension services. These show that it costs about $21 to arrange a single contact between an extension agent and a farmer; about $4-5 a day for a farmer to attend a training centre; about $0.17 per farmer contact hour when a farmer sees one of the Branch's mobile films; and about $0.004 per listener hour for radio. The second case study discusses the work of a non-government organisation, INADES-formation. This was founded by Jesuits in West Africa to promote African development and has been using correspondence courses to train farmers and extension agents since 1962. Its headquarters is in the Ivory Coast but it has separate national offices in some ten African countries. Within Cameroon, where the case study was carried out, it has four separate regional offices. In Cameroon INADES-formation distributes and teaches correspondence courses, runs one-day seminars, and plays a small part in contibuting to government broadcasting on agriculture. The courses are, for the most part, prepared in the Ivory Coast for use throughout the countries where INADES-formation operates but are in some cases adapted by national or regional offices. Farmers following the courses send correspondence work to a regional office where it is marked. For the most part they study individually and INADES-formation has had only limited success in organising group study. The courses do not lead to any formal qualification and are, in practice, effectively limited to literate students. The students are about evenly divided between peasant farmers and extension agents. In some parts of Cameroon the successful completion of INADES-formation courses is a requirement for promotion as an extension agent. There is only limited evidence on the effectiveness of INADES-formation. It reaches a very small proportion of the population but what evidence there is suggests that farmers do learn both better farming practices and the reasons for them from their courses. Two studies showed evidence of changes in farming practice, where local circumstances allowed this, and gains in knowledge for those following INADES courses. - vi The paper analyses the cost of enrolling on correspondence courses and found that the cost per student in 1979 was $224. A comparative study of the costs of INADES-formation in Rwanda showed that there were considerable potential economies of scale as the Rwanda cost per student was $92. The paper concludes by suggesting that an important future role for INADES-formation, and one which is probably highly cost effective, is in the training of extension agents rather than the teaching of peasant farmers directly. The final paper discusses the work of the Lesotho Distance Teaching Centre, which was established as a semi-government agency in 1974. Unlike the institutions described in the two previous parts, the Centre is concerned with a variety of different educational activities and not solely with non-formal or agricultural education. It provides, for example, correspondence courses at secondary level, and for trainee teachers, as well as teaching practical skills to rural people, promoting basic literacy, and helping other agencies working in nonformal education. A combination of factors, including relatively high literacy and poor radio reception conditions in the mountainous two-thirds of the country, have led the Centre to concentrate on the distribution of printed materials for rural education. It has produced a range of printed pamphlets designed to help rural adults acquire practical skills. On the basis of a study carried out by the Centre in 1977, Wolff analyses the production and distribution costs of one pamphlet on how to crochet. The unit cost of producing and distributing this amounted to about $0.65 per copy. There is evidence that the pamphlet was effective in teaching people to crochet or enabling them to crochet better. This makes it possible to estimate a rate of return for buyers of the pamphlets and to show that this is high both for individuals and for society. What we know and don't know We can draw four conclusions from these papers. First, the evidence on extension services is mixed. While existing services are properly criticised on grounds of internal efficeincy and equity, the evidence on extenal efficiency suggests that they are more likely to have a favourable benefit:cost ratio than the reverse. However, budgetary limits make it difficult for extension services to expand so that they reach all farmers. The evidence on the use of mass media is also mixed and, as with orthodox extension, at our present state of knowledge it is seldom possible to attribute results in terms of better agriculture or increased productivity directly to contact with the mass media. There is, however, some evidence of the effectiveness of mass media and clear evidence that they can provide contact between providers of agricultural information and farmers at costs far lower than are possible for orthodox extension agencies. This suggests that there is a case for expanding and improving the use of mass media, both as a support to extension services, and as a way of reaching farmers, including women vii - farmers, whom the extension agents do not meet. There remains much that we do not know, and on which research would be valuable, about the modalities of using mass media, about the best ways of combining different media and about ways of linking mass media with extension services. Hilary Perraton October 1982 References D.T. Jamison and L.J. Lau (1982) Farmer education and farm efficiency (Johns Hopkins University Press, Baltimore) P.L. Spain, D.T. Jamison and E.G. McAnany (1977) Radio for education and development: case studies (Education Department, World Bank, Washington DC) PART I THE IMPACT OF AGRICULTURAL EXTENSION: A REVIEW OF THE LITERATURE Francois Orivel Table of Contents Page No. AN OVERVIEW ......................................................... 3 INTRODUCTION ..................................... 4 I. GENERAL PROBLEMS ..................................... 5 (1) Objectives of extension services ................................ 5 (2) The context ..................................................... 7 (3) The time factor ................................................. 8 (4) Ideology ........................................................ 10 II. INTERNAL EVALUATION OF AGRICULTURAL EXTENSION SERVICES .14 (1) The underlying philosophies of agricultural extension projects 14 (2) Evaluation of the internal effeciveness of agricultural extension services ......................,.,,.,,. 16 Conclusion ........ 31 III. EXTERNAL EVALUATION OF AGRICULTURAL EXTENSION SERVICES .32 (1) Methods .33 (2) Results .33 Conclusion ... 48 Bibliography . ........... , , . ....................,.... 52 -3- AN OVERVIEW This paper provides a review of the literature on the effectiveness of agricultural extension services worldwide. It shows that evaluators of the internal effectiveness of extension services are generally critical; they claim that extension services have poor performance, reach only a small proportion of farmers (those who are better off), have wrong objectives, and are of little help to the agricultural sector as a whole. The author argues that this diagnosis may be excessive, that evaluators often have irrelevant evaluation criteria, and that historically the situation has substantially improved. Evaluators of external efficiency, mainly economists, show much more positive results. Those who use aggregate data find high social rates of return to investment in extension services and high correlations between extension services and agricultural productivity, but the methods used tend to overestimate the specific impact of extension services. Those who use farm-level data obtain mixed results, about half with significantly positive regression coefficients but of low value, and the other half with nonsignificant regression coefficients. -4- INTRODUCTION The purpose of this survey is to take a comprehensive view of the available information on the impact of agricultural extension services worldwide. Its aim is to provide decision-makers with a body of facts and reflections that can assist them in establishing rural development strategies which, in the light of their priorities and financial constraints, will produce the best results. The present review continues the study by Lockheed, Jamison and Lau (1980) on the effects of the educational level of farmers on the management of their farms, which demonstrated a positive, significant, and consistent relationship. The findings of our study are less conclusive, however, in that many of the evaluations of the impact of extension services reveal negligible, or even negative, effects. Another difference from the above-mentioned study is that evaluations and analyses of the impact of agricultural extension services have been more concerned with their effects on equity than have studies of the impact of education. When agricultural extension services are introduced in a poor region, both gross and net agricultural production in the region often increase significantly and the economic value of this increase is greater than the cost of the extension services, while at the same time the disparities between farmers are intensified between the minority who benefit from the extension services and those who do not. There are probably similar negative effects, in terms of equity, in education (disparities between educated and non-educated farmers), but the analysts are not looking at things from the same angle. For the extension services, the comparison is between farmers who are now receiving extension services and other farmers who might possibly, but do not yet, receive such -5- services. In education, the comparison is between farmers who in the past attended, or did not attend, school and the emphasis is no longer on the reasons for the inequality of access to education. Evaluations made by economists are concerned principally with the effects of extension services on production, productivity and income. Evaluations conducted by sociologists, educators and communication specialists are primarily concerned with the problems of the organization of extension services (internal efficiency) and of distribution and equity. Sections II and III of this study deal with these two different types of evaluation and their findings. The first section is devoted to certain problems common to both, viz. objectives, context, the time factor, and ideology. 1. GENERAL PROBLEMS (1) Oblectives of extension services The primary objective of agricultural extension work has long been the dissemination of innovations to increase productivity (cultivation methods, mechanization, new varieties, use of fertilizer, irrigation, new crops, etc.). The aim has been to move from the traditional type of farming, characterized by a small volume of trade (limited purchase of products outside the agricultural sector, limited sales of agricultural products to other sectors), to a new type of agriculture with strong links to the economy as a whole. This is the extension service model that has been an institution for the past 100 years in the industrialized countries. The introduction of this system into the developing countries during the last thirty years has encountered quite a different set of circumstances. The agricultural sector in these countries still includes a large proportion of the working population. Potential exchanges with the -6- other sectors are necessarily limited, since these other sectors exist only in an embryonic state. The extension services have therefore turned their attention towards promoting export crops. The outcome of extension services has been a source of much disillusionment. The market prices of agricultural exports have fluctuated considerably and have generally remained low. Export crops have often been developed at the expense of subsistence crops, sometimes leading to food shortages among peasant families. In other words, the innovations have not been without cost and the corresponding benefits have not necessarily gone to those who have borne this cost. These problems have led to the definition of new objectives and new goals for agricultural extension services. In the 1960s, and more markedly in the 1970s, the extension services have gradually come to concern themselves with subsistence agriculture and with problems of distribution and equity. This new approach presents a formidable challenge to the extension services, since one may well ask whether it can ever be achieved, particularly where the question of equity is concerned. In traditional peasant societies where there have been few disparities, it is difficult for any intervention by the extension services to avoid disequalizing effects; even where these services reach all the farmers and provide appropriate advice, they cannot guarantee that all will benefit equally from their advice nor can they force reluctant farmers to follow this advice. An evaluation made after the lapse of a few years of extension activity will almost inevitably reveal increased disparities among farmers. Where there are already marked disparities, the extension services, even if concentrated on the most disadvantaged, can hardly prevent the better- off farmers from taking better advantage of the services (because they -7- generally possess, on the average, more land, more capital and more education). The objective of strict equality among farmers does not necessarily lead to a general development process, which requires the expansion of non- agricultural sectors. This expansion requires a transfer of labour from the agricultural sector to other sectors, and an increase of productivity by the remaining farmers to feed people working in other sectors. It may be that more equality among poor farmers hinders the necessary mobility from the agricultural sector to other sectors. From a larger equity perspective, temporary inequality among farmers may accelerate the development process. It must not be forgotten that in the developed countries, in recent decades, the achievement of parity of income between farmers and other social categories only became possible when the agricultural sector came to account for less than 10 of the working population, when the birth rate among rural families fell to the national average, and when agriculture became capital- intensive. Rural development, more equal income distribution, and the fight against rural poverty in the Third World, are not strictly agricultural issues. It must be integrated in a wider development perspective where labor mobility is encouraged and not considered a negative outcome. (2) The context The introduction of agricultural extension services is one factor among others that can promote increases in agricultural productivity. Where a combination of favorable factors already exists and only a catalyst is lacking, the project may have spectacular results. Conversely, in a situation where everything is lacking (land, capital, suitable seeds, fertilizer, water, distribution and storage arrangements, markets, etc.), an extension service, -8- however well organized, cannot work miracles. If, in addition, it suffers from weaknesses itself (lack of qualified staff, lack of motivation, lack of the means of transport and of political support, etc.), then the impact will almost certainly be negligible. The problem is thus less one of showing whether agricultural extension services can have an impact, positive or otherwise, than of determining under what conditions a particular extension project has a reasonable chance of success. (3) The time factor Agricultural extension services act by disseminating information. The information is presumed to be of use in increasing agricultural productivity. Extension services are not the only means of disseminating this Information; there are other vehicles such as specialized periodicals, radio, television, training courses, the example of other farmers, organized visits by farmers, membership in professional associations, etc. If extension services are introduced, it is because they are assumed to be a rapid and effective means of disseminating such information and ensuring that it is put to use. It is a means of gaining time. In the 196Os certain economists like Schultz (1964 and 1966) postulated rational economic behavior on the part of the traditional peasant, whereas sociologists like Rogers (1966) saw the traditional peasant as an individual whose behavior was determined by non-economic factors. Rogers drew up a list of character traits that represent the cultural heritage of peasant societies: mutual distrust, lack of inventiveness, fatalism, a low level of expectation, lack of understanding of deferred advantage, a limited horizon, strong family dependency, parochialism, and a lack of empathy. In support of his description of traditional peasant behavior, Rogers quotes studies by Nair (1962): "The lower the level, the more static the aspirations tend to be," -9- and Rosen (1964): "Where poverty is widespread and opportunity for advancement very limited ..., the belief that the individual has little control over his environment is perhaps inevitable and probably psychologically functional." These two analyses are not contradictory. When Schultz observes the economically-rational behavior of traditional farmers, he is not assuming that there are not obstacles, but that the socio-psychological obstacles analyzed by Rogers and others can be overcome when a better environment, adequate infrastructures, and appropriate economic incentives are provided. Schultz also has a longer time perspective and knows that sometimes several generations are necessary to take advantage of certain innovations (Schultz, 1966). It is true that in most cases the evaluation of rural development projects calling for the widespread use of innovative methods is conducted after too short a lapse of time for the benefits of the project to be seen in the case of the majority of farmers. If it takes 20 or 30 years for an innovation to become widely adopted, as in the examples quoted by Schultz, this is beyond the time horizon of decision makers or project evaluators. As Cancian (1972) pointed out (quoted by Barlett (1980)): "The question whether Zinecanticos are economic maximizers or prisoners of tradition ... is a bogus question that leads to scientifically incorrect and politically dangerous descriptions of peasant societies." Barlett adds that during the 1970s, "we have moved away from simple questions with yes or no answers and are now pursuing issues by seeking to determine when farmers behave as one would predict from strict maximization formulae and when they do not .... Our emphasis has moved beyond the consideraton of achieved versus ascribed characteristics of persons to a study -10- of relationships .... Such a perspective leads not only to a measurement of the diversity of behavior within rural communities but also to a clarification of, and sometimes to the measurement of, the variables that interact to produce the behavior outcome." The key variables differ, according to the particular case: access to material and non-material resources, access to land and inheritance, access to water, etc. Barlett concludes that "in-depth studies give conclusive proof that simple solutions and worldwide verities are doomed to disappointment. Local realities will always distort the mathematical curves, and no simple model can hold up past class, kinship, ecological and governmental differences." (4) Ideology Rural development is a sensitive subject from the ideological view- point. Certain writers take a normative approach and indulge in a priori analyses which reveal that their value judgments are not in accordance with the project goals. The major criticism addressed to rural development projects in Third World countries is their "western" character, on the grounds that: - the projects are conceived along the lines of development modes experienced in the West; - they are conceived by Western "experts"; - they are designed to promote the interests of the Western countries. Although in the case of a few writers, the notion of "western" is opposed to "eastern", i.e. the development model promoted and illustrated by the countries of Eastern Europe, particularly the USSR (collectivization of agriculture and state farms), this is no longer true today of the majority of -11- the writers whose approach we have described as ideological. In fact, the "eastern" model has ceased to have any relevance since the results achieved in the domain of agriculture under this system have not been convincing, even in the eyes of its most ardent supporters. Writers who belong to this ideological current nowadays contrast the "western" model with a rather confused model, which combines a clinging to the past, defense of traditional modes of life and production, ecology, opposition to the multi-national cooperations that control the world market for agricultural products, restrictions on trade with the West (domestically- oriented development), and an equal (or more equal) distribution of inputs and outputs. The most brilliant advocate of this line of thought is the agronomist Rene Dumont (1962 and 1980) who has had extensive experience in the field and does not rely solely on a priori criticism, as do many of his followers. The main thesis of these writers is that most of the rural development projects based on agricultural extension services are aimed at promoting the growing of crops intended for the commercial market, and primarily for the world market, and that these crops are developed at the expense of food crops. According to these writers, the beneficiaries of this system are therefore not the farmers, nor even the urban populations, but the governments which collect additional taxes, the commercial companies which make a profit, and the developed countries which obtain their supplies at low cost. It is argued that as a result, the farmers (and even ordinary urban people) and their families are increasingly ill-nourished. The phenomenon as a whole is described as "plundering the Third World". Cf. Dumont (1980). The root of the "evil" lies in the "industrialized society, which is bringing the -12- Third World even closer to starvation." (Dumont, 1980, page 31). It is true that some rural development projects have had, and still have, a negative impact (especially some unwise cash-crop oriented projects). But now, they are improving and the generalization is no longer true. If one takes aggregate data, instead of examples or anecdotes, one finds exactly the opposite: the more a country increases its exchanges with industrialized countries, the more it achieves a high rate of growth and the more it improves the calorie intake of its people. The food consumption data (cf. Annex, Table 1) reveal that in the so-called developing countries, between 1966 and 1977, daily per capita consumption of calories declined in four or five of the poorest countries (per capita income of less than $120); in 16 (i.e. 57%) of the 28 countries with per capita incomes between $130 and $360; in 6 (i.e. 23%) of the 26 countries with per capita incomes of between $390 and $1,050; and in 0 of the 29 countries with per capita incomes of $1,090-$3,500. In other words, the poorer the country the more likely that the food situation will deteriorate. Is this trend the result of "plundering" of the Third World, as claimed by these writers? If we look at the data on foreign trade 1/ we find that exports by the group of 38 low-income countries declined in volume at a median rate of 0.8% per year from 1970 to 1978, whereas exports by the group of 51 middle-income countries grew at a rate of 5.2% over the same period. Thus, the countries whose food situation improved are those who increase their exports the most. Further, the terms of trade index is somewhat more favorable to the low-income countries (98) than to those in the middle-incoae 1/ Data taken from the World Development Report 1980, World Bank, Washington, D.C. -13- group (90) and even to the developed countries (95) for the same period 1970- 1978; the only beneficiary group in terms of trade was the capital-surplus oil exporters (393). It is difficult, on the basis of these data, to maintain the hypothesis of a cause-effect relationship between purchases of raw materials by the developed countries and the deterioration of the food situation in the developing countries. If we take specific examples, as do Berthelot and de Ravignan (1980), we find many agricultural development projects that have had adverse effects on the local population, but on a macro-economic level the cause and effect relationship does not stand up, and more recent projects, lessons of the past having been learnt, are increasingly able to avoid such pitfalls. An obiective analysis of agricultural development projects sponsored by the World Bank, for example, reveals an increasing concern to prevent the unfortunate secondary effects observed in the past. Payer (1979) posited as the real purpose of World Bank loans to small farmers as to persuade them to buy costly industrial supplies (to increase profits in these industries); to expand the technical assistance market (creating opportunities for "experts"); and to oblige farmers to pay taxes; such a statement, whether intended as ideological, polemic or simply naive, has no foundation in rigorous observation of facts. To claim after this that the farmers are happier under a system of subsistence farming and that the agricultural extension services are an instrument of control leading to loss of freedom reflects a value judgment which is no less inspired by western ideas than is the mode of development denounced by Payer. -14- II. INTERNAL EVALUATION OF AGRICULTURAL EXTENSION SERVICES As we noted in our introduction, the internal evaluation of agricultural extension services is dominated by sociologists, educators and communication specialists. It is principally concerned with the interaction between the extension services and the farmers with the obstacles to effective interaction, with underlying philosophies, and with the effectiveness of the extension services in disseminating innovations. (1) The underlying philosophies of agricultural extension projects As Malassis points out (1975), the agricultural extension services follow two principal models: the first, which he calls the "supervisory" system (vulgarisation par "encadrement") employs more or less directive methods imposed from above by technocrats, politicians and "leaders", whose aim is the achievement by farmers of a number of set targets. Active methods of persuasion are used but not coercion. The second method, on the other hand, emphasizes participatory decision-making. In the French-speaking world it is known as animation, whereas in the Anglo-American literature the term "community development" is preferred, and in Hispanic contexts, the term "promotion". It needs hardly be said that an overwhelming number of the writers, including Malassis himself, prefer the second formula. Moreover, although the majority of the extension projects undertaken in the Third World between 1950 and 1970 have only a minimal directive component, they are criticized by a majority of the evaluators for paying too little attention to the participation of the farmers themselves. It should be noted that in this respect Dumont, who in most other ways is regarded as a radical, argues that "democracy comes after development". Paulo Freire (1973) goes so far as to take the view that, however -15- they are conceived, agricultural extension services are a tool for the "manipulation" of farmers, with all the pejorative implications that this has. The paradox is that, in face of so much criticism, the only real effort made in the seventies to revitalize the agricultural extension services--Benor and Harrisson's "Training and Visit System" (1977)--embodies the most radical version of the supervisory system so far. So far as participation by farmers is concerned--empirical evaluators having invariably found that there is too little--the concensus appears to be that the more there is of this, the better. However, to quote Oxenham and Chambers (1978): "Too little may lead to chaos Too much may end in inaction;" or Jamison-McAnany (1978): "The non-directive nature of the animation approach can mean endless dialogue and little action." In our view it is a question of determining the participatory input appropriate to the particular case rather than one of some arbitrary and absolute level of participation. Participation has a price in terms of time and transportation and, over and above a certain point, its impact is undoubtedly negative. Chambers and Oxenham, who are generally very much in favor of increasing the role of the farmers, nevertheless recognize that "representative" participation is frequently the only possible route to take in face of budgetary constraints. Belloncle and Gentil (in Lele, 1977) report that in Cameroon the agricultural extension project has adopted a highly participatory approach and this appears to have been successful although, from the economic standpoint, the complete absence of any analysis of cost-effectiveness leaves the reader -16- in doubt. The vital question is therefore whether, faced with such budgetary constraints, it is better to involve 100 farmers in a highly participatory approach or 1,000 farmers under a supervisory extension system. Ruthenberg (1973) is inclined to favor a third approach, which he describes as "contractual". It simply amounts to a contract between volunteer farmers and the development/extension service agency, under which the farmer agrees to follow instructions, and the agency, to provide him with technical assistance and market outlets. According to the author, this method has been employed successfully in a tea-growing project in Kenya and a tobacco-growing project in Tanzania. Ruthenberg believes that this formula is useful and effective in the case of the introduction of farming practices whose significance is not apparent to the farmer (with long-term implications, such as soil erosion). Moris (1966), having seen the ineffectiveness of extension services in Kenya and Uganda, strongly recommends the solution of Farmer Training Colleges. Two surveys of 171 and 165 farmers trained in colleges in Kenya and IJganda respectively have shown that all the trainees improved their performance as a result. Here again we must object to Moris' conclusions on the grounds that they are not based on cost-benefit analysis. If, for the same budgetary funds, it is possible to train 100 farmers in colleges or assist 10,000 through an agricultural extension service, it is essential to determine the specific benefits attributable to each of these methods. (2) Evaluation of the internal effectiveness of agricultural extension services As we indicated earlier, the literature is extremely discouraging. Rice, for example (1971), who evaluated 13 agricultural extension projects financed with American aid in Latin America, concluded that they had all been failures. We shall now proceed to consider the criteria of effectiveness -17- adopted to determine success or failure, the difficulties identified by writers to explain the failures and, finally, the factors complementary to, and/or that could be substituted for, agricultural extension services. (a) Criteria for effective performance In general, the criteria for determining failure are of two types: either the extension service has contacted very few farmers in relation to the targets set, or those that have been contacted are not those to whom priority should have been given. -- Coverage of farmers Depending on the project, the number of farmers per extension agent varies between 100 and 2,000, but is between 500 and 1,000 in the great majority of cases. The evaluation reports calculate the proportion of this target actually covered by extension agents in a given year as well as the average number of contacts between agents and farmers. It is invariably found that only a minority in the farming community has had effective contacts with the agents and that those covered have been contacted less frequently than would be necessary to ensure the assimilation of all the new farming practices to be introduced. Giltrow and Potts (1978) estimate, in the case of a Nigerian project, that only 3% of the farmers have been in touch with extension agents. Chambers and Wickremanayake (1977) report that surveys made in India and Sri Lanka have shown that the percentage of farmers covered has been relatively high (60% to 80%) but, according to them, these surveys must be subject to reservations as the regions in which they were organized are unrepresentative (more highly developed supervisory structure than on the average, unusually well-served villages, farmers more advanced than the average). Chambers, in correcting these samples, found that the percentages -18- of farmers actually covered were of the order of 16-21%. Perraton, Jamison and Orivel (1981) have shown in a survey in Malawi that 11% of the farmers surveyed were in contact with the extension services where the services were "thin" and 30-60% where they had been reinforced. In fact, the extension system needed to cover a large majority of the farmers on a regular basis would make extremely heavy budgetary demands. It has been found that, when a well-selected minority of farmers has been covered, there is a progressive dissemination of new farming methods amongst the others. Benor's Training and Visit System moreover exemplifies the institutional application of this process, since it is based on a network of contact" farmers explicitly responsible for the spread of new methods amongst the others. This kind of approach should always be made sensitively for, as Belloncle and Gentil (1967) have shown in Cameroon, the criteria used to select such "contact" farmers are not always those that produce generally acceptable leaders. In this particular instance, the choice had been made on the basis of their knowledge of French, with the result that they were sometimes unacceptable to other members of the farming community. -- Type of farmers reached by the extension services These are not only a minority but they also tend to be among the better-off members of the farming community and therefore those who have the least need for extension services. According to Chambers and Wickremanayake (1977), their holdings of land are above average (no data). Similarly, Lele (1977) reports that in the CADU project in Ethiopia, the farmers were on average more affluent (more land and livestock); more often than not they belonged to farmers' associations; they were better educated and younger. Leonard states (1973) that 57% of extension agents' visits are made -19- to the most advanced 10% of the farmers and only 6% to the most traditional 47% of these. According to IDS (1978), the Green Revolution in India has strengthened the domination exercised by large farmers over small ones, since it is the former who have adopted the high-yield varieties, become relatively wealthy and been able to increase their holdings at the expense of poorer farmers. Bhalla (1976) notes that, broadly speaking, the Green Revolution has had an unfavorable impact on the rural landless through an erosion in established clientele patterns (an elementary form of social protection) and a reduction in payments in kind. Lenglet also argues (1980) that it is the more dynamic elements in the rural sector that make the most effective use of the inputs provided by the rural development services in the Ivory Coast. Dubey (1962) has shown that it is in the interest of an extension agent to establish good relations with local persons of consequence (as their hostility could involve him in disputes with his own superiors) and therefore to give priority in his activities to locally dominant farmers. On the other hand, Singh (1979), commenting on the training and visit system, observes that initial evaluations show favorable results in the case of small farmers with only 1 or 2 hectares. Nevertheless, evaluation studies made shortly after the implementation of agricultural extension projects generally show that the initial beneficiaries are amongst those farmers who are least deprived. In the longer term, we can assume that the poorer farmers have finally adopted the innovations (as a result of dissemination and progressive assimilation) or that they have left (rural exodus). It is likely to be cheaper to persuade the more advanced and energetic farmers to adopt new methods of cultivation -20- and there may be a conflict between the goals of effectiveness and equity. (b) Obstacles to the effectiveness of extension services These fall into two categories: those that hinder the efficient internal operation of services, and those that prevent the introduction of innovations. (i) Internal obstacles Ratio of agents to farmers When there are too many farmers to each agent, the effectiveness of the latter is reduced. On the other hand, increases in coverage are costly. The Benor system, which provides one agent for every 600 to 700 families, supported by farmers responsible for training other farmers, appears a good compromise from the economic standpoint. Many of the writers, however, recommend more active supervision (one agent for 150-200 farmers). Training of extension agents Inadequately trained extension agents are incapable of propagating needed innovations. They cannot answer the technical questions raised by farmers, nor can they deal with unforeseen situations (such as crop diseases). On the other hand, a very high level of training does not necessarily produce the best results. Lele (1977) reports that, in Kenya, Leonard showed that extension agents with a secondary education were less capable of handling agricultural problems than those who had only completed primary school supplemented by agricultural upgrading courses, succeeded less well with the introduction of new cultivation methods, were less highly motivated in respect of agricultural extension projects and promoted urban scales of values. -21- Lele also indicates that Harrisson's findings in Nigeria were similar. Location of extension agents Many observers (more particularly Chambers) have pointed out that the further a village is from the place of residence of the extension agent, the less likely it is to be visited by him. Transportation problems are sometimes difficult to overcome. Travel is costly in terms of both time and money. If an extension agent is overloaded, he will tend to see more farmers close by than at a distance. In some agricultural extension systems a group of extension agents resides in an office building located in the administrative center for the district. In others, they work on their own and are scattered amongst the villages for which they are responsible, visiting the head office periodically to keep in touch, obtain training, receive instructions and report on their work in the field. In Lele's view (1977) the empirical studies that have been undertaken on this problem provide no clear answers. Lisenmayer and Agarwal have shown that in the Tanzanian tobacco-growing development project the results were better in the Tumbi region, where the extension agents lived in the villages, than in the Urambo region, where they lived in the administrative center. On the other hand, Leonard found that in Kenya non-resident agents made more visits than resident agents, that they were more effectively controlled by their superiors than those who lived in isolation in the villages, and that those who stayed in villages were exposed to more tribal and family obligations. Rahim (1966) describes the difficulties encountered by an agent -22- living in the village in a project in Pakistan. The villagers regard him with mistrust and suspicion. Living conditions in the village are difficult and austere. There is no health service, no education service and no decent lodging for himself and his family. Hursh, Niels and Kerr (1968) also noted that, in Nigeria, the extension agents living in villages were too cut off from their superiors and very inactive. Remuneration of extension agents What level of remuneration is needed to obtain effective extension agents? Many observers report that wages are frequently low and provide an explanation for the lack of motivation and apathy of many agents. On the other hand, Belloncle and Gentil (1967) state that, in Cameroon, wages were high in relation to the level of earnings of the farmers for which they were responsible and that this disparity in earnings created a psychological barrier that made useful contacts more difficult. Kabure (1979) evaluated the agricultural extension systems in six West African countries (Benin, Mali, Niger, Senegal, Togo, and Upper Volta) and concluded that under-remuneration of extension agents and failure to provide them with the means of transportation are the main cause of their lack of motivation and the rapid turnover of agents, with the consequent poor level of effectiveness. De Vries (1976) made the original suggestion that there should be a link between the remuneration of extension agents and the results achieved by them. It therefore appears that so far as the training of agents, their place of residence, and level of remuneration is concerned, the experience so far obtained does not point to any obvious conclusions. In each case the answer appears to depend on the particular circumstances. -23- The nature of the role of the extension agents There are two conflicting theories here. On the one hand, there are those who believe that an agent should concentrate his efforts on visits, advisory services and demonstrations and, on the other, those who think that he should act as a coordinator between the multiplicity of agencies and departments concerned with rural development, i.e., the "multipurpose worker". He should take the place of a local retailer when the seeds, fertilizers or pesticides he recommends are not available, repair the pump when it breaks down, organize the marketing of crops, undertake surveys, complete questionnaires, distribute contraceptives, etc. Underlying this second theory is the observed fact that rural development agencies act in a disorganized fashion and without sufficient coordination, resulting in costly overlaps or inaction. (Is this or that livestock epidemic the responsibility of the health authorities or the agricultural extension service? In the end, nothing is done to control it.) It was soon found, however, that this idea of a "multipurpose worker" led to overloading the extension agent with bureaucratic functions and allowed him very little time for visits. Most evaluators now consider that the administrative responsibilities of extension agents should be kept to a minimum. (ii) External obstacles Advice given by extension agents falls into two categories: - it either involves the use of new inputs; or - a reallocation of existing inputs. In the first case, we shall ignore for the moment the problem of land, its distribution and the very limited areas available to many farmers in Asia and, to a lesser extent, in Latin America. Many critics observe that, if -24- the results of agricultural extension projects are so limited in scope, it is because the problem of a better distribution of land has not been previously solved. This is probably true in a certain number of cases, although we would neverthess like to point out that: - In South Asia, where, however inequitable the distribution of land may be, the per capita areas available are small, the primary goal of extension activities is to increase average yields and it is by no means clear that a different (more egalitarian) distribution would contribute to this. - In general, a more egalitarian distribution of land, associated with an agricultural extension project, can only increase average yields more rapidly than the initial distribution to the extent that large landowners have been "wasting" the land they possess, i.e., have average yields below those of small landholders. This is frequently the case in Latin America, and more particularly in Brazil, where large landowners leave a large part of their estates uncultivated. But where this is not so, failure of an agricultural extension project cannot be attributed to the inequitable distribution of land. Besides the land, the inputs recommended should be within the grasp of farmers. Purchases of fertilizers, seed and pesticide should cost significantly less than the value of the increased crop it is expected to obtain. A number of writers point out that various recommendations have not been adopted in the absence of a suitable system of credit. Others draw attention to the absence of a marketing network on a scale adequate to dispose of the additional crops expected. Some projects, such as Masagana, in the Philippines, have so successfully increased output and yields that prices have collapsed--a boon for urban consumers but not -25- necessarily for the farmers. Recent agricultural extension projects, more especially the training and visit system, have made a more positive effort to limit recourse to new and costly inputs, whilst focusing more and more on the development of food crops, which present fewer marketing problems and directly benefit the families of the farmers involved. (c) Complementary and/or substitution factors Three factors may complement or substitute for agricultural extension services: - applied agricultural research services; - the education of the farmer; - the mass media. (i) Applied agricultural research services Evenson (1976 and 1978) showed that, in the breakdown of total expenditure on agricultural research and agricultural extension activities, the share of research was significantly smaller in the Third World. He believed that this could in some measure explain the slow growth of agricultural productivity in these regions and might be one of the reasons for the low level of effectiveness of agricultural extension services, which are unable to provide a wide range of advisory services adapted to the needs of different agricultural micro-regions. Malassis (1975) recommends that basic research (belonging as it does to the world at large) be left to the affluent countries and that the research capabilities of the developing countries be concentrated on applied research. In fact, the distinction between applied and basic research is artificial. Where is the dividing line? Can one be certain that the emergence of specific agricultural innovations in tropical regions does not -26- depend on specific choices further back in the field of so-called basic research for tropical conditions? Many of the writers recommend the development of technologies and methods suited to local conditions rather than importing them from the developed countries. They complain that the research stations are excessively concerned with export crops and pay insufficient attention to food crops. This complaint has often been well-founded but the current trend is now toward an increasing orientation of research, of rural development projects, and of advice from the agricultural extension services to food crops. The development of cash crops by small farmers need not necessarily be at the expense of food crops, as Perraton, Jamison and Orivel (1981) have observed in Malawi, where there is a positive correlation between food production per family and cash crop production. (ii) Education of farmers Jamison and Lau (1980) and Lockheed, Jamison, Lau (1980) have examined the relationship between the educational level of farmers and their productivity. They produce solid evidence for the positive effects of education on production, productivity and farmers' income. The rate of return on educational expenditures for farmers is no lower than that noted for other types of employment. Most writers consider that the process by which education makes its impact is better allocation of resources (allocative efficiency). The better- educated farmers make a better choice of inputs, combine them more effectively and judge the appropriate quantity better. Ram (1980) adds with pertinence that this better allocation of resources is due to the fact that education reduces the marginal cost of information (information that is useful in selecting the most productive -27- combination) and thus increases the marginal benefits of such information. The drop in the marginal cost of information stems from the fact that educated farmers communicate better, establish better contacts, have a wider vocabulary, can read written information, etc. Lastly, educated farmers are better at making overall judgments and are more open to innovations. Information Extension Services Radio Allocative Radio ~ ~ Ffiaecyo Prdci Education Printed Matter En]fficiency of Ioction Contacts with more Resources innovative farmers The effect of education is thus twofold: a direct effect, in that it improves the individual's judgment in the choice of the optimal combination of inputs, and an indirect effect, by improving the individual's ability to obtain access to information pertinent to the choice of this optimal combination. On this basis, education and agricultural extension could be regarded as complementary. This is in fact the usual finding of probability analyses, but not in multiple regression analysis. Educated farmers are more active than the others in seeking information, particularly such information as they can obtain from extension services. According to Lele (1977), in the CADU project in Ethiopia, the best educated farmers benefit most from the extension services. Perraton, Jamison, Orivel (1981) arrived at the same findings in -28- Malawi: educated farmers are twice as likely to be in contact with extension agents. Harker (1973) has shown, on the basis of a Japanese sample, that the better educated the farmer is, the more actively he seeks additional knowledge. He also noted that graduates of agricultural schools made more use of the services of extension agents than did graduates of non-vocational schools. Lee (1970) has found the same in the case of a Korean sample. These results are consistent with the assumption that more educated farmers tend to be more active in seeking relevant information, but the kind of information depends on availability and personal experience. Graduates of agricultural schools who have been accustomed to agricultural institutions, have likely met extension people during their training, and go naturally to them when they become farmers. Graduates of general schooling institutions do not have these connections, but prefer to have access to agricultural information through books, journals and newspapers. Coughenor (1960) came to a slightly different conclusion--the better educated farmers are, the more they tend to supplement their knowledge from written material rather than by consultation with extension agents. On the whole, education tends to increase the probability that farmers will seek additional information more actively, and often such information is obtained from extension agents. But this tendency does not necessarily imply a specific combined effect of education and extension, as shown by mixed results from multiple regression analysis. Hayami and Ruttan (1971) report a positive finding, whereas Moock, Jamison and Shrestha (1981) found that the combined effect, although not significant, was negative (a Nepalese sample); this raises doubt as to the multiplier effect of the two factors but does not invalidate the conclusion -29- that each may have a positive effect and that education increases the likelihood of the farmer benefiting from the extension services. Other studies (Baidya, Chou, Jamison, Moock and Shrestha (1980), Hong (1975), Lockheed, Jamison, Lau (1980) and Moock (forthcoming)) tend to confirm that education and extension services are substitutes for one another rather than complementary. It should be noted, however, that the interaction of the two factors, whether the coefficient of regression is negative or positive (- or +), is not statistically significant; it might be prudent to conclude that the interaction of the two factors does not have an effect in itself. (iii) Agricultural extension activities and the mass-media Agricultural extension services employ labor-intensive methods, based essentially on personal contact and demonstration. Whatever the quality of organization and management, this is still a relatively costly approach. There are less expensive means nowadays of spreading information, by the written word, television, radio, films and the telephone, and in the developed countries these are the chief means of access to information available to farmers. These means are only less costly, however, if the necessary infrastructure already exists and if the agricultural information programs they carry only have to bear the marginal cost of the technology involved. Radio and television are only cheap if agricultural programs occupy no more than a small proportion of daily broadcasting time (about one hour), so that the greater part of the amortization cost of the radio network is borne by others. Similarly, radio or television receiving sets must already be available to farmers so that the marginal cost of receiving the programs is negligible. -30- Perraton, Jamison and Orivel (1981) made a study in Malawi of the respective costs of the different channels of information available to farmers: extension services, radio, printed matter, training courses (resident or non-resident), film shows and puppet shows (with mobile teams travelling from village to village). A varied and more or less complete range was thus covered. The costs were compared with the quantity of information obtained by the farmers through each of these media. Radio is far and away the least expensive channel for the amount of information received, and the extension services by far the most expensive. This does not necessarily mean that the extension services are not worthwhile, since certain types of information can probably only be disseminated and understood through such a system, particularly where the farmers are illiterate. Very few studies have been made to try to establish what is the best combination of media, or to determine what type of information is best suited to each medium. It seems probable that, on the average, one hour of listening to the radio is less effective than a one-hour visit with an extension agent. But if the cost of the extension hour is 1,000 times that of an hour listening to the radio (in Malawi the multiple is 3,000), the cost-efficiency ratio for radio is likely to be extremely favorable. Many developing countries do not have the necessary financial means to provide intensive extension services on a regular basis to rural areas, and in some cases radio might be the best substitute from the economic standpoint. Radio, moreover, is a very valuable instrument for transmitting information to the extension agents themselves, reducing the need for visits to headquarters that are costly in terms of both time and money. Two studies (McAnany, 1980; and Jamison, McAnany, 1978) offer numerous examples of the use of radio in rural development projects. They -31- show that this is now one of the least expensive technologies and, if used judiciously, could have a significant impact. Nevertheless, it has not been definitely shown that radio alone can have a decisive effect on the adoption of new agricultural methods. Shore (1980) has found, from a sample of Indian farmers, that although radio was the medium least inequitably distributed (76% of the farmers had access to it), its use had had no significant impact on the introduction of innovations. Rogers, E.M. (1970) emphasized the effectiveness of group listening: "The efficiency of radio forums in diffusing innovations is much greater than either literacy, reading classes or community newspapers, whether cost is measured for peasant receivers or for change agencies." In our opinion, radio is an ideal complement to extension services. It can permit a reduction in the ratio of extension agents to farmers, if the task of disseminating information is judiciously divided between the two systems. Radio should be an integral part of the extension service to ensure the best coordination. This does not mean that the extension service should have its own radio station since it is much less expensive to use the transmitter of an existing station part-time. Conclusion We have seen from this survey of the internal effectiveness of agricultural extension services that certain writers call in question the oblectives of rural development backed by agricultural extension services. Most of these writers have in common a non-directive philosophy, but since they fail entirely to consider the costs of the non-directive approach these arguments remain somewhat rhetorical. Although for a long time the aim of greater equity was not a major concern of agricultural extension services, critics have gone into this aspect in great detail and have pointed to an -32- increase in disparities between farmers, without stopping to inquire whether these greater disparities were an inevitable result of the development process itself or were specifically attributable to the action of extension services. The generally pessimistic conclusions of writers who have engaged in these internal evaluations stem from the fact that they have all too often left out of account the external impact of extension work, i.e. its impact on production, yields and income. By this very omission, however, such critics have played a useful part in drawing the attention of decision-makers to certain obstacles to both internal and external effectiveness of the services, and at the same time to some factors that could increase project effectiveness. III. EXTERNAL EVALUATION OF AGRICULTURAL EXTENSION SERVICES The external evaluation of the impact of agricultural extension services is essentially based on economic analysis. It followed, in point of time, fairly closely on the methodology used to estimate rates of return on education, and it is not just by chance that this field of knowledge, like the theory of human capital, owes its existence to T.W. Schultz. The dissemination of innovative techniques in the agricultural sector to increase productivity implies the prior development of these techniques, and consequently the existence not merely of extension services but also of research institutions. The writers primarily concerned with external evaluation have therefore also examined the state of agricultural research, an intermediate factor necessary for the proper functioning of the extension services. In this third section, we propose to make a rapid review of the methods employed by these writers and the empirical findings to which the -33- different methods have given rise. The description of each method and the related empirical studies is followed by some general comments. (1) Methods Internal rate of return. The internal rate of return equalizes discounted costs and earnings over a period t , the earnings here being measured by the inputs saved as a result of the new production technologies introduced. The cost-benefit method, which is similar, involves the parallel measurement of the costs and benefits of the program. Correlation at the regional level (preferably micro-regional but also intercontinental) between the expenditure on agricultural extension services and progress in agricultural productivity, in order to establish whether the areas that spend most make significantly more progress in productivity than the others. Comparisons of yields (time-series or cross-section). This is a comparison of average yields before and after the introduction of extension services or as between regions with and without such services. Micro-economic production functions, taking as a unit of observation the farmer himself and used to measure the marginal return on use of agricultural extension services. The specification of these functions may vary ("Cobb-Douglas" functions, profit functions, constant elasticity of substitution, etc.). The independent variables may either be introduced directly or as logarithms or as dummy variables. (2) Results Internal return and cost-benefit methods (a) Research alone - Griliches (1958). Rate of return of 700 for hybrid maize and of 360X for hybrid sorghum in the United States. -34- Peterson (1967). Rate of return of 140% for poultry in the United States. - Ayer and Schuh (1972). Rate of return of 90% for cotton in Brazil. Barletta (1967). Rate of return of 75% for wheat in Mexico. - Akino and Ayami (1975). Rate of return of 75% for rice in Japan. - Bredahl and Peterson (1976). Rate of return of between 36% and 40% for various products in the United States. (b) Research with extension service activity - Schultz (1971) has shown that in 1950 in the United States, agriculture sector input savings alone amounted to US$9.6 billion, whilst cumulative research and extension service costs from 1910 to 1950 totaled US$7 billion. Peterson (1971) calculated that the US$9.4 billion devoted to research and extension service activities in the United States between 1910 and 1967 enabled a saving of US$25.9 billion to be made in 1967. (c) Research with or without extension service activities - Araji, Sim and Gardner (1978) estimated the rates of return for 9 agricultural products in the United States. These varied between 33% and 105% (with the exception of cotton). The absence of extension services would reduce these rates by between 25 and 60% according to the product. (The fall is lowest for products controlled by a small number of large producers). (d) General observations on these studies The high rates of return on agricultural research with or without extension service activities are due to the fact that all improvements in productivity in the sector are attributed to them, although it is clear that these are the result of a complex series of factors: -35- progress in the mechanization of agriculture; progress made by industries manufacturing agricultural machinery; improvements in technology and management skills; increased education and higher qualifications of farmers; progress made by the chemical industries producing fertilizers, pesticides and weedkillers; progress in irrigation techniques; impact of the expansion of agricultural structures (increase in the scale of farms); upstream progress in fundamental research. All the improvement in productivity cannot be attributed solely to these programs. Moreover, there must certainly have been some unsuccessful programs which have probably received less thorough evaluation. Method of geographical comparisons The approach adopted here has been to relate expenditure on research and extension services within given geographical limits (county or district, major region, the State, the continent) to improvements in agricultural productivity. The empirical studies have included the United States and India, on the one hand, and the entire world, on the other, divided into major continental regions. The two principal writers are Evenson and Huffman. (a) Results - Evenson and Jha (1973). Analysis of patterns of agricultural productivity in various regions of India between 1953/54 and 1970/71. Agricultural extension programs (with due allowance for their age and experience) combined with research activities are correlated meaningfully and positively with increases in productivity. - Huffman (1974). Analysis of patterns of agricultural productivity -36- in the United States corn belt between 1959 and 1964. Extension and education are seen to be substitutes in the optimum utilization of nitrogen fertilizers. - Cline (1975). Those regions of the United States that have most increased their spending on agricultural extension services are generally those in which the most significant changes in productivity have occurred. - Evenson and Kislev (1975). These writers have made numerous studies relating, by major region, agricultural research--measured in terms of the number of scientific publications--and agricultural productivity. For example, they have found a correlation coefficient of R = 0.93 for wheat (64 countries) and R = 0.85 for maize (49 countries). - Mohan and Evanson (1975). Study of a certain number of districts in India showing that between 1959/60 and 1970/71 those that introduced programs of intensive agricultural development were also those that experienced the most rapid changes in farm productivity. = -Evenson (1976). This study is the first to provide worldwide data on the cost of agricultural research and extension services. It shows that the poor countries devote less effort to research and relatively more to extension services. The author believes, however, that research shows a higher return and that the developing countries have not adopted the best strategy. - Evenson (1978). Regional American data on a number of agricultural products covering the 1949 to 1971 period. Extension service activities are generally regarded as a substitute for education of farmers. Their effectiveness is all the greater when the effort put into applied research is substantial. Estimated internal rate of return on extension service costs: 110x. - Huffman (1978). Data relating to four American states In 1964. The -37- effectiveness of extension activity appears to depend on the nature and combination of crops. Rates of return vary with the product from very ,positive" to negative. The level of effectiveness is higher in the case of crops than in that of stockraising. (b) General observations on these studies These studies show that the United States, which spends more than other countries on agricultural research and extension services, also has a more productive agriculture. Similarly, within the United States, those regions that devote the most energy to agricultural research and extension activities also have the most dynamic agricultural patterns. Outside the United States, only in India has research on regional data been undertaken. There, the results are less clear. These studies raise two fundamental problems: On the methodological plane, there is an element of uncertainty in causation of the same kind as is to be found in studies relating education costs to economic growth (cf. Eicher-Orivel, 1980). Is it the efforts made in support of agricultural research and extension activities that result in a highly productive agriculture, or is it the existence of a productive, rich and strong agricultural base that makes it possible to provide substantial resources for research and extension activities? In practice, the relationship between these two phenomena is very likely to be reciprocal, although it is impossible to say which comes first in the final analysis. It may well be that American farming draws its vitality and dynamic characteristics, in the first place, from a series of factors such as: - more favorable climatic conditions; - good natural irrigation (not everywhere, but in many regions); - systems of land tenure that provide each farmers with a -38- significant land area at low cost; - energy and drive of the farmers (the spirit of enterprise of the immigrants); - highly efficient agricultural machinery industry with low energy costs, etc. Such conditions have from the outset, irrespective of all research and extension activities, made of American agriculture an affluent and powerful lobby, which has subsequently been able to consolidate its advantages through investment in research and extension activities. There is nothing to indicate that a similar effort, undertaken in different contexts, in Africa or Asia, will produce the same results, just as it is clear that the substantial efforts made by Cuba and Albania in public investment in education have failed--despite the predictions of the Denisonian version of the human capital theory--to bring about rapid economic growth. - On the decision-making plane then, even if the relationship between cause and effect applies globally, these studies are of little use in indicating optimum levels of expenditure on research and extension activities. Even if we assume that Africa were to decide to double or triple its expenditure on research and extension activities--which would represent a major effort on its part--and that these measures did not result in any increase in the productivity of African agriculture, this would not in any way modify the ex-post facto conclusions of an Evenson-type study (1976) showing that the regions of the world that spend more have high levels of agricultural productivity, given the magnitude of the disparities that exist on a world place. Method of comparisons of yields (a) Results -39- - Ribeiro and Wharton (1966). This rural development project in four districts in Minas Gerais reports average yields per hectare for maize that are much above the average for the State: Raul Soarez 8,276 kg/ha Forminga 11,262 San Joao Nepomuceno 9,460 Tres Pontas 8,540 Average for Minas Gerais 1,300 - Myren (1966). This also was a project to develop maize growing, but results indicated by a comparison of the overall rate of growth of maize production in the two project regions and in the rest of Mexico were less spectacular and less consistent than the findings in Minas Gerais. Period Jalisco Vera Cruz Mexico Percentage increase in -1957 to 1963 127 147 113 maize production from: 1957 to 1964 160 132 136 1950 to 1964 230 134 143 1940 to 1964 307 136 181 These figures demonstrate dramatically the extent to which statistical results can be affected by the choice of the base year, the observation year, climatic variations, and other random variables. - Benor and Harrison (1977). Seyman irrigation project in Turkey (1967-72) for cotton growing: Yields (tons per hectare) Before the project After the project 1966 1969 1971 1.7 3.0 3.8 Lele (1977). In this remarkable and comprehensive study, Uma Lele -40- examines agricultural development projects in Africa primarily from a qualitative standpoint. Statistical data for yields is provided in the case of three of these: * CADU project (Ethiopia), 1966-71: - Little impact on yields save in the case of wheat (yields having risen from 1.6 t/ha to 2.0 t/ha). - Substantial increase in the areas under wheat, resulting in a significant increase in global production, a fall in wheat prices, the accumulation of substantial stocks and a shortfall in milk production, wheat-growing having taken over from animal husbandry. * WADU project (Ethiopia): Yields in t/ha Before the project After the project Uplands Maize 8.0 20.0 Wheat 9.0 17.0 Teff 5.5 7.0 Lowlands Maize 12.0 24.0 Cotton 1.4 10.0 1.0 9.0 The success of this project was much more marked than in the case of the preceding one. According to the c'mmentators, this is due to: - the organization of a sound credit system; - the natural fertility of the soil; - the motivation and performance of the farmers; -41- - the efficiency of the extension services. * LLDP project (Malawi), 1969-74: This project, which was intended to increase maize and groundnut production, had disappointing results. - Global production: somewhat unreliable data but substantial increase in maize production and practically no improvement in the case of groundnuts; - Yields: very minimal increase in the case of maize. For groundnuts, yields were lower than those recorded in areas outside the project. - Benor and Harrison (1977). Chambal project (Madhya Pradesh- Tndia). Project to increase yields of rice and wheat in Madhya Pradesh, including investments in irrigation and in a Benor-type extension system. Yields in t/ha Before the project After the project 1974/75 1975/76 Rice 2.0 2.4 2.8 1/ Wheat 1.3 2.0 1/ Average yield of 3.7 t for the 17X of the farmers who followed all the advice given by the extension agents. - Benor and Harrison (1977) and World Bank Report No. 2171 a-IN (1978). Chambal project (Rajasthan-India). Project similar to the preceding one in a different region of India. -42- Average yields per hectare Before the project After the project 1975/76 1976/77 1977/78 Rice (irrigated) 2.00 3.35 3.60 4.36 Sorghum 0.44 0.40 0.66 0.83 Wheat 1.17 2.26 2.11 2.31 Grasses 0.69 0.75 0.94 0.72 Sugarcane 40.90 n.d. 59.40 60.50 Yields as a function of the size of farms and of the percentage of new methods adopted Size in ha Rate of adoption up to 50% Rate of adoption over 50% 0 - 0.99 2.37 3.92 1 - 1.99 2.63 4.05 2 - 3.99 2.97 3.79 4 - 7.99 3.18 4.39 8 and up 2.70 4.16 - World Bank Report No. 1288 a-IN (1977). Benor-type extension project plus irrigation in western Bengal. Yields in t/ha Before the project After the project Wheat 1.2 2.7 Rice 0.8 2.0 - Vicente de Jesus (1979). Masagena 99 is a Philippine project designed to increase rice production in three ways: - A network of extension service technicians; - A credit system; -43- - Radio programs. The project has made it possible to move from a position of dependence (rice imports) to one of surplus over a period of four years. Net cash earnings of farmers have doubled between 1973 and 1977. The extent of increases in yields is not known. - Institute of Development Studies (1968). In this special issue, devoted to the result of the Green Revolution in India, it is pointed out that over a period of 10 years (1963-65 to 1972-74), average yields per hectare increased by 60% for wheat and 13% for rice. (b) General observations on comparisons of yields The serious failing in these studies is that they do not permit the identification of the role of each factor in increased yields. In practice, the latter is the result of a complex process involving: - community investments (irrigation); - individual inputs having a monetary cost (pesticides, fertilizers); - individual inputs having no monetary cost but an opportunity cost exceeding 0 (labor); - new methods of cultivation that do not represent an expense to the farmer (most of these are provided by the extension service); - the introduction of high-yield varieties. Moreover, to the extent that these are time series studies, there is an unknown contingency factor (climatic variation, rainfall). They do not therefore permit the measurement of yields specifically attributable to the extension service. It should be noted here that in evaluation studies of World Bank projects in this field, the rates of return are very high, due in part to the fact that increases in yields are attributed -44- entirely to the project, although the latter has merely reinforced a pre- existing extension system that was not itself without cost. Method of individual comparisons These studies focus on the farmer as the unit of observation. They may be time series, if it has been possible to monitor a sample of farmers before, and then after, the introduction of an extension service. This only rarely happens. They are more usually cross-section, i.e. they are based on samples of farmers, some of whom have been in contact with the extension services and others not. So far as statistical verification is concerned, most of these studies use Cobb-Douglas-type production functions, where the coefficient estimated for each independent variable expresses its output elasticity. (a) Results - Patrick and Kehrberg (1973). Cross-section study of five samples of farmers in Brazil, in which only one of the samples is based on more than 100 observations. All these farmers are located in regions with agricultural extension services, but the number of individual contacts with extention agents varies from one farmer to another. The estimated production functions for the five samples show that the more frequent the contacts, the higher the productivity, but this ratio only assumes significant proportions in one of the five cases. - Harker (1973). This is a cross-section study of a sample of 971 farmers in Japan. The use of agricultural periodicals, radio programs and extension agents significantly increases the productivity of farmers. - Moock (1973). 152 maize growers in Kenya. Very slightly positive but not significant impact of extension services. Hopcraft (1974). 674 farmers in Kenya. Productivity increases -45- significantly with the number of visits up to 4-7, then stagnate s. Demonstrations are very effective but not attendance at farmers' training center courses. - Hong (1975). 895 farmers in the Republic of Korea. Extension activities significantly increased the technical and economic efficiency of farmers, a tendency that becomes more marked in the case of older and better educated farmers. Annual rate of return: 449X. - Halim (1976). Three samples of farmers in the Philippines. The number of contacts with extension agents significtintly increases productivity in two instances, but not in the third. - Pachico and Ashby (1976). 382 farmers i-n Brazil. Productivity falls significantly with the number of contacts. - Hay and Jamison (1977). 403 farmers in ThEiiland. Attendance at extension service classes significantly increases productivity. - Jamison and Lau (1978). Two samples of farmers (184 and 91) in Thailand. The number of visits to the village by saxtension service agents significantly increases the productivity of those who use chemical fertilizers. But this factor is negative (althotigh not significantly so), in the case of farmers who do not use fertilizers. - Academy for Educational Development (.1978). The "Basic Village Education" project in Guatemala has experimented with various extension methods for poor farmers in Guatemala (radio programs alone; radio programs plus instructor; and, in some cases, pvesertice of an agriculturist). The experiment has been highly successful in. ttrms of the number of farmers contacted and the adoption of new farming methods, although increases in productivity have remained small. - Baidya, Chou, Jamison, Moock and Shrestha (1980). 683 farmers in -46- Nepal, growing two crops of rice and sometimes winter wheat. Contacts with extension agent s have a non-significant positive effect for the second rice crop and a very slightly sigrLificant positive effect for winter wheat. - Huffman , W.E. (1981). This very interesting study of 295 farmers in the southern Unite,d States shiows, on the one hand, that the farmers receiving assistance from the extension services have a significantly higher productivity than thk others and, on the other hand, that inequality of access, in terms of boath quarntity and quality, by black farmers and white farmers to the extensio)n services explains in part the difference in earnings between these two groupis. (b) General observations on these studies Summary of results Whereas with the' two preceding methods, the postulated relationship was almost invariably positti.ve, the results of these studies of samples of farmers are more divergent: - 9 studies provide significantly positive correlations, although in some instances the scope of the findings is narrow; - 8 others produce results that are positive but statistically without significanace; - 3 produce a negative result that is not statistically significant and one a significanXtly ne!gative result. The method So far as the empirical verification of the human capital theory is concerned, the economic analysis of the effectiveness of agricultural extension services has followed the siame road as educational economics. - Initially, studies of the rate of return of the Becker type (based on averaged data),, -47- - Followed by macro-economic relationships of the Denison type (based on aggregated data); - Estimates of the age-gains profile corresponding to the comparison of rates of return (again based on averaged data); - Finally, an analysis of the variance of individual gains (based on individual data). In both cases, the first three methods provide results that are very positive, although scientifically questionable, as they oversimplify the factors and processes leading to the observed results. For the most part, this objection does not apply to the studies based on individual data, which are far and away the most important among the recent literature. Whereas the first three methods are characterized by overoptimism as to the real impact of agricultural research and extension services on productivity, the fourth tends to underestimate this impact. The improvement in productivity achieved by farmers who have not been in contact with extension agents but who, little by little, come to imitate their fellows who had previously adopted the innovations suggested by the extension services represents an indirect but nonetheless real effect of extension activity. The studies made in Nepal (see Baidya, Chou, Jamison, Moock, and Shrestha (1980), and Moock, Jamison, Shrestha (1981)) indicate that besides the direct impact of extension work (sometimes significant, sometimes not significant), a significantly positive factor with a higher value reflects the effect of the percentage of farmers in the same panchayat who have recently been in contact with extension agents. It is a kind of bandwagon effect, with multiplier potential, that occurs when adoption of an innovation no longer represents marginal behavior. It then becomes the new norm and is adopted by the majority of farmers, whether or not they have been in contact with extension agents. -48- Conclusion Increasing agricultural productivity is a process which takes time, and which sometimes has negative secondary effects (increased disparities between farmers, a decline in food crop production). The second of these effecs can be fairly easily avoided if the project is well designed, but the first is a more difficult problem which calls for special policies on the part of governments. When agricultural extension projects are viewed in the short term, the results are often unequal, inconsistent and sometimes disappointing. There are many Cassandras among the critics. Although some of their warnings have an undeniable value, they are probably wrong in the long term. A diagnosis of disaster may be more an effective incitement to action than studies dwelling on progress already achieved but it can also have a discouraging effect. There is a whole stock of agricultural innovations awaiting dissemination to increase productivity. The extension services are playing, and can continue to play an effective role in this dissemination. The major obstacles to the adoption of new techniques (conservatism, lack of capital, bottlenecks in supply of inputs and marketing of outputs) are well known by now, and have been largely taken into account in new projects over the past ten years. There are still two questions that are not properly clarified and call for further research, namely, the question of the interaction between the educational level of farmers and their access to extension services, and that of the best vehicles for the various types of information. The question of the best vehicle in terms of cost and effectiveness for each type of information has not been sufficiently studied. Certain types -49- of information now disseminated by extension agents could be disseminated more cheaply by radio. Extension workers--the medium which costs the most per hour of contact--should be used only for information that they alone are capable of transmitting effectively. It is well established that educated farmers are more active in seeking information, and that they make better use of the acquired information. But extension services, by themselves, are not a satisfactory substitute for the whole package of education plus information. Education seems to remain a big point in the long run in the rural development process. But several points, here, need to be clarified. First of all, is there a threshold, in terms of a minimum number of years of education, under which the impact of education is not significant? We assume such a threshold exists, but additional research should be undertaken to assess this assumption. Second, there is the problem of the optimal educational policy, given budgetary constraints. Most poor countries have not yet achieved universal primary education, and some countries will have to choose between expansion of primary education and adult literacy programs. What is the most cost-effective strategy? Here again, more research is needed. Finally, there is the question of educational content and curricula. Many observers argue that these should be more oriented toward the needs of rural communities, more practical, more adapted to rural and agricultural life. But if the main impact of education is to make people more innovation-oriented, more able to have access to information, to select it, and make the best use of it, it is not necessarily these changes in curricula that will achieve these objectives. The assumed superiority of vocational, practical curricula versus general one remains to be demonstrated. -50- ANNEX Table 1; Daily Calorie Supply per Capita in 1964-66 and in 1977 1/ Daily calorie supply per capita 1964-66 1977 Kampuchea, Dem. 2,168 1,926 Lao PDR 2,005 2,082 Ethiopia 2,152 1,754 Mali 2,159 2,117 Nepal 2,218 2,002 Somalia 1,778 2,033 Burundi 2,017 2,254 Chad 2,259 1,762 Mozambique 2,108 1,906 Burma 2,011 2,286 Upper Volta 2,060 1,875 Viet Nam 2,134 1,801 India 1,948 2,021 Malawi 2,257 2,066 Rwanda 1,908 2,264 Sri Lanka 2,219 2,126 Guinea 2,075 1,943 Sierra Leone 2,185 2,150 Zaire 2,036 2,271 Niger 2,211 2,139 Benin 2,230 2,249 Pakistan 1,993 2,281 Tanzania 2,170 2,063 Afghanistan 2,000 2,695 Central African Rep. 2,172 2,242 Madagascar 2,375 2,486 Haiti 1,904 2100 Mauritania 1,981 1,976 Uganda 2,179 2,110 Angola 1,907 2,133 Sudan 2,088 2,184 Togo 2,222 2,069 Kenya 2,253 2,032 Senegal 2,348 2,261 Indonesia 1,798 2,272 Ghana 2,136 1,983 Yemen, PDR 1,895 1,945 1/ Countries are ranked by order of GNP per capita, from the lowest to the highest (developing countries only). Data of 1978. Underlined figures indicate that there was a decline in daily calorie supply per capita between those two period. Sources; WDR 1980. Agricultural Commodity Projections, 1970-1980. Vol. II, FAO, 1971. -51- Table 1 (Cont.) Daily calorie supply per capita 1964-66 1977 Cameroon 2,264 2,0 69 Liberia 2,287 2,404 Honduras 1,930 2,02.5 Zambia 2,237 2,002 Zimbabwe 2,551 2,57&6 Thailand 2,226 1,929 Bolivia 1,731 1,974 Philippines 1,895 2,189 Yemen Arab Rep. 2,089 2,192 Congo, People's Rep. 2,151 2,284 Nigeria 2,168 1,951 El Salvador 1,877 2,051 Morocco 2,091 2,534 Peru 2,255 2,274 Ivory Coast 2,433 2,517 Nicaragua 2,253 2,446 Colombia 2,220 2,364 Paraguay 2,736 2,824 Ecuador 1,846 2,104 Dominican Rep. 2,004 2,094 Guatemala 1,952 2,156 Syrian Arab Rep. 2,440 2,684 Tunisia 2,153 2,674 Jordan 2,430 2,107 Malaysia 2,225 2,610 Jamaica 2,243 2,660 Lebanon 2,401 2,495 Korea, Rep. of 2,421 2,785 Turkey 2,858 2,907 Algeria 1,967 2,372 Mexico 2,623 2,654 Panama 2,317 2,341 Taiwan 2,379 2,805 Chile 2,523 2,656 South Africa 2,745 2.83. Costa Rica 2,234 2,550 Brazil 2,541 2,562 Uruguay 3,039 3,036 Argentina 2,868 3,347 Portugal 2,676 3,076 Yugoslavia 3,174 3,445 Trinidad and Tobago 2,361 2,694 Venezuela 2,392 2,435 Hong Kong 2,324 2,883 Greece 2,972 3,400 Singapore 2,454 3,074 Spain 2,703 3,149 Israel 2,855 3,141 -52- BIBLIOGRAPHY ACADEIAY FOR EDUCATION DEVELOPMENT (1978). The Basic Village Education Project-Guatemala. Final Report, Washington D.C. AKIN(), M. and HAYAMI, Y. (1975). Efficiency in Public Research. Rice Breeding in Japan's Economic Development. American Journal of Agricultural Economics. Vol. 57, p. 1-10. AR/lJI, A.A., SIM, R.J. & GARDNER, R.L. (1978). Returns to Agricultural Research and Extension Programs; An Ex-Ante Approach. American Journal of Agricultural Economics. Dec. p. 964-968. PiYER, H.A. & SCHUH, G.E. (1972). Social Rates of Return and Other Aspects of Agricultural Research; The Case of Cotton Research in Sao-Paulo, Brazil. American Journal of Agricultural Economics. Vol. 54, p. 557-569. BARLETT, P.F. (1980). Agricultural Decision-Making - Anthropology Contributions to Rural Development. Academic Press, N.Y. BARLETTA, A.L. (1967). Costs and Social Returns of Agricultural Research in Mexico. Ph.D. Dissertation. University of Chicago. BAIDYA, B.G., CHOU, E., JAMISON, D.T., MOOCK, P., SHRESTHA, R. (1980). Evaluating the Impact of Communications on Agricultural Development: General Observations and a. Case Study from Nepal. The World Bank, DEDPH. BELLONCLE, G. & GENTIL, D. (1967). Pedagogie de l'iniplantation du mouvement cooperatif au Niger, in Acchives internationales de sociologie de la cooperation no. 23, janvier-juin. BENOR, D. & HARRISON, J.Q. (1.977). Agricultural Extension. The Training and Visit System. The World Bank. Washington, D.C. BERTHELOT, J. (1979). Un exemple de contre-reforme agraire au Togo. Le Monde Diplomatique. Juillet. BERTHELOT, J. & DE RAVIGNAN, F. (1980). Les sillons de la faim. Editions l'Hermatton, Paris. BHALLA, S.A. (1976). Farm Size, Factor Productivity and Technical Change in Indian Agriculture. Appendix A in BERRY, R.A. and W.R. CLINE: Farm Size, Factor Productivity and Technical Change in Developing Countries. Study prepared for the ILO and the World Bank. BOWMAN, M.J. (1976). Rural. People and Rural Economic Development. IIEP Seminar Paper. "The Planning Problems in Rural Education". 13-17 Oct. 1975. BREDAHL, M. & PETERSON, W.L. (1976). The Productivity and Allocation of Research; U.S. Agriculltural Experiment Station. American Journal of Agricultural Economics. Vol. 58, p. 684-692, -53- CANCIAN (1972). Change and Uncertainty in a Peasant Economy. Stanford, Stanford University Press. CHAMBERS, R. & BELSHAW, B. (1973). Managing Rural Development; Lessons and Methods from Eastern Africa. Institute for Development Studies - Discussion Paper no. 15, University of Sussex. CHAMBERS, R. & WICKRE MANAYAKE (1977). Agricultural Extension. Myth, Reality and Challenge. In Farmer (ed.) Green Revolution? Westview Press, Boulder, Colorado. CHINAPPA, B.N. (1977). The Adoption of the New Technology in North Arcott District. In Farmer (ed.). CLINE, P.L. (1975). Sources of Productivity Change in United States Agriculture. Ph.D. Thesis, Oklahoma State University. COUGHENOUR, C.M..(1960). The Functioning of Farmer's Characteristics in Relation to Contact with Media and Practice Adoption. Rural Sociology, XXV - no. 3, 293-4. DENISON, E.F. (1964). La mesure de la contribution de 1'enseignement (et du facteur residuel) a la croissance economique. In OCDE; Le facteur residuel et le progres economique. Paris. DUBEY, SUTTON & GALLUP (1962). Village Level Workers; Their Work and Result Demonstrations. National Institute of Community Development. Government of India - Manager of Publications. Delhi. DUMONT, R. (1962). L'Afrique Noire est mal partie - Le Seuil, Paris. DUMONT, R. (1979). La Guinee-Bissau peut encore echapper au sous-developpement. Le Monde Diplomatique. Juillet. DUMONT, R. (1980). L'Afrique Etranglee. Le Seuil, Paris. DUMONT, R. and PEIRIS, D. (1980). Spokes in the Peasant Revolution. Interview published in the Guardian, Dec. 24, p. 6, London. EICHER, J.C. & ORIVEL, F. (1979). L'Allocation des ressources a l'education dans le monde. Office des statistiques, UNESCO. EVENSON, R.E. & JHA, D. (1973). The Contribution of Agricultural Research System to Agricultural Production in India. Indian Journal of Agricultural Economics. Vol. 28, p. 212-230. EVENSON, R.E. (1976) a. Agricultural Research and Extension in Asia; A Survey Preliminary Draft. Second Asian Agricultural Survey Team. September. EVENSON, R.E. & KISLEV, Y. (1976). b. A Stochastic Model of Applied Research. JPE Vol. 84, no. 2 - April. The University of Chicago Press. EVENSON, R.E. (1978). Research, Invention, Extension and Productivity Change in US Agriculture: An Historical Decomposition Analysis. Paper presented at Symposium on Research and Extension Evaluation. Moscow. Idaho, 21-23 May. -54- FARMER, B.H.., ed. (1977). Green Revolution? Technology and Change in Rice. Growing Area of Tamil Nader and Sri Lanka. Mac Millan, London. FLIEGEL, F. (1956). A Multiple Correlation Analysis to Factors Associated with Adoption of Farm Practices. Rural Sociology, XXI, 284-92. FREIRE, P. (1973). Extension o Communicacion; la Conscientization en Medio Rural. Montevideo - Tierra Naeva. GRILICHES, ZVI (1958). Research Costs and Social Returns; Hybrid Corn and Related Innovations. Journal of Political Economy, 66, p. 419-431. GUTELMAN (1979). L'utilisation des media modernes pour l'enseignement en milieu rural dans les pays en voie de developpement: les problemes d'organisation. Paris, IIEP. HALIM, A. (1976). Schooling and Extension and Income Producing Philippines Household. Bangladesh: Department of Agricultural Extension and Teachers Training, Bangladesh Agricultural University. HARKER, B.R. (1973). The Contribution of Schooling to Agricultural Modernization: An Empirical Analysis in P. FOSTER and R.J. SHEFFIELD (eds.). Education and Rural Development - London; Evans Brothers. HATCH, J. (1976). The Corn Farmers of Motupe; A Study-of Traditional Farming Practices in Northern Coastal Pern. University of Wisconsin - Madison. HAY, J. & JAMISON, D.T. (1977). The Effects of Education on Farmer Productivity in Malaysia. Unpublished Manuscript. Washington, D.C., The World Bank. HONG, K.Y. (1975). An Estimated Economic Contribution of Schooling and Extension in Korean Agriculture. Unpublished Doctoral Dissertation. University of the Philippines at Los Banos. HOPCRAFT, P.M. (1974). Human Resources and Technical Skills in Agricultural Development; An Economic Evaluation of Educative Investments in Kenya's Small Farm Sector. Unpublished Doctoral Dissertation. Stanford University. HUFFMAN, W.E. (1978). Returns to Extension; An Assessment Paper presented at Symposium on Research and Extension Evaluation. Moscow, Idaho, 21-23 May. HUFFMAN, W.E. (1981). Black-White Human Capital Differences; Impact on Agricultural Productivity in the US South. AER. March. HURSCH, NIELS & KERR (1968). Innovation in Eastern Nigeria: Success and Failure of Agricultural Programs in 71 Villages of Eastern Nigeria. Michigan State University. INSTITUTE OF DEVELOPMENT STUDIES (1978). Green Revolution, Counter-Revolution or Evolution? in Development Research Digest, no. 2, Autumn. Complementary Issue on the Green Revolution and Rural Technology. -55- JAMISON, D.T. & LAU, L. (1978). Farmer Education and Farm Efficiency. Washington, D.C., The World Bank, mimeo. JAMISON, T.D. & MC ANANY, E. (1978). Radio for Education and Development. SAGE, N.Y. JOSHI, P.C. (1979). Naissance d'une nouvelle classe d"'indesirables" en Inde. Le Monde Diplomatique, juillet. KABORE, J. (1979). Rapport de mission de consultation sur les constraintes. des services de vulgarisation agricile dans six pays fancophones d'Afrique de l'Ouest. Haute Volta, Ministere du.Developpement Rural. 146 p. LELE, U. (1977). Le developpement rural; l'experience africaine. Economica, Paris. LENGLET, F. (1980). The Ivory Coast - When Benefits from Education Information in Rural Television? In Mc Anany. Communications in the Rural Third World. Praeger, N.Y. LEONARD, D. (1970). Some Hypotheses Concerning the Impact of Kenya Government Agricultural Extension on Farmers. Institute of Development Studies, Staff Paper no. 7, Nairobi, University of Nairobi. LEONARD, D. (1973). Why Do Kenya's Agricultural Extension Services Favor the Rich Farmer? Sixteenth Annual Meeting of the African Studies Association, Syracuse, N.Y., Oct., Nov. LOCKHEED, M.E., JAMISON, D.T. & LAU, L.J. (1980). Farmer Education and Farm Efficiency: A Survey. Economic Development and Cultural Change. Vol. 29, No. 1, Oct. MC ANANY, E. (1979). Agriculture Communication Field Guide. Institute for Communication Research. Stanford University. MC ANANY, E. (1980). Communications in the Rural Third World. The Role of Information in Development. Praeger, N.Y. MALASSIS (1975). Ruralite, Education, Developpement, UNESCO, Paris. MOHAN, R. & EVENSON, R.E. (1975). The Intensive Agricultural Districts Program in India; A New Evaluation. Journal of Development Studies. Vol. 11, p. 135-154. MOOCK, P.R. (1973). Managerial Ability in Small Farm production; An Analysis of Maize Yields in the Vihiga Division of Kenya. Unpublished Doctoral Dissertation. Columbia University. MOOCK, P. (forthcoming). Education and Technical Efficiency in Small Farm Production. Economic Development and Cultural Change. MORIS, J. (1966). Farmer Training as a Strategy for Rural Development. Conference of Kericho. 25 Sept. - 10 Oct. -56- MYREN, D.T. (1966). The Rockefeller Foundation Program in Corn and Wheat in Mexico, in WHARTON, C.R. (ed.). Subsistence Agriculture and Economic Development. Aldine Publishing Company, Chicago. NAIR, K. (1962). Blossoms in the Dust; The Human Factor in Indian Development. Praeger, N.Y. O'SULLIVAN, J. (1980). Guatemala; Marginality and Information in Rural Development in the Western Highlands. In Mc Anany Communication in the Rural Third World. Praeger, N.Y. OXENHAM, J. & CHAMBERS, R. (1978). Organizing Education and Training for Rural Development; Problems and Challenges. Paris IIEP. PACHICO, D.H. & ASHBY, J.A. (1976). Investments in Human Capital and Farm Productivity: Some Evidence from Brazil. Unpublished Manuscript. Cornell University. PATRICK, G.F. & KEHRBERG, E.W. (1973). Costs and Returns of Education in Five Agricultural Areas of Eastern Brazil. American Journal of Agricultural Economics. Vol. 55, p. 145-154. PAYER, C. (1979). La Banque Mondiale et l'integration du paysan au marche. Le Monde Diplomatique. Juillet. PERRATON, H., JAMISON, D.T. & ORIVEL, F. (1981). Mass Media for Agricultural Extension in Malawi. The World Bank, mimeo. PETERSON, W.L. (1971). The Return to Investment in Agriculture in the United States, in W.L. FISCHEL (ed.) Resource Allocation in Agricultural Research. Minneapolis. University of Minnesota Press. RAHIM, S.A. (1966). The Comilla Program in East Pakistan in WHARTON, C.R. (ed.) Subsistence Agriculture and Economic Development. Aldine Publishing Company, Chicago. RAM, Rati (1980). Role of Education in Production; A Slightly New Approach. The QJE, Sept., Vol XCV, no. 2. RAVIGNAN, F. de (1979). Defense et illustration du "modele". Le Monde Diplomatique, juillet. RIBEIRO, J.B. & WHARTON, C.R. (1966). The Acar Program in Minas Gerais, Brazil, in WHARTON, C.R. (ed.) Subsistence Agriculture and Economic Development. Aldine Publishing Company, Chicago. RICE, E.B. (1971). Extension in the Andes; An Evaluation of Official US Assistance to Agricultural Extension Services in Central and South America. AID Evaluation Paper no. 3, Washington D.C. ROGERS, E.M. (1966). Motivations, Values, and Attitudes of Subsistence Farmers: Toward a Subculture of Peasantry in WHARTON, C.R. (ed.) Subsistence Agriculture and Economic Development. Aldine Publishing Company, Chicago. -57- ROGERS, E.M. and others (1970). Diffusion of Innovations in Brazil, Nigeria and India. East Lansing. Michigan State University. Dept. of Communication Report. ROGERS, E.M., BRAUN, J.R., with VERMILION, M.A. (1977). Radio Forums: A Strategy for Rural Development in Spain. JAMISON, AC ANANY (eds.); Radio for Education and Development - Case Studies. Vol. II, World Bank Staff Working Paper no. 266. ROSEN, B.C. (1964). The Achievement Syndrome and Economic Growth in Brazil Social Forces. Vol. 42, no. 3. RUTHERNBERG, H. (1973). Types of Organization in Agriculture Production Development. In Zeitschrift fur Auslandische Landwirtschaft, Jahrgang 12, Heft 3/4. SCHULTZ, T.W. (1964). Transforming Traditional Agriculture. Yale University Press. SCHULTZ, T.W. (1966). New Evidence of Farmers Responses to Economic Opportunities. From the Early Agrarian History of Western Europe, in WHARTON, C.R. (ed.), Subsistence Agriculture and Economic Development. Aldine Publishing Company, Chicago. SCHULTZ, T.W. (1971). The Allocation of Resources to Research in W.L. FISCHEL (ed.) Resource Allocation in Agricultural Research, p. 90-120. Minneapolis, University of Minnesota Press. SHORE, L. (1980). Mass Media for Development: A Reexamination of Access, Exposure, and Impact. In MC ANANY (ed.) Communications in the Rural Third World. Praeger, N.Y. SINGH, I. (1979). Small Farmers and the Landless in South Asia. World Bank Staff Working Paper, no. 320. VICENTE de JESUS (1979). Talk given in Jamaican's Seminar on Agricultural Extension. 1-7 June. WORLD BANK REPORT (1977). Appraisal of the West Bengal Agricultural Extension and Research Project. March 10, no. 1288 a IN. WORLD BANK REPORT (1978). India Staff Project Report. Appraisal of the Composite Agricultural Extension Project. No. 30, no. 217 a IN. Washington D.C. PART II MASS MEDIA, BASIC EDUCATION AND.llAGRICULTURAL EXTENSION Hilary Perraton Table of Contents Page SUMMARY .......................... 61 INTRODUCTION ..................................... 63 II AIMS AND CONSTRAINTS OF BASIC EDUCATION AND AGRICULTURAL EXTENSION ... - ..65 A. Aims .65 B. Constraints on education .68 C. Constraints on agricultural extension X 70 III THE ORTHODOX APPROACH ........................... 73 A. The quality of the evidence .................. 73 B. Literacy: methods, effectiveness, costs ....... 75 C. Agricultural extension: methods, effectiveness and costs ........ ........................... 79 IV USING MASS MEDIA ................................ 83 A. Mass media with group study ..... .............. 83 B. Mass media for individuals with feed back ..... 100 C. Radio ........... ............................. 102 D. Rural newspapers ............... ............... 106 V THE COSTS OF MASS MEDIA ............ ............. 110 A. Mass media with group study ..... .............. 110 B. Mass media for individuals with feedback *... 116 C. Radio and print ....... ........................ 118 D. Cost summary ....... .......................... 120 VI CONCLUSIONS ......... ............................. 124 A. Implications for policy ...... .................. 124 B. Implications for research ..... ................ 126 APPENDIX A - Mass media projects . t ........ 128 REFERENCES ..141 -61- SUMMARY This paper surveys the evidence on using mass media to provide basic or agricultural education and to support agricultural extension in third world countries. A systematic search of the literature makes it possible to draw some conclusions about the methods, effects and costs of such use of mass media. The paper begins by analysing the aims of agricultural extension and basic education. While the principal aim of agricultural extension is to increase agricultural productivity, the aims of basic education are often broader. They include: studying the 3-Rs; helping people to benefit from rapid change; the ability to solve problems which rests on an understanding of the relationship between the individual and society; the mastery of generative concepts; and efficiency in acquiring information. The paper next examines the constraints within which basic education and agricultural extension work and which apply regardless of the methods of communication used. The problems of comparing orthodox with unorthodox methods of teaching are discussed. In the light of that discussion the evidence on the costs and effectiveness of orthodox literacy teaching are examined as an example of a conventional approach to basic education. This suggests that adult literacy teaching has generally been neither particularly successful nor particularly cheap. There is more limited evidence on the effectiveness of agricultural extension and it often has major problems in reaching women farmers and in reaching poorer fatuers. As conventional agricultural extension is labour intensive, budgetary constraints make it difficult to improve agricultural extension by raising the ratio of extension agents to farmers. The main body of the paper, section IV discus8es the use that has been made bf mass media. Mass media have been linked with group study in various ways, notably through the radio schools of Latin America, farm forums, and radio campaigns of Tanzania and Botswana. While group study necessitates a ground organisation, as well as an organisation to produce broadcasts or printed materials, it increases the effectiveness of those materials. In some countries mass media projects have been organised on the scale where, over a period of time, they are reaching an important proportion of rural adults, which we can assume to be a larger proportion than is reached by other agencies. Mass media have also been used for the teaching of individuals with arrangements for organised feedback from them. One of the techniques that has been used here is the employment of correspondence courses. These -62- have proved effective for teaching agriculture and are of particular importance for the in-service training of agricultural extension agents. Mass media, including both radio and rural newspapers, have also been used for education without any formal arrangement for regular feedback. While the evidence does not permit close comparisons of effectiveness between this approach and the others discussed, it does confirm that radio has been effective in communicating information to farmers in a wide range of countries. The costs of mass media are discussed in section V. Radio schools in Latin America, teaching the equivalent of a primary school curriculum, have done so at a cost per,head ~of betweenfU978US$3Q and 65 a_year. _The large radio campaigns in Tanzania had costs of between $1 and $5 per participant for a campaign of two to three months. The cost of radio programmes by themselves, without any arrangement for feedback, is modest. While the costs vary from country to country,evidence from Malawi suggests that the cost of a one hour contact between a radio programme and a farmer is only 1/3000 of the cost of a face-to-face contact with an extension agent. The paper concludes by discussing the implications of its findings for policy and for research. The effectiveness of mass media combined with group study, the potential of training agricultural extension agents by correspondence, and the relative cheapness of radio suggest that these are of considerable importance for basic education and for agriculture. Further research is suggested on alternative ways of linking radio with orthodox agricultural extension, on the effectiveness of extension work generally, and for further studies of the use of mass media in population, health and nutrition. -63- I INTRODUCTION 1.1 This paper examines the role of mass media in supporting basic educa- tion and agricultural extension. It addresses the question, 'how far, and in what way, can mass media be used as a supplement or as an alternative to orthodox methods of basic education for adults or of agricultural extension?'. 1.2 Its origin is two-fold. First, in very broad terms, it takes as a starting point the recognition that orthodox methods of extension, as of basic education, are coming up against the limits of money and manpower. The burden of school budgets means that ministries of education cannot afford to teach adults in the same way, incurring the same unit costs, as they do children. Similarly, in many countries, it is not possible to expand the agricultural extension service and improve the ratio of extension agents to farmers. Existing ratios and existing methods of working are pressing hard against financial barriers. This makes it appropriate for educators and agriculturalists to seek alternatives to their orthodox, face-to-face, methods of communication. 1.3 Second, the study follows the Bank's earlier examination of the educa- tional uses of radio (reported in Radio for education and development:case studies)and of distance teaching for formal education (reported in Alternative routes to formal education). These suggested the potential value of using mass media and distance-teaching methods for out of school education and for extension which are examined in this paper. 1.4 Adult education and agricultural extension have different roots and are often quite separately administered. In the eyes of some practitioners they have different objectives. They are considered together here, however, because of their interdependence. Basic education for rural people will concern itself with agriculture if it is to speak to their condition. Agricultural extension, if it demands more than an unthinking response to the extension agent's remembered instructions, relies on the understanding which is the main aim of basic education. In practice, too, many peasants receive their basic education, about agriculture at least, principally from the agricultural extension agent. And in practice the methods developed by educational agencies in using mass media are relevant to agricultural extension and vice versa. 1.5 This paper therefore begins by examining the aims of basic education and of agricultural extension and the limits within which both have to work. In order to assess the value of both orthodox and alternative methods we need to look at the major problems which arise in evaluating both adult educa- tion and agricultural extension. As a starting point to that, the experience -64- in the third world of the costs and effects of adult literacy programmes using conventional methods, and of orthodox extension, are examined. The paper then discusses the different ways in which mass media have been used to support rural education. It then examines the comparative cost of different approaches and concludes with recommendations both as guidelines for policy and on further research issues. -65- II AIMS AND CONSTR4INTS OF BASIC EDUCATION AND AGRICULTURAL EXTENSON 2.1 This paper concentrates on methodology. But techniques are seldom neutral and so, before attempting to compare orthodox and mass media approaches, we examine the aims of basic education and'of extension. In order to see how realistic these aims may be, they are examined in the light of the constraints on education and on extension. A. Aims 2.2 The aims of agricultural extension are straightforward. They are to communicate information, mainly about better farming practice, in order that farmers can produce more food. Often, of course, an extension service does more than this: it may provide inputsaforlarmers, such as seed, fertiliser or credit, or advice on activities beyond cultivation, such as post-harvest technology or marketing. Its central aim, however, is most often to provide advice about better farming techniques. Agricultural extension services therefore assume that knowledge is available, usually from agricultural research stations, about improved methods of farming. In practice, although not in theory, extension services have often worked on the assumption that one-way transmission, of messages to thi farmer, is all that matters. 2.3 The effectiveness of extension has been limited in various ways, which we consider in more detail below. One of the important limits on what an extension service can do is the background knowledge and educational level of the farmers to whom it is addressed. To take a trite example, if farmers are illiterate, an extension service can make little use of print to provide information to them. In developing countries, therefore, there is an over- lap between the aims of extension and the aims of basic education. 2.4 The aims of basic education are more complex. It is, in fact, easy to make a case against any attempt to educate adults. Many adult education programmes have failed to attract students, to retain their interest or to teach them anything. There is scant evidence that adult education programmes have any great effect on students' lives. In contrast there is, in most third world countries, a strongly expressed demand for primary and secondary schools. And there is evidence of the long term effects of primary schooling on the attitudes and even the productivity of those who have been to school. Against that demand, and that kind of evidence, it is necessary to weigh carefully the case for devoting resources to the education of adults. For, unless education's share of the natio6al budget is to grow, then devoting resources to adults means diverting them from children. -66- 2.5 Basic education for adults has, in fact, attracted support partly for ideological reasons and partly for practical ones, where it was seen as a pre-requisite for development. The ideological reasons have been varied. Some have been religious in origin: the Roman Catholic church, for example, in both Latin America and Africa has played a leading role in some programmes of adult education. Often, too, equity has been cited as an argument for basic education to redress some of the imbalance between those who went to school and those who did not. Such education has also been seen either as a necessary preliminary to revolution against oppressive regimes, enabling poor peasants to analyse the relationship between themselves and their oppressors, or as a proper consequence to it. 2.6 For practical reasons, new regimes, and newly independent countries have often seen the education of adults as a necessity for the changes they seek within their own societies. Nyerere, for example, soon after Tanzania's independence, argued for adult education because Tanzania's development needed educated adults and could not afford to wait till the children then at school had grown up. Such arguments are still potent: in Ethiopia in the late 1970s, 'after the revolution, the government was deter- mined to make education relevant to the needs of the rural poor who made up the mass of the people' and launched a programme of basic education for adults. (Unesco/UNICEF 1978, p.68) 2.7 The arguments for educating adults converge, whether they start with ideology or with the intention of developing society. Nyerere speaks of liberating man as well as of developing Tanzania. INADES-formation, a church-based educational agency in West Africa has, as its general aim, 'to work for the social and economic advancement of people, laying particular importance on their free and responsible participation in transforming their societies' (INADES, n.d., p.2) Radio schools in Latin America see their role as to help towards the transformation of society as well as the educa- tion of individual citizens or families. 2.8 Moving on from these broad arguments, it is possible to distinguish five qualities which mark basic education. 2.9 First, it includes teaching about the 3-Rs. These are tools, necessary for people to cope with the modern sector, and becoming necessary to cope with changes to the traditional sector. The balance between the 3-Rs will vary and literacy is not always the most important. In Ethiopia again: 'people who had formerly been nomadic gave priority to numeracy over literacy since the counting of livestock was an essential skill which they needed to acquire'. (Unesco/UNICEF 1978, p.77). But literacy and numeracy alone are not adequate components for a programme of basic education. -67- 2.10 For, second, if basic education is to be of value to individuals or to societies, it needs to foster the ability to benefit from rapid change. Mary-Jean Bowman argues that 'the essence of development man or woman is a readiness to adapt to change and to grasp or to create new opportunities.' She goes on to suggest this requires various capabilities including adapta- bility and, more specifically, 'ingenuity in the creation of new opportunities or new ways of doing things in response to new opportunities'. (Bowman 1980, p.8) This ingenuity can be exercised either in transforming rural society or in moving out of it. Thus education needs - to quote an analysis from the Dominican Republic - 'a content and methodology which permits individuals to adapt themselves to a rapidly changing society and to prepare themselves for creative participation in the process of change'. (White 1976a, p.4) 2.11 Such ingenuity or adaptability depends, third, on 'balanced personal growth (intellectual, moral, emotional, social, physical, aesthetic, and psychomotor functions) ... the ability to solve problems, and personal creativity (critical consciousness)' (ibid.) More specifically, it demands the ability to articulate relationships between self and society and between nature and culture. That ability is central to Freire's method of literacy teaching with its concern to develop the student's critical consciousness of himself and his society. Freire found that his students rapidly grasped the distinction between nature and culture, explaining which things in a picture belonged to which category and why. The ability to articulate this is necessary if adult learners are to move on and consider ways in which they can change their society or their family, social and economic life. The articulation involves extending the learner's awareness of a distinction from a limited to a more general context, necessary, as Bruner et.al (1966) pointed out, 'because so much of the meaning of what is being learned is intrinsic in the context in which the learning occurs, there is very little need for verbal formulation'. 2.12 That ability to generalise leads us to our fourth quality which concerns the mastery of generative concepts - those which enable us to recall or order information. Once we have learned some higher level concepts we can generate the detail that we would otherwise forget. Once we know the prin- ciple of transpiration, for example, we can recreate all the detail we might otherwise forget about how to keep plants alive. The choice of such concepts will vary from place to place and will relate to the curriculum of basic education. 2.13 Finally, basic education needs to develop what Bowman calls 'efficiency in acquiring information'. The ability to locate and benefit from varied sources of information is crucial if basic education is to be more than the -68- equivalent of a board school, with its stress on rote learning of a received code of knowledge. 2.14 To sum up, therefore, the 8ims of agricultural extension are fairly simple and clear cut. Their achievement may depend upon the level of education of participating farmers. And that is the concern of basic educa- tion whose aims are broader and less easy to define. B. Constraints on education 2.15 We have only limited evidence on the effects of education on adults, although there is enough for us to sketch out the constraints within which such education must work. There is a much less clear relation between literacy and productivity, for example, than between schooling and producti- vity. And, in an intriguing piece of research, Scribner and Cole found that literacy in the Vai script, used in Liberia but learned only out of school, 'did not substitute for schooling. Vai literates were not significantly different from non literates on ... cognitive measures (and) ... there was no across-the-board evidence of enhanced performance associated with this traditional, non-schooled literacy' (Scribner and Cole 1978, pp 27-28). We cannot simply equate the education of adults with primary schooling, about whose effectiveness we know rather more. Attitudes, for example are, it seems, permanently changed by even a short period at school. (Cole et.al, 1971). Adults who have been to primary school achieve higher scores on the 'overall index of modernity' developed by Inkeles and Smith (1974). And they tend to produce more as peasant farmers (Lockheed, Jamison and Lau, 1978). This contrast may be explained by the suggestion that the effects of schooling are the result more of the hidden curriculum - the values of order, time- tabling and an imposed structure to one's work - than of the overt curriculum (cf Inkeles and Smith, 1974, p.133). 2.16 How far, then, can we expect the education of adults to affect their lives?' The adults'own interests, and the amount of time they have available, set limits to the effectiveness of any educational programme. "In many villages of the third world there is no postal service, no banking facilities, no access to newsprint, no all-weather road, no outpost of the government's extension services ... At international conferences, there is often discussion of the value of literacy for understanding written instructions on the use and application of fertilisers: but there is very little mention of the fact that the written word is nowhere in public in many of these villages where people are meant to be diligently reading fertiliser instructions. One sees no village name; no public announcement board; no shop names, except a plot number." (King, 1979, pp.38-9). It is not surprising that many adults choose not to spend their time learning to read and write in a literacy campaign. There is little for them to read -69- and reading is not important to them. Again, in many societies, women are particularly short of time and are therefore unable or unwilling to spend their precious time on educational programmes. To have any chance of success education needs to be geared to the interests of adults and the questions of importance to them and it must be available at a time and in a way which does not put unreasonable burdens upon them. 2.17 And that means there is a constraint on the curriculum. We cannot just start with the 3-Rs and assume that they will excite and attract adults. There have been two different approaches here. Some early literacy programmes have assumed that their aim is simply to teach a narrow technique. Similarly much agricultural extension and some health education programmes have assumed that their function is simply to instruct peasants in a narrow range of tech- niques which they must follow but not question. Even if they master all the techniques, there is no curriculum which puts together what they have learnt into an understanding of health or agriculture which is greater than the sum of its parts. More recently, as at Radio Santa Maria in the Dominican Republic, for example, educators have begun to talk more broadly about course content and argued that this needs to be 'adapted to the psychology and life-context of the rural, low-status adults' (White 1976a). If we are to follow this line of thought then we are likely to have 'an emphasis on human rights and awareness of socioeconomic inequalities' and include 'educa- tional programmes in agriculture, nutrition and health, home-making, and childcare'. (ibid.) Few rural educators who want to go beyond instructions in a handful of techniques would quarrel with that. But it is very different from what happens at primary schools. With a captive audience of small children one can start with the traditional teaching of the 3-Rs of which schools have decades, if not centuries, of experience. (It is only in well developed schools with highly trained teachers that we begin to get a curriculum more closely based on the local environment.) (cf Beeby 1966). And there is little experience anywhere in the world in building a curriculum which starts from those interests in health or agriculture and moves on to the abstractions which capitve primary school children, with lots of time, take in their stride. Curriculum building for the education of adults is more difficult. 2.18 Perhaps even more important than these practical and intellectual con- straints are the political and economic ones. There are sharp limits to what education can do in isolation. Learning about fertilisers is of little value for someone who cannot buy them. Learning improved techniques of cultivation, which depend upon land reforms, is of little value if the power -70- structure makes any land reform improbable. The effects of educating adults need to be assessed within the political framework of the country concerned. For education is seldom the prime mover. Where education is working against the grain of its host society, we cannot expect it to cut sharply through the major problems faced by peasants who have never been to school. 2.19 So much for the constraints on adult learners. There are two further sets which limit the work of teachers and organisers. The first concerns finance. The cost of ordinary schools, and the continual pressure to expand them, means that, usually, there is little money left for the education of adults. Lavish and expensive solutions to their educational problems are unrealistic. Second, there are few staff available to teach adults. In a very small number of countries there is a cadre of adult educators. Elsewhere adult education is likely to be staffed by primary school teachers in their spare time, by community development workers, and by agricultural extension agents or other extension workers. These groups have one thing in common: for the most part they have themselves received a relatively limited educa- tion. A common feature of the literature, on agricultural extension and that on primary education is the limit which the inadequate educational background of teachers or agents sets to their own work. 2.20 Despite these constraints, there is some evidence that educational programmes outside school can have an effect on their participants. Adults do, for example, learn from some agricultural broadcasts. In Latin America, adults who have studied by radio and correspondence do make changes in their everyday life, as well as learning to read, write and pass examinations. More broadly, exposure to the co-operative movement in the Comilla project, in what is now Bangladesh, had a greater effect on 'overall modernity' than attendance at school. (ibid. p 20) The point of quoting isolated examples here is to demonstrate that the education of adults may be justified by works as well as by faith. We will look in more detail at the evidence from a variety of educational projects in part IV, below. C. Constraints on agricultural extension 2.21 Agricultural extension, too, works within definable limits. It has a long history, going back at least as far as the 13th century (Brown 1971, p.2), and a simple theoretical basis: agricultural research workers develop better farming techniques and the extension service exists to make that information available to farmers. One might therefore expect a well documented and high correlation between extension activity and productivity. Agricultural extension services in the third world have, indeed, burgeoned, encouraged by that assumption, by the belief that extension has had a history of success -71- in metropolitan countries, and by ministries of agriculture wanting to raise productivity. 2.22 There are two sets of reasons why the relation between extension and productivity is, in fact, much less straightforward. First, extension is only one among a number of determinants of agricultural production. Some of these are set out in Figure 1. If an extension service can affect only the availability of information to farmers, its effects may be swamped by those of other factors. Fig. 1 What determines agricultural change Main political social an culture soil and features of framework economic climate society rstrctur e decide: |ntfaing practice] Facilities market availability perceived availability available to structur of inputs risks, and accepta- farmers e.g. credit, costs and bility of seed, benefits information ----T fertiliserI . I also affect: acceptance of changes in farming 2.23 Second, extension services have often themselves not been particularly efficient. They have been criticised on both broad and narrow grounds. The broadest-based criticisms have shown that much agricultural extension is socially regressive, concentrating its benefits on richer rather than poorer farmers, and tending to widen the gap between the rich and the poor. Almost universally it has concentrated on the needs of men, even in societies where women do most of the agriculture. From a narrower perspective five major difficulties have been identified as confronting agricultural extension: - Extension agents have often been too young and under-trained so that they could persuade neither through the authority of age nor through superior knowledge. - Often they have lacked an appropriate organisational structure so that they have been inadequately supervised, been given ill- considered tasks and have had to waste their time on bureaucratic trivia. -72- They have been too thin on the ground to have any significant impact on many farmers. In northern Nigeria, for example, one study showed that only 14% of farmers had any contact with ex- tension agents (Giltrow and Potts 1978, p.D2) Too often agents have had little to teach that would benefit farmers. Often there have been neither simple, (Norman 1976, p.174) nor even complex, (Mellor 1973, p.95) innovations they could offer, despite the hope that there would be in the future. There have been inadequate channels for farmers or extension agents to communicate with research workers: 'We witnessed what can happen when research experts and farmers get together - generally a rare event - on a visit to the lush rice-producing District of Tanjore in the South Indian state of Tamil Nadu. In a series of roadside seminars, farmers beseiged with sharp technical questions a group of experts, including India's leading rice research expert ... In many instances, the experts had no ready answers but, having seen for themselves what the farmers needed, they promised to try to get answers' (Coombs with Ahmed 1974, p.119) -73- III THE ORTHODOX APPROACH 3.1 We can use the evidence of orthodox basic education and extension as a benchmark from which to survey the use of unorthodox approaches which rely on mass media. The best documented work in adult education is on literacy. 3.2 There are, however, three difficulties in making such a comparative analysis. The first is the lack of control groups. While we can find groups who do, or do not, participate in literacy programmes or have con- tact with extension agents, the groups are seldom comparable in other ways. We are therefore forced back to more tenuous evidence, or to multivariate analysis, which leaves unclear the direction of causality. The second difficulty is of particular imnportance for agriculture: we lack a typology for extension. Extension work varies, from place to place in many ways: in the ratio of farmers to agents, the variety of crops, the relative scarcity of land or of labour, extension's own administrative structure and so on. But the literature on extension and its effectiveness often assumes a homo- geneity which neglects these differences. The third difficulty is posed by the quality of the evidence itself. A. The quality of the evidence 3.3 To assess the value of orthodox approaches to education or extension we need some agreed measures of their effectiveness. These are not easy for either sector. There is no simple agreement, for example of what we mean by literacy. Unesco's Experimental World Literacy Programme seems to have defined literacy as the level achieved by school children after four to five years at primary school, while the Tanzanian literacy programme of the 1970s regarded the second year primary level as the criterion. (Unesco 1976, p.185). At another extreme, in Yugoslavia functional literacy has been defined as 'that educational level comparable to the eight years of schooling required of children in the Yugoslavian elementary system' (Savicevic 1973), while those responsible for the British national literacy campaign and its evalu- ation in the late 1970s have been sedulous in avoiding a definition of literacy, or success in literacy. In the Tanzanian national campaign, four levels of achievement are recognised: I Enrol and attend 2/3 of the sessions. in one year; II Recognise words and symbols; write simple short sentences, add and subtract one-figure numbers (semi-literate); III Read a simple text with understanding, write a simple text, multiply two-figure numbers and divide one-figure numbers (literate); IV Continuously use literacy skills, can read and write letters, read a newspaper, keep accounts (functionally literate). (Unesco/UNDP 1976) -74- Thus, even where we have figures for literates, we may not be able to com- pare like with like. 3.4 Much literature on agricultural extension has used the acceptance or adoption of innovations as a measure of effectiveness, without going on to investigate the relationship between such adoption and agricultural pro- ductivity. 3.5 Another body of literature has looked at the correlations between agricultural research or extension and productivity on country, region, or world wide bases. Most such studies have been based on rich countries rather than poor and, as Orivel says: On the methodological plane, there is an element of uncertainty in causation of the same kind as is to be found in studies relating education costs to economic growth.... Is it the efforts made in support of agricultural research and extension activities that result in a highly productive agriculture, or is it the existence of a productive, rich and strong agricultural base that makes it possible to provide substantial resources for research and extension activities? In practice, the relationship between these two phenomena is very likely to be reciprocal, although it is impossible to say which comes first in the final analysis. It may well be that American farming draws its vitality and dynamic characteristics, in the first place from a series of factors .... ... There is nothing to indicate that a similar effort, undertaken in different contexts, in Africa or Asia, will produce the same results. (Orivel 1981, pp.34-5) In contrast, as we shall see, there is a shortage of data on productivity where the unit examined is the individual farm. 3.6 The evidence about costs, like that about effects, is limited and often unsatisfactory. Where costs of teaching adults, or making them literate, are quoted in literature, it is often unclear what is included or excluded. In Somalia, for example, the costs quoted for the literacy campaign are taken from the Ministry of Education's budget and look as if they exclude the major staffing cost, the salaries of teachers taken out of school to teach adults (UNICEF 1978, p.35). As a result we sometimes sim- ply do not know which figures to believe. Ansari and Rege (1964), for example, suggested that in India in 1947-51 it cost 12 rupees to make an adult literate while by 1965/66 Roy et.al (1969) thought it cost 120 rupees. While inflation would account for a change by a factor of two, it would not account for this change by a factor of ten. The evidence on the costs of agricultural extension is no better for here there are remarkably few studies which have looked at the costs of exte-l;sXon and fewer -e;11 which have relied on figures for actual expendlturL .s omposed to buagiL *L&.ac.. -75- 3.7 Finally, there are a number of inescapable methodological problems in evaluating extension or basic education. Where objectives are long-term, one cannot rely on short-ternm indicators. As far as basic education is con- cerned, its objectives are seldom stated in terms that lend themselves to measurement; we usually lack the dubious but convenient yardstick of examina- tion results which are so useful for educational evaluation at higher levels. Furthermore, the very word evaluation implies that we can compare like with like, yet different methods inevitably reach different audiences and make rigorous comparisons impossible. Nevertheless, in the country of the blind the one-eyed man is king, and it is better to base policy on what sketchy evidence can suggest, than to neglect that evidence totally because of its sketchiness. B. Literacy: methods, effectiveness, costs 3.8 The one area of basic education on which there is a reasonable amount of evidence is adult literacy. Four approaches to this can be distinguished: mass campaigns; the naive postwar campaigns of the 1940s and 50s; Paulo Freire's methods; functional literacy as exemplified by Unesco's Experimental World Literacy Project (EWLP). 3.9 The first mass campaigns were developed in the Soviet Union. Between the revolution and the second world war, the USSR apparently achieved near universal literacy, at least in Russia and probably in the Ukraine. The programme involved the mobilisation of large numbers of part-time tutors, the production of primers, supplementary readers and newspapers for new literates in Russian and a range of other languages, and encouragement, through many of the agencies of a totalitarian state, to participants. More recently, Cuba, Tanzania, Somalia and Vietnam have used a similar approach. In the case of both Cuba and Somalia, schools were closed for the duration of the campaign and both teachers and older school students sent to the countryside as literacy tutors. All these campaigns have in common very strong national support: and encouragement to students, and enrolments running into hundreds of thousands or millions. There are parallels between them and the work of Mobral in Brazil: while this is not a state agency, it works on a similar scale and in a country where there are comparably strong pressures to become Literate. The illiterate minority in Brazil are left in a cultural ghetto, unable to vote, to get a bank loan, to join the army or to have any chance of a good job in the modern sector. 3.10 After the second world war both Unesco and the British government (for its dependent territories) felt impelled to launch literacy campaigns, but did so with little success. The Soviet achievement - with its stress on supporting materials and on the context within which new literates were -76- studying - seems to have been overlooked in the planning. In 1964 Unesco's World Literacy Project was abandoned, as in the late 1950s it had been found that only 10-20% were willing to enrol on such campaigns, with subse- quent wastage rates up to 50% and a tendency to lapse into illiteracy. (Blaug 1966, p.408). The nearest to an exception in this sad catelogue came in Ghana where the Laubach campaign was the most successful among British dependen- cies with 129 000 literacy certificates awarded between 1952 and 1957. (Foster 1971, p.590). The period coincides with Nkrumah's building up of the Convention People's Party. 3.11 The main ideological basis for these approaches was a concern for equity - to allow those who had not been to school the opportunities to gain literacy which were open to those who had. In the case of the Soviet Union that con- cern was associated with the demand for a more educated and more adaptable work force. Paolo Freire's ideological arguments are stronger: that in a society where the powerful have had formal education and the powerless have not, the achievement of literacy is a precondition for increasing the latter's control over their own lives. His method depends upon a culture by culture, even village by village, analysis of learners' interests 'Being able to read successfully three traditional literacy primers will not alter ... a neo-literate's place within that structure' of the peasant's or worker's world.' (King 1979, p.38). Almost by definition, this approach does not lend itself to large-scale literacy programmes, at least in non-industrialised countries. There is little hard data available on the effects of the Freirean approach - although there is plenty on its attractiveness to learners and offensiveness to right-wing governments - and less on its costs. 3.12 In the late 1960s Unesco launched a new literacy programme - the Experimental World Literacy Programme - designed to learn from its earlier failures. A set of carefully planned and monitored experiments were run in a number of countries, with the apparent intention that Unesco or member states would launch into large-scale literacy work again once these had yielded their lessons. The new programme assumed adults wanted not just literacy but functional literacy. The phrase had been used in two senses. First, it implied that literacy primers should be about issues of functional importance to learners (health, the growing of local crops, money). Second, it (later?) came to be used to imply a level of literacy so that, for example, literacy students in Tanzania are classified as literate if they achieve one level in the national tests and functionally literate if they have achieved a higher level. (Kuhanga, 1976/77, p.34). The latter meaning has crept into the dictionary so that Urdang defines 'functional illiterate' as 'a person -77- whose literacy is insufficient for most work and normal daily situations' (Collins English Dictionary 1979). The Unesco programme is remarkable both because it seems to have cost more, to teach less, than its predecessors and because its spectacular failure has been noticed, monitored and published. That sort of honesty deserves applause. 3.13 The fullest data on the costs of literacy is available from EWLP, from Mobral in Brazil, and for the national campaigns in Somalia and Tanzania. Table 1 sets out the basic facts. 3.14 Leaving aside the Somalia figures, which, as we saw, may well omit staff costs, there are two clear conclusions. First, given the necessary national commitment, it is possible to run a literacy campaign large enough to achieve great economies of scale. In Tanzania, for example, a national and party commitment to the literacy campaign ensured both en- couragement to the students, and an organisation to train tutors, distribute materials and even hold national examinations for students numbered in their millions. In the Soviet Union in the 1930s the 8-10% annual growth rate, with the consequent demand for literate workers, coupled with socio- political pressure, provided similar motivations, and a similar political drive to establish an infra-structure for literacy teaching. Given such support, the evidence from Brazil and Tanzania suggests that the cost of making someone literate may be as low as $20-40. The cost per learner in Tanzania in 1975 was estimated at 17s/30 or $2.42. Assuming that a student takes 2 years to become literate we get a cost per student of about $4.80. 15% of those who took the national tests in 1975 reached level IV so, if we make the conservative assumption that none of the rest went on to achieve that level or gained anything from their lower performance, we can estimate the 1975$ cost at 100/15 x $4.8 or $32, or 1978$39. If this figure for success istoo modest by a factor of two, we would get a cost between 1978$20 and $40. Mobral showed a cost per final participant of $15 which suggests that their costs per new literate were of the same order. All these figures omit the opportunity costs of participants' time and may well omit some capital costs and the cost of volunteer tutors. 3.15 But, of course, there are relatively few occasions on which a government, or any other national agency, wishes to launch a campaign on that scale. And, second, the evidence from small scale literacy programmes is dramatically different. In 1966, in the Light of Unesco's earlier experience, it was forecast that EWLP would cost: '$39-$35-$40'. The figures achieved are sadly different and suggest that costs per successful literate, in 1978$$, may be as high as $525, with costs per final participant of between $120 and 350, while dropout rates are likely to be high. TABLE 1: Costs and results of literacy projects Project Intended Actual Completion Cost per Cost per final Cost per audience audience rate enrolment participant examined literate Current Constant Current Constant Current Constant Cost $ 1978 $ Cost $ 1978 $ Cost $ 1978 $ Cost Cost Cost EWLP PROJECTS a/ Algeria 55 000 54 000 70% cycle 1 71 112 199 156 n/a n/a 1967-74 49% cycle 2 Ecuador 33 000 17 410 50% 70 111 123 194 300 474 1967-72 Ethiopia 73 000 2 000 12.5% 54 85 92 145 212 335 1968-73 Iran 80 000 66 000 40-50% 49 77 76 120 332 525 1967-72 Zambia b/ 10 928 60% (phase 169 267 n/a n/a 1969-72 1-2) Sudan c/ 1970-71 Textile workers 598 n/a 72 117 221 360 n/a n/a Oil and soap workers 553 35% 66 108 192 313 n/a n/a OTHERS Brazil - Mobral d/ 4.3 m 52% 4.83 7.63 9.50 15 n/a n/a (1972 figures) Somalia 1973 e/ 1.3 m 73% 0.66 0.95 0.9 1.3 1.1 1.6 Tanzania 5.2 m 77% 4.54 5.58 5.90 7.25 16 20 (1975 figures) f/ a. The figures for Algeria, Ecuador, Ethiopia and Iran are from Unesco (1976); a 1972 deflator has been used to convert to 1978 $$. The figures include capital costs but it is not known how these have been calculated. b. We have taken K141 as $169. c, Literacy work (1974), pp 103-5 d. Unesco Regional Office (1975) p.27 e. UNICEF (1978) pp 34-5 f. Kuhanda (1976-77) p.34 -79- C. Agricultural extension: methods, effectiveness, costs 3.16 As we saw, agricultural extension services vary widely, and along a variety of different dimensions. As a step towards a typology it is useful to distinguish four models: the master farmer system; the traditional or take it or leave it approach; a structured system; and contract farming. 3.17 (a) Under a master farmer system, extension agents concentrated on a small, selected number of master farmers and trained them, working on the assumption that their influence would introduce new pratices to their neighbours. It did not work and in practice meant that more resources went to the rich farmers than the poor, without much general raising of the standard of agriculture. Probably no-one liked it: 'the Service began a vigorous campaign of designating good farmers and citizens in communities as Master Farmers. Often these had been effective leaders in their groups and people never thought of themselves as exceptional, except that they were good farmers and always helpful to others. When they were publicly acclaimed as Master Farmers, it sort of set them apart. They were often not much good as local leaders any more.' (Loomis and Beegle 1957, p. 134). (b) Under the traditional approach 'peasants are brought innovations and information which they are free to accept or reject' (Lele 1975, p.64). Critics of this approach suggest that extension agents are themselves too free to visit or stay away from farmers. (c) As a result, various attempts have been made to impose a more rigidly structured systema of visits and of reporting on extension agents. The PIM system in Kenya (cf. Chambers 1974) and the Training and Visit system developed by the World Bank are examples. (cf Benor and Harrison 1977). (d) Even tighter supervision is exercised under contract farming. The 79 000 farmers of the Kenya Tea Development Authority, for example, are selected, trained, graded and controlled by the Authority who provide seedlings and buy the green leaf. (cf. Phillips and Collinson 1976, pp.342-3) 3.18 As we are concerned with the effects of extension, rather than of rural development projects more broadly, it is necessary to seek evidence on these four approaches from studies which focus on the farmer as the unit of observation. For, as Orivel has pointed out, more general studies com- paring yields in development projects do not 'permit the measurement of yields specifically attributable to the extension service'. (Orivel, 1981 p.40). Table 2 sets out the findings of such studies in third world countries. -80- Table 2: Studies of effectiveness of agricultural extension Country Date of Type of Results study extension service Botswana a/ 1967/8 Master Benefit/cost ratios of: farmer 3.33:1 at 8% discount rate in Barolong 1.66:1 at 15% rate area. 1.33:1 at 8% discount rate in SE generally. 0.67:1 at 15% discount rate. Contact mainly with wealthier farmers. Correlation of +.194 output and exposure to extension. Brazil b/ 1973 Traditional Extension visits positively related to productivity in 5 samples but significantly so only in one. Brazil b/ 1976 Traditional Productivity falls significantly with number of contacts. India (Punjab and Haryana) c/ c 1972 Traditional Research could not find yard- stick by which to measure extension effort India d/ 1971 Agricultural Negative correlation between extension the length of time an extension generally service has existed and output Kenya e/ 1970-72 PIM Agents spent 57-66% of time (structured) visiting farmers, contacting 20.3 per month. More visits paid to wealthier farmers. Kenya (Tea c 1973 Contract Three-fold increase in produc- Development farming tion; agents visiting each Authority) f/ farmer 6 times annually. Kenya b/ 1973 Traditional Positive but not significant im- pact of extension services. Kenya b/ 1974 Traditional Productivity increases signifi- cantly with number of visits up to 4-7, then stagnates. Malawi f/ 1972-73 Agricultural Extension visits positively and extension in significantly associated with intensive higher yields of maize but not project of groundnuts. Malawi / 1980 Traditional Extension visits positively but not significantly associated with production of both maize and groundnuts. -81- Nepal b/ 1980 Traditional Non-significant positive effect for 2nd rice crop and very slightly significant positive effect for winter wheat. Philippines b/ 1963/73 Traditional Number of contacts with farmers significantly increases pro- ductivity in two instance but not in third South Korea b/ 1975 Traditional Extension activities signifi- cantly increases technical and economic efficiency of farmers. Thailand b/ 1977 Traditional Attendance at extension classes significantly increases productivity. Thailand hI 1978 Traditional Significant positive correlation between visits and productivity for farmers using fertilisers; negative but not significant correlation for farmers not using fertilisers. Notes: a. Lever 1970 b. Orivel 1981 C. Chaudri 1979 d. Brown 1971 e. Leonard 1977 f. Lele 1975 and Phillips and Collinson 1976 g. Perraton, Jamison and Orivel 1981 h. Lockheed, Jamison and Lau 1978 -82- 3.19 There are two striking omissions from table 2. First, it omits evidence from the Training and Visit system, although this appears to have had a striking impact on yields. It does so because the available studies, on Madhya Pradesh, Rajasthan and West Bengal, for example, do not permit a judgment to what extent these yield increases and tech- nological improvements were due to extension alone and how much to other inputs or investments. In each of these cases, however, inputs' were available at more or less the same level before the introduction of the reformed extension service but were not widely used. Nor did the availability of irrigation increase substantially. The only major change was in the extnesion service itself, strongly suggesting that extension had a catlytic effect on boosting yields. (Benor and Harrison 1977, p.45) 3.20 Second, with few exceptions, the studies have little data on the costs of extension. 3.21 Four conclusions emerge. They should be seen in the light of the quantity and quality of the information available. In view of the scale of agricultural extension, and the significance of extension service to many third world countries and in the budgets of many ministries of agriculture, there is remarkably little hard evidence on its costs and on its effects. 3.22 First, agricultural extension can be successful in carrying information to farmers and in raising their productivity. Evidence of its success is, however, seldom dramatic and it is by no means clear that it will always have a favourable benefit:cost ratio. 3.23 Second, the Training and Visit system has had successes in a number of cases, but it may not be of universal relevance. Where it has achieved successes so far, there have been relatively simple, and cheap, changes which farmers could make and which would yield clear benefits in a short time. Many desired agricultural changes are more complex, or have benefits available over a longer time scale. 3.24 Third, the tendency of extension to contact mainly the richer farmers is confirmed in a number of studies. 3.25 Fourth, the scale and costs of much agricultural extension is such that it could not be scaled up so that all farmers got the attention which is at present lavished on a fortunate minority. As Lele pointed out, in relation to the intensive Kenya Tea Development Authority and Malawi Lilongwe Land Development Project, 'without a very major commitment of resourcea by national governments to rural development and, equally important, to suit- able training of the manpower necessary for the expansion of these services, intensity of most of these programs cannot be replicated on a large-enough scale to reach a mass of the rural population in the foreseeable future.' (Lele 1975, p.69) -83- IV USING MASS MEDIA 4.1 We now turn to the experience of the third world in using mass media as a support, and as an alternative, to orthodox basic education and exten- sion. 4.2 The evidence on mass media is patchy. It covers a variety of projects, ranging in their intentions from teaching narrow skills, as in much agricul- tural extension, to a basic curriculum, as in some of the radio schools of Latin America. Where results seem applicable to this study, we examine projects whose aims went beyond basic education as we have discussed it. A further difficulty in interpreting the literature arises from ways in which mass media have been used in different countries and different continents. Some patterns of use, for example, have developed only in a single continent which forces us into extrapolating from the experience of one continent to another, and making intellectually uncomfortable comparisons between continents. 4.3 Findings are also limited by the measures which researchers have used in considering the effects of mass media. In ascending order of sophistica- tion, researchers have used measures of four kinds: - effort: measuring output in such terms as the number of programmes broadcast or the number of publications printed; - reach: measuring the size of the audience reached and, sometimes, its nature (e.g. age, sex, socio-economic status); - knowledge gain: measuring whether participants in a programme have learned from it; - effect: measuring the direct effects of a programme (e.g. by crop sampling to see whether participation in an extension programme raises agricultural productivity). 4.4 In this chapter we look in order at the use of mass media when linked when group study, at its use with feedback for individuals, at radio, and at rural newspapers. A. Mass media with group study 4.5 Learning groups, usually supported by radio, have been organised in a number of different ways. The earliest pattern of organisation came from the radio farm forums of Canada. Groups of farmers met together once a week over a number of months or years to listen to the radio programme, discuss it, and then go on to act on its advice. Later farm forums were established in India and Ghana and in other countries of Africa and Asia, with mixed success. A different kind of organisation was developed in Latin America with radio -84- schools. Each school is a group, usually of adults and often based on a single family, who study an organised curriculum by listening to the radio and following printed lessons. The schools are supported by central, non- government, organisations such as ACPO (Cultural Popular Action) in Colombia, which have, for the most part, been established by the Roman Catholic Church. Both farm forums and radio schools are expected to run for a number of years. A different kind of group study has been developed in Tanzania - the radio campaigns. Here, efforts are made by a wide number of government and party organisations to set up a very large number of learning groups, embracing a significant proportion of the adult population, but for these to work together only for two or three months. Campaigns have been used for political, health and agricultural education in both Tanzania and Botswana. 4.6 Radio schools. There are some 25 radio school organisations in Latin America. Their intention is to teach people to read and write and these provide a basic education, geared to the interests of rural adults. Agriculture is an important component of much of their teaching. Detailed case studies have been published on three radio schools, ACPO in Colombia, ACPH in Honduras and Radio Santa Maria (RSM) in the Dominican Republic. We concentrate mainly on these three institutions, but look also at evidence from others, notably from ERBOL in Bolivia and Radio ECCA in the Canary Islands. All radio schools have broadly similar methods of working, which derive from those developed by ACPO, which was established in 1947. 4.7 ACPO was set up by a priest, (now) Mgr Salcedo, a radio amateur, who wanted to use radio in order to reach his congregation. ACPO's aims are: '(1) motivation of the campesino or rural farmer for development; (2) human promotion or education of the whole man; (3) integration of the campensino into society; (4) organisation and development of the community especially by participation in local organisation; (5) ;productivity of the campesino in his agricultural work; (6) spiritual development of the campesino in his personal life.' (McAnany 1975, p.245) 4.8 ACPO, and other radio schools with similar sims, have interpreted these to mean that their curriculum should combine the secular - basic reading and writing and the elements of health and agriculture - with the religious. 4.9 Each radio school organisation consists of a network of learning groups, which are supported by monitors, and provided with teaching material both in print and over the radio. The groups are often based on a single family. They meet once or more often a week, listen to a radio programme, and work through correspondence lessons sent to them from the headquarters of the radio school. Each group has a leader as a member of the group, who must be able to read and write, and who helps the others learn. As the radio school movement was started by the Roman Catholic church, and remains supported if - -85- not dominated by it, the local priest often plays a key role in encouraging the establishment of radio schools and in helping to maintain them. In Colombia, for example, 'groups are more successful where the parish priest is sympathetic, and where he is relatively young. If he is too old he can- not easily climb up the mountains to visit nearby schools.' (Young et al. 1980, p.98). Monitors also support groups: in most cases these are volunteers, who are paid a small amount by the radio school, and trained in the work of helping groups. The RSM monitors meet their groups once a week, act as a channel for sending school-members' work back to headquarters, and retain a proportion of the fee which members pay and which they collect. In Colombia there is a more complex structure with a hierarchy of monitors and supervisors, who are trained in short residential sessions by ACPO at its headquarters. But in all cases the monitor's function is different from that of a teacher: it is not to teach the subject matter of the courses, but to guide a learning group, with its own member/leader, on using the radio programmes and printed study materials. 4.10 How successful have the radio schools been. McAnany has suggested that there are two common problems in answering this question; '(1) The goals of the projects are often so vaguely stated that it is difficult to pin them down to see whether they have been achieved. If a radio school proposes to create the "new Latin American man" among its audience, the task of defining the goal and estimating its accomp- lishments often prevents any real'performance evaluation. (2) Project directors frequently (and often sincerely) confound effort and performance; thus they will say they have accomplished a good deal in creating a radio station, developing programming, printing study guides and training monitors and supervision. All this is true and an evaluator seems un- grateful if he presses on to ask how many people have become literate through the process; yet that is a performance question.' (loc. cit.) 4.11 A third difficulty arises from the radio schools' different perception of their role. In many Latin American countries there are severe limits to the change which education can make to campesinos' lives in the absence of land reform. Radio schools have therefore been faced with questions about their role as reforming or revolutionary institutions. For its part, 'ACPO considers it beyond its jurisdiction to promote peasant organisations with vindictive purposes. "We leave politics to the parties," says Monsenor Salcedo.' (Beltran, 1976, p.239). In contrast, MEB in Brazil was committed to structural change aimed at economic and social reform: after the military take-over it was effectively closed down. In Honduras, ACPH has tried to steer a path closer to structural change than ACPO's without jeopardising its actual existence. 4.12 It is therefore difficult to gauge how far radio schools have met their own objectives but we can try to answer three simpler questions: who have -86- they taught? did their students learn anything? did it make any difference to their lives? 4.13 First, as to numbers. Where, as in Colombia, radio schools are broad- casting and distributing a newspaper, it is important to distinguish between the numbers enrolled for courses and the, generally larger, numbers listening to radio programmes or reading the newspaper. Enrolment figures for ACPO, ACPH and RSM are given in Appendix A. Enrolments have varied from year to year but as a rough figure we can assume that they lie between 110 000 and 180 000 annually for ACPO, between 10 000 and 17 000 for ACPH, and between 12 000 and 23 000 for RSM. Using figures for 1979/80 we can make a very crude comparison between enrolment and target audience, as in table 3. Table 3: Proportion of rural adult population reached by radio schools Country Total % aged % Estimated Radio school population 15-64 rural total rural enrolments as (millions) population proportion of aged 15-64 rural adult (millions) population Colombia 25.61 56 30 4.3 2.6 - 4.2% Honduras 3.4 49 64 1.07 0.9 - 1.6% Dominican Republic 5.1 51 49 1.27 0.9 - 1.8% 4.14 If we assume an enrolment of between 110 000 and 180 000 for ACPO, of between 10 000 and 17 000 for ACPH and 12 000 and 23 000 for RSM this suggests that they are, in any one year, reaching the following proportion of rural adults: ACPO 2.6 to 4.2% ACPH 0.9 to 1.6% RSM 0.9 to 1.8% *These figures tally with the more limited figures from Radio ECCA in the Canary Islands: in 1975 they had 20 263 students in a total population of 1 167 524. If we assume that 40% of the population are rural adults, then Radio ECCA is reaching some 4.3% of them. (For its part, it estimates that it taught 10.12% of the total population between 1965 and 1975.) (Espina 1976). 4.15 Brumberg (1975 p.35) estimated that 'ACPO may have reached roughly 4.5% of its potential rural audience in 1968'. The figures are necessarily crude, and assume - wrongly - that radio school students are all adults. But they are near enough the same order of magnitude for us to be fairly sure that radio schools can, in any one year, enrol between about 1% and 5% of their target population, the rural adults. 4.16 Unfortunately, we know rather less about who these students are, in terms of their age, sex or income. The literature on ACPO confirms that the majority -87- of its students are peasant farmers but are not the poorest members of society. We have fuller information about RSM in the Dominican Republic. The students there have a median age of 18.4; 56.9% of them are women and 84% of the total live in the countryside. 4.17 We know something about their wealth. In Colombia ACPO attracts very few landless peasants (Brumberg 1975, p.39). Radio Santa Maria, in the Dominican Republic, had considerable though limited success in reaching the landless and those with the smallest holdings. White concludes that RSM students have a somewhat higher socio-economic status within rural communities (though not within the larger national social status system). (White 1976a, p.25) Evidence from Bolivia confirms that radio schools can reach the poorer members of society; 63% of ERBOL's literacy students in 1976 and 83% of its second and third year students had an Amerindian language as a mother tongue and we can therefore assume that these were generally poorer than speakers of Spanish. (Detailed figures are in Appendix A.) In general, therefore, radio schools have considerable success in reaching poor members of peasant communities but seldom reach the poorest. 4.18 How far did the students benefit? The fullest information is from ACPO. Their own figures suggest a fair success in their literacy teaching: between 1960 and 1968 they enrolled 800 385 illiterate students, 48% of them women; 278 948 or 35% of the total learned to read and write. (Brumberg 1975, p.37). Over the period from 1961 to 1968 between 61% and 85% of those who enrolled on the correspondence course and took their examination succeeded in passing it. Perhaps more important, from 1962 to 1966 the proportion of students taking the examination remained constant at between 53 and 56% - a high completion rate. (ibid. p.38)' For reasons that Brumberg does not explain, the completion rate fell to 26% in 1967 and 22% in 1968. 4.19 There is suggestive evidence that ACPO does change its students' atti- tudes: studies have been consistent in reporting that ACPO students score higher on measures of modernity, on knowledge of and attitudes towards innovation, and on their integration in the community as measured by such things as membership of co-operatives. The evidence is not conclusive, and differences between ACPO students and others cannot be attributed solely to membership of ACPO. Musto (1971, p.180), for example, found that peasants' modernity was more closely related to the socio-economic level of the region than to the influence of ACPO, while their innovativeness was as closely related to the socio-economic level of the region as to ACPO. Nevertheless, in the same region, ACPO students were consistently more -88- innovative than non-students. Brumberg compared two communities, one highly influenced by ACPO and one hardly touched by it and, while he did not have the data for a rigorous comparison, found similar differences. In the area influenced by ACPO he found more household improvements, such as latrines and fences and more children vaccinated than in the other. He also found 57% of farmers claimed an increase in agricultural productivity in the previous three years, compared with 34% in the other community. Thus, such evidence as there is all points the same way in suggesting that ACPO does have an influence,.necessaily difficult to quantify, on its students. 4.20 This must be seen in context. Schramm and Lerner have suggested that, while ACPO members have learned from their courses, they have not taken the lead in educating and mobilising others. ACPO has been more successful in fostering individual development of its students than in helping towards structural change in society (cf Beltran 1976, p.236 ff) and its appeal seems to have been to the more traditional rather than more radical peasants. 4.21 The evidence from Honduras and the Dominican Republic is consistent with that from Colombia in suggesting that radio schools can more easily teach the subject matter of basic education than promote more far-reaching changes in their students' lives. For four years in which Radio Santa Maria was offering a basic education course roughly equivalent to primary school, its students achieved the results shown in Table 4. Table 4: RSM students 1971/2 to 1974/5 Enrolled Took examinations Passed examinations no. % no. % 1971/72 12 238 6 823 55,7 5 813 85.2 1972/73 22 375 15 694 -70.1 13 915 88.7 1973/74 20 171 11 690 58.0 10 476 89.6 1974/75 20 109 13 790 68.6 11 936 86.6 Source: White 1976a, pp.54-5 Completion and pass rates are high. These are similarly high for ERBOL in Bolivia and Radio ECCA in the Canary Islands. Moreover, White established that 'Students in grade 6 and 8 in the basic education programme of Radio Santa Maria have higher median scores on the certificate examinations than equivalent grade levels in the official system using conventional classroom methods when the field teachers of Radio Santa Maria are functioning at least adequately.' The basic education programme had broader objectives but was less successful in meeting them than in teaching the content of the basic education programme and had little effect on students' social, political, religious and family attitudes. (White 1976a, p.82) -89- 4.22 In Honduras, ACPH had limited success in teaching literacy through radio school methods. Where students had no previous schooling, less than 30% achieved even minimum functional literacy in one term although 75-90% of those who had been to school - even if they had later forgotten to read and write - did reach this level. (White 1976b, pp.44-7). ACPH set up a Popular Promotion Movement (PPM) which was teaching about health and agricul- ture, as part of basic education. PPM had some success in teaching peasants about health practices and agricultural methods but less in encouraging their adoption. 'the PPM, through the radio schools and other means of communication, has been successful in communicating a level of agricultural and health knowledge to its participants which is above that of the general campesino population, even though the non-PPM control group in this case included some relatively modernised communities. However, the discrepancy between agricultural knowledge and practice among radio school students and PPM participants in general suggests that the overall conclusion that participants are gaining knowledge skills but not applying them in concrete circumstances of life appears to be true also with regard to agricultural education' (White 1976b, p.87) Health education, on the other hand which was linked with the activities of homemakers clubs, appears to have been more effective and White found a correlation of .445 between the length of radio school membership and health practice (ibid. p.94). He concludes: 'First, the campaigns of the radio schools, in conjunction with the homemakers' clubs, have emphasized preventive practices which are within the reach of most socioeconomic statuses and do not require any credit support or more sophisticated technical knowledge, as do yield- increasing agricultural inputs. Secondly, the presence of the health centres in the rural central-place towns and the occasional visits of vaccination brigades from the regional hospital supported the radio school and homemakers' club efforts and made vaccination and prenatal care more feasible. Thirdly, the homemakers' clubs have encouraged the direct implementation of health practices by creating a social force in the community motivating women the implement health practices, by maintaining a supervisory and technical assistance system directed toward solving the concrete problems of rural women, and, in some cases, by providing credit, medicines, and other supplies in the local community'. (ibid. pp.94-5) 4.23 kadio schools, then, have succeeded in attracting between about 1% and 5% of rural adults annually. In addition to these, their radio programmes and newspapers have reached much larger eavesdropper audiences. They have established organisations which can serve students in their thousands, and have developed a method of teaching which achieves a fair measure of success in educational terms. In other words, radio school students have relatively high results in terms of examination results or promotion from the equivalent -90- of one school grade to the next. The schools appeal to a significant proportion of relatively poor peasants, though seldom to the poorest. They have been less successful in achieving any major transformation of their societies. 4.24 Farm forums. As we saw, there is another major style, or tradition, of linking broadcasts with group study - that of the radio farm forum. The idea goes back to the Canadian farm forums which existed from 1940 to 1965. Their motto was 'Read, listen, discuss, act'. Groups of farmers met together to read an explanatory pamphlet, listen to a radio broadcast, discuss it, and act together in the light of their discussions and of the information they had received. The intention was to combine local energy and activity with information and support which came by radio and by print. The original groups were set up with advice and support from the Canadian Adult Education Association and the Canadian Federation of Agriculture while the broadcasts were made by the Canadian Broadcasting Corporation. Feedback from groups was used in making programmes. 4.25 Since then forums have been widely established in Asia and Africa as a way of teaching farmers. In Francophone West Africa radio clubs influenced by the French tradition of animation have been established and run on very similar lines. Forums or clubs have been reported from India, Pakistan, Thailand, Nepal, Ghana, Benin, Nigeria, Zambia and Senegal, though there is rather less information about their extent and their effects. Table 5 sets out the reported facts about clubs and forums. Table 5: Radio Forums and Clubs Country Date forums Number of forums established Dahomey a/ n/a 600 (1972) 700 (1975) b/ Ghana 1962 200 (1968) 300 (1971) c/ India 1956 22 500 (1977) d/ Niger e/ 1962 42 (1966) 78 (1972) / Senegal 1968 57 (1969) 270 c(1971) h/ Togo 1964 152 (1966) i/ Zambia 1967 630 (1973) i/ Notes: a. McAnany 1972a b. Dicko 1977 c. Donkor 1979 d. Khan 1977 e. McAnany 1972b f. Lefranc 1967 g. 28 clubs and 50 learning groups h. Cassirer 1973 i. Kahnert 1967 j. Edington et.al. 1974 -91- Forums generally seem to have attracted 10 to 20 members each. But they have not grown into a mass movement: nowhere does the membership of farm forums reach anything like the proportion of the target population reached by radio schools. A partial explanation for this seems clear. There is ample evidence that the field organisation to support forums is crucial to their success. In his analysis, for example, of the Indian forums, Schramm points out that the trial forum project in Puna in 1956, which was much more successful than its longer follow-up, had much more generous field staffing. (Schramm 1967, pp.110-120). In Canada the forums could rely on support from both the farmers' union and the adult education association. In Zambia extension agents have had the job of supporting farmers while in Ghana a specialised staff travelled round to support existing forums and encourage the establishment of new ones. But this kind of support inevitably imposes another burden on government field staff and on government finance. Govern- ments have not generally been so convinced of the value of forums that they have been willing to finance extensive field support for them. In Uttar Pradesh in India, for example, the government was not able to appoint even one forum organiser for each clistrict with a population of nearly 2 000 000. (ibid. p.120). At the same time, as forums have usually been established by government agencies, they have not built up the networks of voluntary and part-time workers who support radio schools. In practice, then, govern- ments have not been convinced of the value of forums. They may or may not be right. To assess that, we need to ask questions about the nature as well as the size of forum audiences and about the effects of forums. 4.25 The original forums in India attracted illiterate as well as literate members and members who were willing to go on attending meetings. Two-thirds of them attended 18 out of a possible 20 meetings. (Studies of the larger scale forums, set up after the first trial, suggested that only half of the nominal members then attended regularly. This is a similar finding to that in Ghana where about half the members attended half or more of the meetings.) (Abell 1968, p.44) A study in Ghana in 1964-65 found that 54% of forum members were women and 37% had never been to school; 53% were not literate; threequarters of them were primarily engaged in farming. Village leaders had been asked to recruit forum members with the intention that they should represent a cross-section of the population. As in 1960 literacy throughout Ghana is reported as 27% it seems that forums did have some success in appealing to the least educated people in villages but only limited success (ibid. pp.29-32). In Senegal it was found in 1969 that literacy groups were broadly representative of their villages although no fuller details of the membership were provided (Cassirer 1973, p.42). Membership is of course -92- dependent on the local culture. In Senegal men and women sit separately as they listen to broadcasts and the reports do not say how far each group con- tributes to discussion. In Niger, however, less than 10% of forum members are women. (Rondeau 1973, p.71). So far as it goes, the evidence is con- sistent that forums are as successful as other methods of education in teaching a cross section of the adult audience. It seems likely that their membership is less skewed towards the wealthy than, for example, the work of extension agents. 4.27 We turn to,their effects. Some forums have none; reports on forums in India for example, make it clear that some of the forums which exist on paper no longer actually meet. And of course effectiveness depends on how well they are run. Experience in Zambia has suggested that 'The Chairman chosen is usually an elderly man who commands respect of the villagers whilst the Secretary is generally chosen from the ranks of the younger literate men. Experience has shown that these appointments are often critical and that it is as unsatisfactory for the villagers to elect their officials as it is for the Extension officer to appoint one. A selection compromise seems to be the answer.' (Edington et.al, 1974, p.208) We can distinguish two sorts of effect. First where forums have been estab- lished with an effective two-way communication system then 'a channel of communication exists between the Ministry and many of the farmers which is immediate and reliable'. (ibid. p.212) Similarly, in Senegal Radio Educative Rurale has been seen as valuable because it provided a channel through which the views of peasants could be made known to the highest level of government. It has also been criticised as a palliative which reduced peasant opposition while little was in fact done about their economic and social disadvantages. Nevertheless the evidence is clear that a forum organisation can facilititate communication between government and the field. 4.28 Second there is some evidence about more direct results in the country- side. The first experiment in India showed that forum members did learn as a result of their attendance: there were significantly greater gains in learning among forum members than among a control group, while illiterate members gained more than literate. Forum members also reported on decisions they had taken about action. Groups intended for example to make compost, or to dig a well, or sell milk through a co-operative, or to use rat poison while conserving rat-killing wild animals. (Schrsmm 1967, pp.114-6). In Ghana, surveys were carried out before a 20 programme series of farm forum meetings and six months later after the series. Villages with farm forums were compared with a control group. It was found that 'ten of the fifteen subject matter broadcasts showed evidence of the positive influence of local -93- village listening discussion groups (The Forums). Desired knowledge, action or attitudes were present among 10% to 27% more of the respondents' in forum villages than in the controls (cf. Appendix A and Abell 1968, p.35). While this was a modest study, conducted for a short period, it does seem to con- firm the efficacy of farm forums, both for increasing the farmers' know- ledge and for promoting the adoption of innovations, though some of their findings might be attributed to a Hawthorne effect. The evidence is thinner than we would like. As Edington et. al (1974, p.211) say of the Zambia farmers: 'whether the service helps to increase agricultural and livestock production is a question at present impossible to answer ... there is no doubt that this service is making an important contribution to the morale of the farming communities. What might, perhaps, be undertaken one day is a survey of activities which individual farmers claim were inspired by farm forum broadcasts. One thing which is clear is that many farmers find it difficult to sustain these activities. In Zambia again for example, of 29 forums examined in 1973, 14 were meeting regularly, 7 'seemed to be urgently in need of considerably more super- vision' and 4 'appeared t:o be virtually defunct'. 4.29 We can sum up: - Forums have reasonable success in attracting a cross-section of their potential audience. - They have failed to attract sufficient support from governments for them to expand to the point where they reach significant proportions of the population. - They provide a two-way channel of communication between government and farmers. - There is evidence that farmers learn through attending forums and make changes in their methods as a result. - We do not know enough about the attractiveness of the farm forums where most farmers own their individual radios. 4.30 Radio campaigns demonstrate another way of linking group study with the use of radio and print. The campaign method was originally developed in Tanzania in the late 1960s and early 1970s. Groups met for between two and four months at intervals of once or twice a week and listened to a radio programme. Each group had a group leader who had been trained in his role of guiding the discussion and a secretary who sent back reports on the group discussion to the campaign organisers. Whereas radio schools and farm forums have reached only relatively small proportions of the population, the aim of radio campaigns was always to involve very large numbers. The conse- quence of this was that a large number of political and national organisations were pressed into service to support the work of campaigns. A large organisation -94- had to be created in order to deliver printed materials to group leaders, to train those leaders, and to handle feedback from the groups. Tanzania used both its adult education organisation and its party organisation for this purpose. In Botswana, which is a multi-party state, political organisations were not used but a whole variety of government departments co-operated in supporting the work of groups. Table 6 sets out the basic facts about the campaigns in Tanzania, and in Botswana which followed the Tanzanian example. 4.31 The early campaigns in Tanzania were all on political or civic education. The first large scale campaign in 1969 aimed at explaining to people the second five-year development plan and this was followed by campaigns on national elections and on a variety of civic issues which seemed of import- ance to government at the time when it was celebrating 10 years independence. In 1973 with the campaign on health there came a change: the emphasis now was not only on education but also on practical activities which people could undertake. The campaigns on health education in 1973, on nutrition in 1975, and on forestry in 1981 were all intended to leave behind practical monuments in the countryside in terms of better health practices, better nutrition and more trees. 4.32 The model was transported to Botswana in part as a result of one adult education professional moving from Tanzania to Botswana. Here, too, the first campaign in 1973 had a political theme and its intention was to inform people about the national development plan for 1973 to 1978. The 1976 campaign in- troduced a variant: the Botswana government was developing a new policy for the use and management of tribal grazing lands, which covered a very large proportion of the country's surface area. One of the purposes of the campaign was to consult the public about the way in which the policy should be implemented, area by area and district by district. Feedback from groups was used to guide the developing policy by regionally and nationally. This national campaign was followed three years later by a local campaign in the western Kalahari, whose aims included both informing people about the district development plan and consulting them about local development issues. The purpose was to provide feedback to local rather than central government. 4.33 There is a fair body of literature on the campaigns, so that it is possible to write with some certainty about their audiences, the extent to which participants continued to attend group meetings, how much they learned, and what results the campaigns had in villages. We look at each of these questions in turn. -95- Table 6: Principal radio learning campaigns in Tanzania and Botswana Title and content Date Length No. of No. of % age groups people women Tanzania To plan is to choose 1969 8 x 1 600 n/a n/a (Kupanga ni Kuchagua) hour Second 5-year develop- program- ment plan mes The choice is yours 1970 1 month n/a 4 000 n/a (Uchaguzi ni Wako) National elections A time to rejoice 1971 10 n/a 20 000 38 (Wakati wa Furaha) Celebration of 10 years of independence Man is health 1973 12 75 000 2 million 49 (Mtu ni afya) weeks Health education Food is life 1975 4 n/a over n/a (Chakula ni uhai) months 2.5 million Forestry 1981 5 n/a n/a n/a months Botswana The people and the 1973 10 pro- 1 432 12 000 n/a plan (Setshaba le grammes Togamaano) in 5 National Development weeks Plan 1973-78 Our land (Lefatshe 1976 10 pro- over 55 300 50 la rona) gramnes 3 000 Government plans for in 5 use of tribal grazing weeks lands Learning about govern- 1979 10 pro- 257 3 060 n/a ment (Lesedi la puso) grammes Regional campaign of in 5 civics education weeks -96- 4.34 First it is clear that the campaign technique can reach very large audiences indeed. Starting on a fairly modest scale, the Tanzanian campaigns built up to the point where they were numbering participants in their millions. With figures of this size one cannot expect precision but the Tanzanian experience suggests that one can reach between one in five and one in ten of the adult population in this way. In Botswana the second campaign, which had very much fuller support from government than the first, reached 18.5% of the adult population. 4.35 The campaigns also had reasonable success in reaching the less educated and poorer members of society and reaching people in the countryside. And as time goes on campaigns seem to have become more successful from this point of view. Thus in Tanzania while 'A time to rejoice' had 72% of farmers among its audience the figure for 'Man is health' was between 90 and 95%. In Botswana the campaign in 1975 reached 9% of the total rural population as compared with 3% of the total urban population. We get a similar picture if we look at the schooling which radio campaign participants had received. 31% of the participants in 'A time to rejoice' had received five years or more of education while this figure drops to 18% for 'Man is health'. In Botswana we have figures just for 'Our land' where 37% had received no education, 31% had been to school for 1 to 4 years and 32% had been to school for 5 years or more. 4.36 Evidence about the economic status of participants is thinner. However questions were put to participants in Botswana in 1976 on their ownership of cattle. This is an issue so sensitive that the results must be treated with some caution but 40% of participants claimed they had no cattle at all. This exactly reflects the proportion of households without cattle in the country generally. 4.37 Where we have evidence, the campaigns seem to have been more successful in attracting men than women. 'A time to rejoice' had only 38% women in its audience but this figure rose to 49% for 'Man is health'. In Botswana in 1976 the sexes were more or less evenly divided. However, as 58% of the population in Botswana were women, recruitment was in fact biased towards men. 4.38 Thus we can conclude that the campaigns had considerable success in reaching the poorer members of society but not complete success. In Botswana for example in 1976, 61% of group members were literate - a higher percentage than one would expect in a random sample of the adult population. In the regional campaign there in 1979, while we have no detailed breakdown of the audience, the report does show that relatively few of the Basarwa, often the poorest group in society in the western Kalahari, participated in it. .-97- 4.39 Of course not all participants attended all meetings. Generally it seems that people went to about two-thirds of the total. In 'A time to rejoice' the average was 65%; it was 63% for 'Man is health' and 61% in Botswana for 'Our land'. More seriously Matiko (1976, p.14) reported that from a small sample 34% of the groups working on 'Food is life' in fact broke up. And there is evidence that even over the short period during which campaigns were running participants stopped coming to meetings. There is a significant drop-out rate even for short campaigns. With 'Man is health' the audience fell most steeply between the first and second weeks and there was then.a decline by a further 10% in all from the second to the tenth week. There was a rather higher drop-out in the regional campaign in Botswana in 1979 with 49% of the audience dropping out between programme I and programme 10. While the figures are not out of line with those for adult education classes in Botswana or Tanzania generally, they are high enough to be worrying in view of the very short period of time over which campaigns were running. 4.40 The evidence is consistent that participants did learn from taking part in radio campaigns. In a number of campaigns pre- and post-tests were admini- stered and it was found that there were learning gains among participants and, where significance tests were applied, that there were significant differences between those who had attended learning groups and control groups. In a number of cases, however, we have no such data. For 'Our land' for example, information about people's background knowledge of the subject of the campaign was collected before it began, but no follow-up study was done. 4.41 Where we have figures they are consistent and impressive. Thus for 'A time to rejoice' comparisons were made between groups using a pre-test and post-test measure of knowledge. There was a difference in mean score between pre- and post-test results of 11% (means of 72% and 83%). This is significant at a better than .01 level. (Hall 1972, p.47). For"'Man is health' Hall ex- amined both pre-test and post-test scores of groups participating in the campaign and compared these with control groups who were not participating. He found a rise of score from 42% to 63% (significant at the .01 level) for pre- and post-test scores of participating groups. He also found a difference between the scores of 28% for the participating control groups which was sig- nificant at the .05 level (Hall 1978, p.52). Pre- and post-tests were also carried out in Botswana for 'Learning about government' and the report concludes that 'most questions asked in the surveys revealed a moderate gain in knowledge. The proportion of people giving correct answers to the factual questions generally rose by 10 and 20% points. Gains in knowledge were greater for those who stayed with the campaign until the end'. (Boipelego Education Project, nd, p.112). Unfortunately no significance tests were carried out here but the evidence is consistent with that from Tanzania. -98- 4.42 The later campaigns in Tanzania aimed at making changes in everyday village life and tests of knowledge gain are thus an inadequate measure to assess their success. We have the fullest information on the results of 'Man is health'. The evaluation team in Tanzania carried out a detailed survey of health practices in sample villages before and after the campaign with the results shown in Appendix A. The team established that the cam- paign did result in measurable changes in health practices in the villages surveyed. Hall goes on to report that 'on a national basis, the digging of pit latrines proved the most successful single activity of mtu ni afya' (Man is health). His table is reproduced in Appendix A. He also noted that 'the average rate of latrine use increased 123% in the eight villages'. (Hall 1978, p.58-9). A later survey suggests that after four years latrine usage generally had increased to include 59% of the population and suggests that some of this increase, taking place over a period of years, should probably be attributed to 'Man is health'. (Maliyamkono et.al 1978, p.79). 4.43 Campaigns,then,can leave physical monuments in the countryside as well as reports of learning gains in the researchers' documents. 'Our land' in Botswana had one quite different aim: to consult people about a government programme and provide information from those people to district and central government. The evidence is clear that the campaign did result in feedback from the learning groups to government: 25 000 report forms, 70% of the maximum possible number were submitted during the campaign. The Ministry of Local Government and Lands Report on the campaign records that 'Government, which required that there should be a public consultation on the TGLP policy proposals, is satisfied with the campaign's achievements and is making use of the data produced by the consultation'. (Ministry of Local Government and Lands 1977, p.124). No further details are given of the extent to which government policy at local or national level was affected by the consultation. 4.44 Experience of campaigns is limited to two countries. Undoubtedly in Tanzania where they have been most extensively used they have been helped by the political structure which Tanzania has created for communication through- out the country and for adult education. But we can sum up with four con- clusions which would apply more generally: The campaign method enables educators to reach extremely large audiences although, to do this, it is necessary to set up a large organisation to service groups and train group leaders. This imposes major burdens on the agencies responsible. Campaigns have some success in reaching the poorest members of society and in reaching non-literate students. Campaigns appear to have a successful record both in enabling people to learn from the radio, from print and from group discussion and in -99- stimulating activity on the part of participants. As it is of the essence of such campaigns that they run for a short time, there are severe limits to the educational content of a radio learning group campaign. Even the successive campaigns in Tanzania did not add up to a curriculum comparable to that of the Latin American radiophonic schools. 4.45 Other groups. Although most of the experience of linking mass media with group study has come from radio schools, farm forums or radio campaigns, a number of organisations have used group study in other ways and used mass media to support the work of existing groups. In Lesotho, for example, the Lesotho Distance Teaching Cenitre has used radio and print to help existing groups to learn. In Botswana, the Botswana Extension College provided radio and print support to the work of Village Development Committees, and of agricultural youth clubs (4B clubs). There is also well documented experience from Tanzania of using mass media to support co-operative education at the Co-operative Education Centre. 4.46 Such projects have more modest aims than those of radio schools, for example. They are usually using mass media to teach a narrower range of skills and there is little evidence which compares their effects with the effects of alternative approaches. Studies in Botswana suggested that the techniques were of value for teaching about village development and for teaching young people about growing fruit trees. There was a higher survival rate for fruit trees planted following a print and radio educational programme than there had been in a previous scheme without that support (BEC Evaluation Unit 1975). Some radio schools, too, have run agricultural courses with more modest aims than their programmes of basic education. In Chile, for example, radio school methods were used to teach about horticulture, about goose keeping, and about rabbits. Participants introduced the recommended methods horticulture and goose keeping but not of rabbit keeping, where they could not afford the breeding pairs. (Amtmann et.al 1977). 4.47 In Lesotho, one meticulously researched experiment compared the use of printed matter by groups with its use by individuals. Lesotho Distance Teaching Centre prepared pamphlets on how to crochet and made these available both to individuals and through existing women's organisations. Group leaders were given a limited amount of training in how to use the booklets and how to conduct their groups. Under these conditions, where groups used the booklet as recommended, 85% of group members performed an end of course test correctly; 96% of beginners learned to crochet. Four times as many group members were successful in learning to crochet as compared with private indi- viduals who received the booklet. (LDTC 1978, pp.36-42). -100- 4.48 The Co-operative Education Centre (CEC) in Tanzania has used group study, supported by radio and correspondence courses, for co-operative education. Between 1965 and 1970 CEC used these methods to teach about 15% of all co-operative committee members. Many of these were illiterate so that it was only through the group method - with at least one literate member in each group - that they could follow the course. During the period 22% of. group members completed this course and received theircertificate, compared with 28% of individual - and necessarily literate - students. The ultimate completion rate was expected to be higher (Grabe 1975, p.603). 4.49 Limited though it is, there is, then, some evidence for the effectiveness of this mode of group study. It has one important limit: it reaches those students who are members, or potential members, of the organisations on which it is based and its reach into the community is likely to be no longer than theirs. 4.50 There are theoretical grounds for believing that group learning, linked with mass media, should be effective. It allows personal encouragement and discussion to support the suggestions made by specialists in print or by broadcast. And it allows learners to discuss ways in which the, necessarily general, information from these media can be applied in their own community. Practice tends to confirm the theory and we can conclude, on groups generally, that they do provide a mechanism for large numbers of people to learn both educational material in general and agriculture in particular. The wide variation in styles of group learning between different countries, and different continents, make it more difficult to generalise about the best pattern of organisation for a particular country or a particular educational problem. Where such an organisation exists, group study makes possible communication from learners to educational or agricultural establishments. But their success depends on the existence of an effective supporting organisation on the ground, as well as the production of effective teaching materials. B. Mass media for individuals with feedback 4.51 Mass media are often used as one-way channels of communication. Usually, there is no ready channel for a newspaper or a radio station, for example, to receive regular feedback from listeners or readers. Of course both inevitably receive some feedback: producers travel the countryside seeking reactions to their programmes, newspapers get letters from their readers, figures of listenership or sales provide some information to managers. It is useful, however, to distinguish between that kind of feedback and attempts to provide it in regular and systematic ways. A number of broadcasting - 101 - projects have created structures which assure them of regular feedback from listeners and have used this feedback as a basis for their programmes. In a rather different way INADES-formation uses feedback on a regular basis in teaching better farming techniques through print. Two long-running projects using regular feedback have been well documented - Dialogo and INADES-formation. 4.52 In Costa Rica the Centro De Orientacion Familiar (COF) began broadcasting programmes about human sexuality and family planning in 1970. It broadcast for 20 minutes, six days a week, at 8 p.m. (except during presidential addresses and crucial soccer matches) (Risopatron and Spain 1979, p.17). Over the next nine years Dialogo received between 1 000 and 1 700 letters from listeners annually. A survey in 1978 showed that 53% of adults in Costa Rica heard from the programme and 36% did so regularly. More striking: 'The Dialogo audience is a representative cross-section of Costa Rican society. With the minor qualification that Dialogo listeners were a bit more rural, and less educated, and less likely to read the newspapers daily, we can still say unequivocally that Dialogo's audience includes substantial proportions of every sector of Costa Rica's society ... ... Dialogo's strongest grip was precisely on the poor, rural, and less educated.' (ibid. p.vii) The study also found that there were 'strong associations between ialogo's listenership in knowledge, attitudes and practice in human sexuality matters' (ibid. p.67). It speculates too on the relationship between Dialogo and the continuing decline in the birth rate in Costa Rica during the period it has been broadcast. 4.53 A more modest experiment was carried out in Botswana, where a radio question and answer programme Re Botseng (Ask Us) was broadcast by Botswana Extension College from April to December 1977. A follow-up study showed that the programme had a limited success in reaching rural audiences: 59% of the audience was rural as against between 80 and 90% of the total population. But the programmes were more popular in the countryside than in the towns. In contrast with Dialogo, 65% of those who sent questions were men. The popularity of the programme was shown by the fact that 68% of a sample tried to listen to the programme regularly. Perhaps more significant, by the time it- had been running for about six months, some 60 letters a week were arriving - equivalent, if that figure were continued, to a rate of 3 000 per annum, or a response by letter from 1% of the adult population. (BEC Evaluation Unit 1978). (In practice the rate was not maintained as the programme came to an end in December.) 4.54 INADES-formation, which is based in the Ivory Coast but offers courses in some fourteen African countries, uses correspondence courses to teach -102- basic agriculture. The courses are for the most part in French, though in some countries, such as Ethiopia and Rwanda, they have been translated into other languages. INADES-formation also organises face-to-face teaching to support these students. While in the past it has made a number of attempts at group study, it has had relatively little success with these and more recently it has concentrated on teaching individuals. Here feedback serves two purposes. First INADES-formation students complete written worksheets which test their understanding of the course material. This gives them a link with their tutor and enables the tutor to encourage them and help them if they have not understood something. Second, through this information about students and through information obtained at face-to-face seminars, INADES- formation revises its courses and checks that they are in fact meeting the needs of their students. 4.55 INADES-formation courses are, in practice, available only to literate students, and the organisation has not attempted to enrol the maximum pos- sible numbers on its courses. Their figures show that INADES-formation enrolled 4 459 students in 1979/80, but that in only two countries, Burundi and Rwanda, were there over 1 000 students. 4.56 INADES-formation has so far had little success in reaching women or in reaching non-literates through group work. But for those it does teach it appears to be extremely effec-tive. A study in Cameroon found evidence of students learning from their courses and applying that knowledge (Jenkins and Perraton 1981). Its effectiveness might be increased dramatically if it moved its focus of attention from teaching individual farmers to the training of extension agents - something which it has already begun on a fairly small scale. 4.57 On the, admittedly narrow, basis of these three case studies we can conclude that the existence of feedback links to producers of mass media can - help ensure their effectiveness by keeping them in touch with their audience; - lead to very strong feelings of involvement by learners with teaching agencies; - if radio is used, attract large audiences and ones which may have a higher proportion of women, and of the poor, than are reached by other forms of adult education. C. Radio 4.58 The great majority of broadcasts are aimed not at groups but at indivi- duals, and without any attempt to organise feedback. But, while there is a considerable amount of educational and agricultural broadcasting, often -103- produced for ministries of agriculture, community development, education or health, little research attention has been directed to it. 4.59 It is worth putting that educational broadcasting in context. First there is not very much of it in contrast with broadcasting generally: 'less than 5% of the Third World's total hours of radio programming were defined by UNESCO as educational' in their 1974 Statistical Year Book, (Gunter and Theroux 1977, p.288). Second, while educational broadcasts may say one thing, the other 95% may be saying something quite different. The problem is particularly acute in Latin America where broadcasting is dominated more by comm erce than by government. In many third world countries the mass media communicate the values of the industrialised metropolitan countries: 'The media in developing countries take a high percentage of their cultural and entertainment content from a few developed countries, and chiefly from a few large producers in those countries ... The developed countries get the selected best of the culture (chiefly music and dance) from developing countries; the latter get a lot of what on any objective standard is the worst produced by the former. Transnational companies are playing an ever more active role in the worldwide provision of communication infrastructure ... When these influences become dominant in very different cultures, the effect is to impose uniformity of ltaste, style and content' - (MacBride et.al, 1980, p.163) And that uniformity is one which reflects the values and aims of the trans- nationals and their metropolitan basis. These may be completely at variance with the avowed aims of government and development agencies. In Latin America for example, print and broadcasts alike have this character. 'In comparison to the non-Latin American dailies, those of Latin America allocated much less space to developmental information, considerably more to sports and amenities, and a little more space to racially negative information ... ... those studies available on radio and television content suggest that both these media, but particularly television, pay considerably less attention to developmental information; they seem to play up trivial matters to a muc]h higher extent than it occurs in the print media. Studies specifically concerned with mass communication as related to rural development have found that even those messages specially addressed to farmers are geared - in content and in form - exclusively to a minority among these, the city-based large land-owners.' (Beltran 1974, p.374) 4.60 Thus in considering the educational effects of mass tiedia on individuals we have to keep in mind that we are dealing with a small proportion of the total output - whether in print or through broadcast - concerned with values which are likely to conflict with those reflected in the rest of the mass media. -104- 4.61 Nevertheless, radio is important. It can reach the non-literate as well as the literate and today many people even in the poorest countries have access to a radio. In 63 out of the 91 Third World countries surveyed by Katz and Wedell (1978) there was, about ten years ago, already at least one radio to every twenty people and in nearly half, already one to every ten. The figures are summarised in Table 7. It seems reasonable to assume that today, even in the poorest countries, between 25% and 50% of the population have access to radio. Table 7: Distribution of radios in 91 countries Number of countries with one radio to: over 20 11-20 6-10 5 people people people people or fewer In Africa 16 13 11 1 In Asia a/ 9 5 8 4 In Latin America and Caribbean 3 4 6 11 Source: Katz and Wedell 1978, Table A4 Note: a. Including Western Samoa 4.62 We know a little about the use made of radio broadcasts. In India a study in 1967 showed that radio was the 'most commonly attended medium' with an audience of 76% of those surveyed. There was little difference in socio- economic status between those who did and did not listen to the radio but a rather greater difference for those reading the press. Despite these high listenership figures, the study suggested that only 8.6% learned about agri- cultural innovations through mass media, while 52% did so through a change agent such as an agricultural extension officer. A parallel study in Nigeria found that neither the change agent nor radio seemed 'to play an important intervening role'. (Shore 1980, pp.40-43). Other studies do suggest that radio can be an important source of agricultural information, though many of them rely on farmers' reporting about the use they have made of mass media. In the Philippines radio is said to be the 'most preferred medium' for information on agriculture and it was found that 83% of farmers' co-operators used radio. (Gomez 1976, pp.198-9). In Guatemala, radio has been found to be effective both by itself and when linked with the work of an agricultural extension agent (Rich et.al 1978). The Basic Village Edu- cation Project there found that radio was 'extremely effective in bringing about changes in knowledge, in attitudes, and in behaviour' (Academy for Educational Development 1977, p.xiii) and that radio was 'the most important -105- source of new information reported by farmers'. (ibid., p.75) In Bangladesh in the late 1970s one survey found that 57% of village listeners listened to the agricultural programmes occasionally and 8% regularly. (Khanam 1976/77). Evidence from Nigeria is less straightforward. Two studies in the late 1960s found that rural people devoted an 'extremely low proportion of time to such activities as listening to the radio and reading newspapers' and that only 2 to 8% used mass media channels for information on agriculture (Axinn and Thorat 1972, p.164). In contrast a study in Western Nigeria found, though with some uncertainty about the data, that 38% of farmers head broadcasts regularly. (ibid. 165). Another study in Western Nigeria found that many farmers were 'being reached by agricultural programs on the radio. Only one- third of the sample knew their local extension worker and only one-fifth knew him well' (ibid. 166). Similarly in Northern Nigeria in 1972 Yazidu found that 47% of farmers reported using radio as a source of agricultural informa- tion as compared with 20% who got information from their extension worker and 16% from friends and neighbours. (Yazidu n.d. p.47). In Eastern Nigeria 38% of farmers used radios as a source of agricultural information. (ibid. p.49). The difference between the various Nigerian figures may be partly explained by the different dates on which studies were made. In Malawi 27% of farmers reported radio as a source of agricultural information (Perraton, Jamison and Orivel, 1981); a finding that is consistent with the Nigerian data if we bear in mind the difference in wealth between the two countries. 4.63 The evidence is consistent enough for us to be able to conclude: - Radio broadcasts do reach a rural audience which even in the poorest countries is likely to be 30% or more. - They are regarded by listeners as being an important source of agricultural information. - Because they are s9 widespread they probably reach poorer families than are contacted by extension agents, and reach people who cannot read. - Despite this knowledge about their reach, we know little about their effectiveness. 4.64 There is one variant of open broadcasting on which we do have more infor- mation. Three projects have attempted to use advertising techniques with radio spot advertisements repeated frequently over a period of up to a year. These have usually been done in an attempt to change people's health or diet practices. These advertising campaigns had very narrow objectives: as Coombs has pointed out, some communications specialists 'lean heavily towards adver- tising and propaganda techniques rather than authentic educational efforts' -106- (Coombs 1980, p.38). Drilling people, day-by-day, over the radio on what they must do is far removed from the idea of education which lies behind many of the projects discussed. The results are equivocal. In Ecuador for example the National Institute of Nutrition broadcast radio spots for a year with five groups of messages: - on protein and calorie malnutrition - on breast feeding - on the need to boil drinking water - on parasites and diarrhoea - on iodised salt. Many of the messages did result in changes of peoples' knowledge but not in their practices, with the exception of using iodised salt. This was available for sale at the same price as ordinary salt: 'the iodised salt message was immensely successful. Whereas only 5% of highland Mestizos used iodised salt before the campaign, fully 98% of those samples after the campaign were using iodised salt. On the other hand, few of the other messages - which were so successful at the level of awareness and knowledge - achieved significant behaviour change'. (Gunter and Theroux 1977, p.292) 4.65 The result is consistent with parallel experiments in Nicaragua and the Philippines. After a year of similar radio spots a quarter of mothers in Nicaragua claimed to be using an appropriate lemon drink for treating diarrhoea. In the Philippines a quarter of mother used some oil to enrich their babies diets. (We look at the costs involved below.) (Manoff et.al, 1977). The evidence does not suggest that the advertising spot has major advantages over other radio formats. Generally, we need more research on the comparative effects of different styles and formats of broadcasts. D. Rural newspapers 4.66 Rural newspapers of two kinds have been used to encourage and help rural development. Some have been centrally produced, usually by a ministry of education or of information, and distributed throughout a country or a region. Others have been village newspapers, produced in a village or a small town, and circulating only in a small area. Alongside such papers, a number of organisations produce magazines, often with a narrower focus of interest: Za Achikumbi in Malawi, or Agripromo in francophone west Africa, for example, are almost entirely concerned with agriculture. 4.67 The aims of rural newspapers (of the first kind) usually include the following: '1. to produce reading materials for new literates; 2. to ensure the continuing education of the rural masses; to give them practical advice on produce and civil rights and responsibilities; -107- 3. to give the masses information about events concerning their environment, their region, their nation and the outside world at regular intervals; 4. to ensure a "dialogue" between the leaders and the rural masses; 5. to help ensure the participation of the rural masses in the economic, social and cultural development of the nation; 6. to proceed to set up a local, decentralised press and to show the rural masses how to express themselves in the press;' (Bhola 1980, p.35) 4.68 The balance between providing something for new literates to read, and carrying information about development, obviously varies from paper to paper. In the Soviet Union in the 1930s, for example, "The Peasant Newspaper for New Readers was published as a study aid for pupils who had passed the primer stage. It came out six times a month. The subject matter in this newspaper was designed for country people and was of interest to adult collective farmers. For example, the paper's fifty-fourth issue for 1934, published on the eve of the eighteenth anniversary of the Revolution of 1917, carried on its first page such items as these: 'What are your Collective Farm's Accomplishments on this Anniversary of the Revolution?' and 'On the Eve of the Great Holiday'. On page 2, there was an article headed 'Order and Culture in our Communal Affairs'. On page 3, under the heading 'Learn to Count', there was one of a series of arithmetic lessons. And finally, on page 4, there were items of interest to farmers: 'How to store and preserve Seed','Autumn is the time to see to your Seed Stores', 'Store up Seed of Fodder Grasses'. Reading such items broadened the pupils' knowledge of farming and encouraged them to study. Issue No. 67 (1935) carried current political news on its first and second pages, and on page 3 there was another arithmetic lesson on multiplying whole numbers by decimals. The items of farming on the last page included 'Efficient Use of Fertilizers', and 'We started with Fertilizers'. The latter was an article by an outstanding woman collective farmer, M. Demchenko, who was well known throughout the country. It described how she succeeded in obtaining a record yield of sugar-beet per acre. Among the other items on the page were: 'What to do with Fertilizers in Autumn and Winter', 'How best to use Fertil- iz'ers', 'How to make Compost', 'How to use Peat', 'Don't forget Ashes and Lime'. There was also a bibliography on these subjects." (Fundamental and adult education, 1959, p.174) 4.69 Rural newspapers have been extensively developed in francophone west Africa in recent years: Kibaru in Mali, for example, has been widely publicised. These papers generally contain materials of five kinds: news - international, national and regional; information - on agriculture, health or other developmental issues; literature - stories, sometimes recorded from existing oral literature, history and poetry; entertainment - which may include puzzles and riddles, sometimes con- tributed by readers; -108- a means for dialogue - between citizens and between citizens and the newspaper. This often involves including a readers' letters sec- tion in the paper. 4.70 It is not easy to include everything one would like to in such a news- paper. It is difficult to be topical. Kibaru, for example, found that weather news had no interest for readers who received it three months late. (They also failed in their attempt to include lottery tickets in the news- paper as they tended to arrive after the draw had taken place.) 4.71 There are few estimates of the effective audience for rural newspapers. Such basic information as there is appears in Table 8. We can assume that each copy is read by a fairly large number of people. On that assumption, the better established papers probably are reaching a fairly high proportion of their target audience. If, for example, 20% of the adults in Mali are now literate, then there is one copy of Kibaru for every 55 of them. (Though this is only 1.9 copies per 1 000 population compared to an average of 14 copies of newspapers per 1 000 throughout Africa.) 4.72 The resources needed to publish a rural newspaper are of three kinds: the staff and equipment needed to print the paper; a network of correspondents to provide news and information from the countryside; a distribution system to ensure that the paper reaches its target audience. Most are printed at government presses, using spare capacity. The Rwanda newspaper Imveho works with a staff of 5 people but often people work partly on a rural newspaper and partly on other projects. Thus el campesino in Colombia and Agripromo in west Africa are produced by staff from ACPO and INADES-formation respec- tively, but by staff who also work on the agencies' other activities. Similarly, newspapers which have government backing can use government information and field officers as a network of correspondents: Kibaru has a network of 42 which provides it with news from the field. 4.73 Distribution too, may use government networks, such as those used to supply literacy centres. But it inevitably presents problems: 'In localities where the infrastructure of roads, trains and buses does not exist and postal services are infrequent or non-existent, the only choice available to the newspaper publisher is to develop his own infrastructure. That can be very expensive even in projects with scores of landrovers and jeeps shuttling around in the bush or on the campo or in and around the villages.' (Bhola 1980, p.37) 4.74 Naturally, fewer production and distribution resources are needed for the smallest, village-based newspapers which are likely to have circulations measured in hundreds rather than thousands. The other side of the coin is that the smallest papers may have greater difficulty in gathering news, -109- Table 8: Audience for rural papers Country Paper Date Frequency Circulation Date Established Tanzania Elimuhaina mwisho 1974 monthly 45 000 1978 f/ 20 000 Oct.74 a/ Mali Kibaru 1972 monthly 8 000 1976 b/ 12 000 1979 c/ Rwanda Imvaho n/a weekly 7 500 1976 d/ Liberia Wozi Rural Press 1951 weekly 1 300 1979 e/ Togo Game Su (Ewe) 1977 monthly 2 500 1978 f/ CAR Linga (Sango) 1976 2 months 1 000 1978 f/ Ghana Kpodoga (Ewe) 1976 monthly 5 000 1978 f/ Kenya Kisomo (Kikuyu) n/a monthly n/a Bumanyati (Swahili) 1976 monthly 4 000 1978 f/ Notes: a. Bhola (1980), p.38 b. Dicko (1977) c. Hein and Kanyogonya (1979) p..62 d. Dicko (1977) p.76 e. Hein and Kanyogonya (1979) p.70 f. Literacy work 1978-79 p.29 especially of regional, nationaL and international events. Bhola sounds another warning note: We are beginning to realise that for groups to handle media as producers ... they must first develop routine experience with the media as consumers ... Rural newspapers produced by communities them- selves may be possible only after substantial routine experiences with newspapers produced outside the communities' (ibid.) 4.75 To sum up, we can assume that rural newspapers are of value in societies where new literates, and rural people generally, are short of printed material and value what is available. Tley can be set up and run on a modest scale, and can be used to carry information about agriculture or about the content of basic education more generally. There is little information available to assess their effectiveness. -110- V THE COSTS OF MASS MEDIA 5.1 There are a number of difficulties in calculating the cost of using mass media for education. First, where information about costs is avail- able, it is often not in a standard form. Different reports have used different principles in separating direct from indirect costs, in calculating depreciation, and in using, or not using, a social discount rate. This is explicable partly because, till recently, there has been no universal agreed practice, and partly because some costs are unimportant to managers. Project managers, as opposed to macro-economic planners, often have a quite rational lack of interest in costs which can be borne by a donor agency or by another department's budget. 5.2 Second, where information is available, it is often difficult to dis- aggregate the costs of different media. Even though it may be possible to calculate the costs of a project, it is often impossible to attribute them to a particular educational course or a particular medium. 5.3 Third, there are often difficulties in comparing the costs of education through mass media with the costs of orthodox education. These difficulties are compounded in basic education, where we lack even the unsatisfactory but useful benchmarks of public examinations. 5.4 Where evidence on costs is available, however, we may be able to measure costs of three kinds: - cost per unit of production (e.g. cost per hour of radio broadcast, or cost per book produced); - cost per student reached; - cost per successful student (e.g. measured in terms of learning to read and write or completing a course or passing a test). In a few cases attempts have also been made to measure the impact of an educational programme (e.g. in terms of increased agricultural production or of the number of latrines built), and then to assess that in financial terms. 5.5 We look in turn at different ways of using mass media. A. Mass media with group study 5.6 We have some information about the costs of radio schools, campaigns and farm forums, and a more limited amount about other forms of group study. 5.7 Information on costs is available for three radio schools - ACPO in Colombia, Radio Santa Maria (RSM) in the Dominican Republic, and Radio ECCA in the Canary Islands. A detailed analysis of the cost information available for ACPO is in Appendix A. This makes it possible to distinguish ACPO's expenditure on its commercial activity, and on its publication of -111- the magazine el campesino. At the time of the key study, ACPO had 170 000 students while the weekly priint run of el campesino was 70 000. Including capital costs, using a 10% social discount rate, ACPO's costs in 1978 US$$ were as follows: Fixed costs of ACPO's educational activities 2 247 770 Variable costs of ACPO's educational activities 2 709 430 Total costs of ACPO's educational activities 4 957 200 Annual cost per student 29.16 Cost per copy of el campesino 0.21 Using the approach proposed by Jamison, Klees and Wells (1978) the cost function for ACPO is then: TC = AC(N) + F where TC is the total cost, AC is the average cost, N is the number of students, F is the fixed cost, 4 957 200 - 15.94(170 000) + 2 247 770. Unfortunately the figures available do not make it possible to estimate with enough confidence the dropout or progression rate, so we can present only a cost per student taught. Nor do we have comparisons with the cost of conventional education in Colombia. 5.8 Similar figures to these are found at RSM and Radio ECCA. In his study of RSM, White showed that the cost per student in 1975 was $24.31 (equivalent to 1978 $29.90) if one assumed there were 20 000 students, roughly the actual number of enrolments at the time of his study. His figure included an allowance for depreciation and maintenance but did not include an annualised capital cost (White 1976a, p.90 ff). If we assume that the cost should be increased by 20% to allow for capital costs, we get a cost per student in 1978 $$ of $35.88. As some 60% of students are promoted from one grade to the next, the cost per student promoted is 1978 $59.80. His figures (see Appendix A) suggest that, including capital, RSM's costs in 1978 $$ would be of the following order: Fixed costs 342 187 Variable costs 375 519 Total costs 717 706 The cost function, TC = AC(N) + F is: 717 706 = 18.78 (20 000) + 342 187. 5.9 The RSM courses are, to a degree, comparable to those offered at ordinary schools in the Dominican Republic. It is therefore possible to make a comparison between the cost of education through RSM and in the conventional way. Figures are available both for costs per student, and costs per student promoted: on either measure the figures for RSM are below -112- those for conventional schools. The costs, omitting capital in both cases, are $29.90 per student enrolled and $49.84 per student promoted for RSM and $47.86 and $76.80 respectively for orthodox students. 5.10 Espina Cepeda (1980) examined costs for Radio ECCA which, like ACPO and RSM, provides basic education through radio and print. He calculated that, in 1978, it cost 5 000 pesetas ($64) to teach a student through ECCA for a year, compared with a cost of 40 000 pesetas ($509) for a year of primary school education. The promotion rate at the end of each course was between 72% and 75% so that the cost per student promoted was between $83.11 and $88.89. His paper suggests that students took six months to work through each course so that, over a 2½ year period, they would complete the equiva- lent of a five-year primary curriculum. If we assume, therefore, that 23% of those starting will complete the course, then the cost per completing student is (100/23) x (64 x 2.5) = $696 while the cost of a primary school course at school is given as $2 545. 5.11 As table 9 shows, the evidence from radio schools is consistent in that their costs are of the same order of magnitude and that, where comparisons can be made, they are cheaper than the orthodox equivalent. Table 9: Costs of radio schools 1978 US $$ Radio school + Annual cost per student Annual cost per student Cost for full promoted course at radio at conventio- at radio at conventio- at radio at school nal school school nal school school conven- tional school ACPO Colombia 33.60 n/a n/a n/a n/a n/a 70 000 RSM Dominican 29.90 47.86 b/ 49.84 76.80 b/ n/a n/a Republic a/ 20 000 Radio ECCA 64 509 86 c/ n/a 696 2 545 Canary Islands 23 000 Source : text Notes: a. Omitting capital for both radio and conventional classes b. This is the figure for 40 000 students at conventional evening classes: a figure for 20 000 students would probably be higher. c. Taking the mean of the figures in the text. -113- 5.12 In contrast with radio schools, few costs have been published for farm forums. The notable exception is Schramm's (1967) report on Indian farm forums. He estimated the costs shown in Table 10. These omit capital costs for the radio station (and apparently for other buildings) but include costs for staff, training sessions, print, air-time and travel and half the costs of the radios used. They also assume that forums would expand from 500 to 2 000 over a period of 5 years. Table 10: Costs of Indian farm forums 1965/66 Currency : US $$ No. of forums 1st year 2nd year 3rd year 4th year 500 1000 1500 2000 1965/6$ 1978$ 1965/6$ 1978$ 1965/6$ 1978$ 1965/6$ 1978$ Total cost a/ 53 510 107 020 82 920 165 840 114 360 228 720 143 750 287 460 Annual cost per forum 107 214 83 166 76 152 72 144 Annual cost per member b/ 5.35 10.70 4.15 8.30 3.81 7.62 3.60 7.20 Cost per meeting c/ 2.67 5.34 2.08 4.16- 1.92 3.84 1.81 3.62 Cost per meeting per member 0.13 0.26 0.10 0.20 0.10 0.20 0.09 0.18 Note: a. Following Schramm we have taken Rupees 4.8 = $1.00 and then used American deflators (1965/6 $2.00 = 1978 $1.00) to convert to 1978 $$ b. Assuming 20 members per fornm c. Assuming 40 meetings a year 5.13 In contrast with these figuses the Poona forums had actual costs of $4.18 (1978 $9.40) per meeting - significantly higher than the costs here. At 20 members per forum and 40 meetings a year this would give an annual cost per member of $18.80. If we assume that Schramm's predictions for the first and second year, and the actual costs for Poona, give the most reliable guide, then we get an annual cost per participant of between $8.30 and $18.80 for farm forums in India. 5.14 Reports, which include statements about costs, have been published for five radio learning campaigns, three in Tanzania and one in Botswana. Un- fortunately the reports are not comparable in the level of detail they include, or in the way they analyse costs. Only the report on the 1976 Botswana campaign 'Our land' discusses indirect as well as direct costs, and all omit capital costs, other than those incurred for buying new equipment, such as radio sets, for campaigns. (As all campaigns used existing facilities for -114- a very short period of time, the effect of including capital costs would not have been great.) Furthermore, campaigns were on strikingly different scales, varying in the number of participants from 3 500 to two and a half million. Nevertheless, the evidence which is discussed more fully in Appen- dix A makes it possible to summarise the costs of campaigns as in Table 11. Table 11: Costs of radio campaigns in Botswana and Tanzania Currency : 1978 US$$ Campaign and No. of Total cost Total direct Direct cost per date participants (direct and cost participant indirect) Tanzania A time to rejoice 20 000 n/a 19 224 0.96 1971 Man is health 2 million n/a 405 180 0.20 1973 Food is life 2½ million n/a 810 857 0.32 1975 Botswana Our land 55 300 488 683 143 651 2.60 1976 Learning about 3 500 n/a 14 993 4.28 government 1979 Source Appendix A 5.25 The Botswana costs are significantly higher than those for Tanzania. Two reasons for this are evident. First, the later Tanzanian campaigns were on so much larger a scale that some economies of scale are to be expected. Second, the reports make it clear that Botswana's costs for research, eval- uation and consultation are far higher than those for Tanzania. A major reason for this is that the Botswana campaigns had, among their aims, to consult the public about central and local government policy. 5.16 Taking that into account, the figures for 'A time to rejoice' with 20 000 participants and 'Our land' with 55 300 are of the same order. If we omit the follow-up activity, which accounted for some $0.64 of the $2.60 per par- ticipant in 'Our land', this suggests one could run a campaign at a direct cost per participant of about $1 to $2. To allow for indirect costs, which in the Botswana case meant that the total cost was about 3.3 times the direct cost, the figure should be multiplied by 3.3, suggesting a total cost per participant in the range $3.30 to $6.60. We can then derive the cost function for the Botswana campaign 'Our land' -115- TC = AC(N) + F 488 683 = 6.13(55 300) + 149 451. 5.17 If these figures for smaller campaigns are right, then the costs for the larger campaigns are under-reported. It seems (cf. Appendix A) that the total cost per participant, including both direct and indirect costs, is of the order of $1.30 to $4.30. 5.18 All these figures omit the opportunity cost of the participants' time. Hall (1978, p.49) did, however, put the expenditure on campaigns into con- text by reference to the value of one of their products: 'The total number of person-hours that went into activities inspired by the 'Man is Health' campaign cannot be counted. Consider latrine- building alone. If the average latrine took 50 hours to construct, the estimated 750 000 latrines (based on district reports) built in Tanzania represented 37.5 million person-hours. If the Ministry of Communication and Works had paid workers one Tz shilling an hour to do the same, the cost would have been 37 500 000 Tz shillings. The campaign itself cost only 1 942 000 Tz shillings.* The gains from a single activity among the many pursued by Mtu ni Afya participants are staggering. *This figure represents "externally financed" campaign costs (that is, those not absorbed by the participating ministries and agencies).' Even if we assumed that the total cost of -'Man is health' was of the order of $2.00 per participant, rather than the externally financed figure of $0.20, this still gives us a favourable benefit:cost ratio. 5.19 We have little information about the costs of supporting other groups through mass media. In Botswana the costs of running courses which lasted several months for Village Development Committee members and members of 4B clubs (agricultural youth clubs) were estimated at R49.14 and R2.35 per participant respectively. (Perraton 1977 p.46) These figures are for 1973-76 and are approximately equivalent to 1978 US $85 and $4. The figures include overheads but exclude any element for capital. The numbers involved were very small - with between 1 500 and 3 000 taking part in the 4B clubs but only some 25 committees at a time following the course for village development committees. 5.20 Rather fuller figures, again for southern Africa, exist for the Lesotho Distance Teaching Centre. They experimented with teaching women techniques of crocheting and compared the costs of doing so by distributing booklets to individuals and encouraging existing women's groups to study the booklets together. The calculations in LDTC (1978), suggest that there were consider- able advantages in making printed materials available both to individuals and to groups, as this reduces the unit cost of the materials, while, in a -116- society with a high literacy rate, a significant number of individuals do learn simply from printed materials. Their figures, which include overheads but exclude capital, give us the results in Table 12. Table 12: Costs of Lesotho crochet project per participant 1977 Rand 1978 $ a/ Cost per individual recipient of a booklet (N = 10 800) 0.56 0.70 Cost of successful individual learner b/ 1.60 - 3.71 2.01 - 4.65 Cost per individual group member (N=1 500) 1.85 2.32 Cost per successful group learner 2.39 - 3.30 3.00 - 4.13 Source : LDTC (1978) Notes: a. Taking $1.15 = Rl and multiplying by 1.09 as the 1977 deflator b. Two criteria of success were used: 'learnt almost everything' and 'learnt at least something', giving us the two values. 5.21 The LDTC figures also make it possible to calculate a cost function for the project. Their fixed costs were $6 044. They had 10 800 individual students, to whom only printing costs are attributed, and 1 500 group mem- bers. We can therefore derive a cost function: TC = AC(N ) + AC(N ) + F S g where N is the number of individual students and 5 N is the number of group members. $11 271 = 0.15 (10 800) + 2.40 (1 500) + 6 044 B. Mass media for individuals with feedback 5.22 Distance teaching methods, in which print and/or broadcasts are used for teaching individuals with a feedback element, have been used on a small scale for basic education in the third world: there is far less evidence about the costs here than there is for the use of distance teaching aimed at groups of students. Figures are available from INADES-formation, which uses correspondence courses, supported by some face-to-face sessions to teach agriculture, and from Radio educative rurale (RER) in Senegal which made extensive use of feedback in planning and running radio programmes. Table 13 sets out the costs for INADES-formation in Cameroon and Rwanda. -117- Table 13: Cost of correspondence education through INADES-formation Currency: 1978/79 US $$ a/ Cameroon Rwanda Fixed costs 72 791 85 439 Variable costs 64 222 22 565 Total 137 013 108 004 No. of students 573 1 010 Cost per student b/ 239 107 Source: Jenkins and Perraton 1981 Notes: a. Although these figures are from 1978/79 accounts we have not deflated them and assume they are reasonably accurate for 1978. A 7.5% discount rate is used. b. The cost to INADES-formation is reduced by some $15 because of fees paid by students. 5.23 To the costs of correspondence teaching one may add the cost of attend- ing face-to-face sessions: the cost per person/day in Cameroon was $58.85. If we assume that each student should attend two such sessions each year, then we get annual costs of $357 in Cameroon and, assuming comparable face-to-face costs there, of $225 in Rwanda. We can calculate the cost functions for studying by correspondence in both cases: TC = AC(N) + F $137 013 = 112.08 (573) + 72 791 (Cameroon) $108 004 = 22.34 (1010) + 85 439 (Rwanda) Thus, while the fixed costs for Cameroon and Rwanda are very similar, there are striking differences in the variable costs. The figures also suggest that there may be potential economies if INADES-formation could increase its enrolment substantially. At present INADES-formation clearly does not expect to enrol students on the scale of the Latin American radio schools, although their experience suggests that such an expansion might perhaps be possible. 5.24 In contrast with INADES-formation's concentration on print, RER was essentially a radio project but with extensive staff time and travel used for collecting feedback. The costs of this are reflected in their production costs. The costs of production, transmission and reception for 1968-71 were calculated by Trabelsi (1972) and Table 14, for 1970, is based on his work. The cost includes estimates for the amortisation of capital. -118- Table 14: Costs of RER Senegal, 1970 Currency a! CFA 1978 US $ Transmission cost per hour b/ 5 900 39 Production cost per hour 355 200 2 326 Total 361 100 2 365 Cost per listener per hour c/ 0.451 0.0030 Reception cost per listener per hour 0.018 0.0001 Source: Trabelsi 1972 Notes: a. Taking 250 CFA = $1.00 and US deflator of 1.637 b. For 46 hours of broadcasts c. Assuming an audience of 800 000. C. Radio and Print 5.25 Where radio is used for education, its costs are of three kinds: for production, for transmission, and for reception. We have more estimates of the costs of production and transmission than of reception. And, where listeners already own radios, it is reasonable to include only the marginal cost of listening to educational broadcasts - often a figure so low that it will make little difference to calculations or to decision-making. Neverthe- less, the figures for transmission and production vary widely. Table 15 sets out some recent costs which are probably comparable and in which some attempt has been made to include capital as well as recurrent costs. Table 15: Estimated costs per hour of open broadcast Costs per hour Country Date of Production Transmission Production and informa- Current Constant Current Constant Transmission tion US $ 1978 US $ US $ 1978 US $ Current Constant US $ 1978 US $ Tanzania a/ 1979 464 408 124 109 588 517 Nepal b/ 1975 82 99 53 64 135 163 Mexico b/ 1972 115 179 14 22 129 201 Shetland Isles (Britain) c/ 1978 n/a n/a n/a n/a 427 427 Malawi d/ 1980 n/a n/a n/a n/a 1037 830 Notes: a. Mbunda and Edington (1979) using a 5% discount rate and taking 8.074/- = $1.00 b. Unesco (1977) using a 7.5% discount rate c. Postgate et al. (1979) d. -119- 5.26 Two recent studies enable us to look in more detail at the costs of broadcasts for farmers. Mbunda and Edington (1979) looked at the costs of school broadcasting but considered also the costs of broadcasting to farmers. Perraton, Jamison and Orivel (1978) looked at the cost of farm broadcasting in Malawi. We summarise their findings in Table 16. As, with open broad- casting, we can never be sure how large the audience is, figures for costs per listener are necessarily uncertain. But these figures suggest that, at least in East and Central Africa, production and transmission costs are likely to be in the range of $500 and $1 000 per hour and annual costs per listener between $0.10 and $0.75. Table 16: Costs of radio broadcasts in Malawi and Tanzania Currency: 1978 US $$ Malawi a/ Tanzania b/ Transmission cost per hour 806 c/ 113 d/ Production cost per hour 111 428 Total cost per hour 917 541 Cost per listener per hour e/ 0.0031 0.0005 Cost per listener per year f/ 0.740 0.1353 Reception cost per listener per year g/ n/a 2.0709 Notes: a. Taking $1.20 = K1.00 and a deflator of 0.8 to convert from 1980 Malawi Kwacha to 1978 $$, and using a 7.5% discount rate. b. Taking 8.074/- = $1.00 and a deflator of 0.88 to convert from 1979 Tanzania ahillings to 1978 $$. Mbunda and Edington use 5% and 10% discount rates: for the sake of comparability we have taken the mid-point between their 5% and 10% figures as giving a good enough approximation for a 7.5% rate. c. The actual discount rate used for calculating this figure is not available. d. These are actual costs; if the radio transmitter network were extended to cover the whole country, the cost might rise as much as three times. e. Assuming audiences of 300 000 in Malawi and 1 000 000 in Tanzania. f. For 232 hours of farm broadcasting ayear in Malawi and 250 in Tanzania. g. Assuming a 5-year life for a set and 10 listeners per set. 5.27 Where production costs are high, the costs of broadcasting can, of course, be much higher than this. Even excluding capital costs, the Manoff advertising campaigns appear to have cost $1.75 to $2.50 to reach each family in their target audience. (1978 $2.03 - 2.90) (Manoff 1977). -120- 5.28 Costs for television have not been calculated but such figures as are available suggest that these would be between ten and thirty times as high as for radio. Furthermore, television presents major problems of reception and maintenance, especially in the poorest and most remote areas. There is little experience of television for basic education or agricultural extension in developing countries where factors are favourable to its use - such as a compact agricultural area, with common linguistic and agricultural features, a high proportion of villages with mains electricity, and a television service already developed for non-educational purposes. 5.29 We have some information on the costs of producing educational magazines or papers, though none on the cost of village newspapers. As we saw above, (para. 5.7) with a print run of 70 000 the sixteen-page el campesino costs $0.21 per copy. In Malawi we found that the bi-monthly Za achikumbi, also of 16 pages and with a circulation to farmers of 30 000, costs K1.86 to K2.13 per farmer per year. This gives a cost per issue of US.$0.30 - 0.34 per copy. With a print run of only some 15 000, the Lesotho pamphlet on crocheting cost $0.70 per recipient, but some 26% of this cost was for dis- tributing the pamphlet, a figure omitted from the costs for el campesino and Za achikumbi. Where, as with the radio campaigns, print has been used to support broadcasts, the costs are often not disaggregated. D. Cost summary 5.30 The information on costs is summarised in Table 17. It is limited, and provokes difficult questions about the legitimacy of comparing costs across national and cultural frontiers. For some types of project, we are extremely short of even the sketchiest information: we have for example, very little information on the costs of radio farm forums, despite their existence in more than a dozen different countries. 5.31 The costs in Table 17 need to be seen, too, against the information available about the costs of radio broadcasting and the costs of print. The hourly cost of radio per listener, in particular, is so low that much of the expenditure on multi-media projects must be on elements other than radio. A high proportion of the costs appear to reflect administration, the develop- ment of teaching material, and the support of group learning. 5.32 Some further generalisations are possible. 5.33 The large-scale Tanzanian campaigns appear to have achieved costs which are very low as compared with other forms of multi-media education, or, indeed, with any other form of education. Assuming the figures are reasonably accurate, these low costs are explained by four factors: the campaigns were big enough to achieve major economies of scale; they made extensive use of unpaid group leaders; -121- printed materials were supplied on a group basis and not to every individual; government and party staff devoted time to setting up and supporting groups, as part of their regular duties, and the value of this time is not adequately reflected in the costs. Samller campaigns, with participants measured in tens of thousands rather than millions, achieved costs between two and ten times as high as large campaigns. Judging by the single example of the costs of farm forums, their costs, for an intensive campaign reaching a larger number of people, are comparable to those of a radio farm forum, reaching a smaller number but over a much longer period. 5.34 We are on surer ground with radio schools. Here, despite the range of students from 14 500 to 170 000, annual costs per student are within the range of $29 to $64. As their work in basic education can be compared with primary schooling, it is also possible to show that costs per student, and costs per student promoted, are lower for radio schools than for orthodox education. The scale at which the radio schools operate, their use of volunteer group leaders, their support from the church, and their extensive use of radio all help to produce these results. 5.35 The evidence from Lesotho shows that, where an organisation is running a number of different teaching activities, unit costs can be kept down for more modest projects, as with the crochet course. But attempts to teach a whole curriculum - as with INADES-formation or the radio schools - do demand a body of students probably in the tens of thousands rather than the thousands. The fixed costs, in Rwanda and Cameroon alike, are high in relation to the number of students, when compared with those for other multi-media projects, while the variable costs in Cameroon are also high. Multi-media teaching can reduce educational costs but it is not necessarily a recipe for doing so. 5.36 If we compare the costs for multi-media projects with those for ortho- dox literacy work, discussed in section III, two conclusions follow. First, it is possible to teach people either a broad curriculum, as with the radio schools, or narrower skills, as in Lesotho, at costs well below those often achieved for literacy projects. But, second, the extremely large literacy campaigns show economies of scale which appear otherwise only to have been achieved by the Tanzanian radio campaigns, themselves also on a very large scale. TABLE 17: Costs and cost functions of multi-media proiects Project Units for N at time Cost function Average cost Evidence on effects Evidence on compar- measuring N of study in 1978 $ for given N able costs in 1978 $ Radio schools ACPO students p.a 170 000 TC(N) = 15.94(170000) 29.16 35% success in liter- n/a + 2 247 770 a/ acy; some evidence of changes in attitudes and practices in health and agriculture RSM students p.a 20 000 TC(N) = 18.78(20000) 35.88 60% of students promo- Cost per student + 342 187 b/ ted annually; 57% of lower than cost students are women; in conventional moderate success in classes reaching poorest. ECCA students p.a 14 623 n/a 64 72-75% of students Cost per student promoted annually lower than cost in conventional I class Farm Forums India Membersp.a. ? -20000 n/a 8.30-18.80 Some evidence of learn- n/a ing. Government not clearly convinced of value of forums. Radio campaigns Our land Participants 55 300 TC(N) = 6.13(55300). 8.84 Women less well repre- n/a Botswana in campaign + 149 451 c/ sented than men in audi- ence. 40% participants with no cattle. Tanzanian large Participants 2 - 2.5 m n/a 1.30-4.30 Changes in health prac- Costs of projects campaigns in campaign tices. Evidence of undertaken by learning. peasants lower than current market rate 'A time to Participants 20 000 n/a 3.20 n/a n/a rejoice' in campaign Tanzania 'Learning about Participants 3 500 n/a 14.27 Some knowledge gains government' in campaign Botswana Other groups Crochet Participants 10 800 indi- TC(N) = 0.15(10800) 0.70 47% of individuals Teaching groups project - in campaign viduals + + 240(1500) + 6044 d/ (individual) and 96% of group may be more cost- Lesotho 1 500 group 2.32 members learned to effective than members (group crochet teaching indi- members) viduals. Courses for individuals INADES-formation Students p.a 573 TC(N) = 112.08(573) 239.12 Subjective evidence High variable Cameroon + 72 791 e/ of courses' effec- cost; high fixed tiveness cost in relation to number of students H INADES-formation Students p.a 1 010 TC(N) = 22.34(1010) 106.93 n/a n/a Rwanda + 85 439 e/ Source : text Notes: a. Using a 10% discount rate b. Including an estimated figure for capital c. Omitting capital d. Omitting capital e. Using a 7.5% discount rate -124- VI CONCLUSIONS A. Implications for policy 6.1 We look in turn at the general conclusions from this study, at its implications for basic education, for agricultural extension, and for costs. 6.2 General. Mass media have proved an effective way of teaching people who meet together in groups, in order to study material taught through broad- casts and print. The combination of mass media and group learning has been used for teaching the content of basic education, for teaching certain techniques, and for stimulating action within a community. The success of such methods depends on the existence of an organisation to support and en- courage learning groups. 6.3 Either radio or printed materials can be effective in communicating information to individual learners, who are not organised into groups. It is reasonable to assume that they are-less effective used in this way, but they are also much cheaper. Radio is a particularly cheap medium, in terms of cost per listener, in a number of developing countries, provided that the audience is large enough for the production and transmission costs to be spread widely. We know little about the limits to the effectiveness of mass media used in this way. 6.4 Under the right circumstances, mass media used with or without a group organisation, can reach a large proportion of adult populations, and can reach both women and other disadvantaged groups. Group organisations have, in Latin America, reached between 1% and 5% of rural adult populations each year. Radio broadcasts, even in countries where radio ownership is not yet universal, can reach 30% or more of target, rural, audiences. 6.5 Correspondence education, has provide an effective method of teaching both basic and specialised subject matter. It can be assumed that it is more effective when linked with radio. The major problems which arise, as with other forms of part-time education for adults, concern retention and dropout rates. 6.6 Basic education. Mass media have proved more successful at levels above the most basic, less successful at teaching literacy or numeracy, even with some face-to-face support, to adults who have never been at school. (In this regard they are probably comparable to orthodox forms of education.) The radio schools of Latin America, for example, have had more success at such levels. 6 .7 Of the separate components of basic education, mass media have been successful in developing skills in literacy and numeracy, in enabling people to benefit from change, and in increasing their efficiency in acquiring information. There is insuffucient evidence to show how far they achieve broader goals. -125- 6.8 Many mass media programmes have had relatively narrow aims, of teaching particular skills or methods, or about particular aspects of agriculture or health. A smaller number, mainly in Latin America, have attempted to go beyond this and teach a curriculum of basic education. We can infer, from both sets of experience, that as the methods can be effective for narrow objectives, so there is no reason to suppose they will not be effective for broader ones. What is lacking - for the radio campaigns of Tanzania for example - is the idea of a curriculum, or the intention to relate together different elements of adult education so that the whole is greater than the sum of its parts. 6.9 Agricultural extension. Mass media, and especially radio, have demon- strated in a wide variety of countries, their ability to carry information to farmers about agriculture. Radio is often seen, by farmers, as an im- portant source of agricultural information. It is also a very cheap way of communicating with farmers. Correspondence courses can teach farmers and it appears that farmers can introduce changes into their practice as a result of following such courses. Mobile cinema vans in at least two African countries appear to be successful in providing information to farmers. Where distance teaching techniques have been used, feedback from farmers to extension or education agencies is possible. 6.10 There is an unrealised potential in using mass media to support agricul- tural extension agents, and to train them in-service. Oddly, while this use of radio and print is so widespread for upgrading primary school teachers as to be almost the norm in many developing countries in Africa, it has been little used to upgrade extension agents. The experience of INADES-formation suggests that this may be an important future role for the mass media. 6.11 Generally, the evidence about the effectiveness, and about the costs, of using mass media for communicating agricultural information is such that its increased use may be of major importance as extension grapples with the problems of budget limits and of scaling up its level of activity. 6.12 Costs. The use of mass media, unsupported by a group activity in the field, is very cheap, especially when compared with the costs of face-to-face contact. Comparisons of effectiveness are more difficult. Existing data does not permit us to make a comparison between the effectiveness of, say, listening to a radio programme for an hour and talking with an extension agent for an hour. But evidence from Malawi suggests that the costs are in the ratio of 3000:1. Thus the cost for each contact between an agricultural extension agent and a farmer was between (1978) $15 and $21; in contrast the cost per farmer of attending a puppet play was about $0.08, of watching an -126- agricultural film for an hour was $0.17, and of listening to an hour's agricultural radio programmes was $0.004. Costs obviously vary from place to place, but it seems likely that they are often of this order of modesty. Similarly, where the scale is large enough, printed materials can be very cheap: the 16-page Colombian magazine el campesino costs only $0.21. 6.13 The costs of teaching individuals with correspondence courses, either with or without radio support, depend so much on the numbers involved and the overall costs of the agencies running those courses, that few generalisa- tions are possible. There are circumstances under which such teaching can be cheaper than alternatives. If it increases the effectiveness of,say, ex- tension agents, the potential benefits may be very large indeed. 6.14 Where a group organisation exists to support learners, costs are, of course, much higher than they are for mass media alone. Radio campaigns, reaching audiences of about 2 million, for a period of two to three months, probably had costs of between $1 and $5 per participant. Radio schools in Latin America, with audiences of between 20 000 and 200 000, can teach the equivalent of a primary school curriculum for between $30 and $65 a year. (In both cases the opportunity costs of the participants' time is not included.) While these audiences are far smaller than those for the Tanzanian campaigns, they are nevertheless significant. If a radio school, or something like it, recruits 2k% of the rural adult population in a year, then, over twenty years it may have taught a quarter or a third of all the adults in a country. The costs for such an achievement can be modest. If the annual cost is of the order of $50 per participant, then one could reach 1I% of the population each year for $0.625 per head of the population. Expressed as a percentage of GNP per head, this would mean the devotion of the following resources: GNP of $180 (e.g. India) 0.35% $400 (e.g. Ghana) 0.16% $600 (e.g. Papua New Guinea) 0.10% 6.15 The case for educating adults cannot rest on that sort of arithmetic, involving, as it does, huge assumptions and uncertainties. But the evidence suggests that a case can sometimes be made for investment in well-designed multi-media educational projects for adults. B. Implications for research 6.16 In assessing the value of mass media, we are hampered by the shortage of comparable data on the effectiveness of orthodox methods of education and especially extnesion. Despite the level of government, and international, investment in agricultural extension services, there is little data which -127- shows how valuable extension services are in raising productivity. Along with further research and experiment on the use of mass media, which is discussed below, we therefore need more on orthodox extension,and its costs and effects. 6.17 Considerable further research is needed about the ways in which mass media, and especially radio, can best support the work of agricultural extension. The need is not for further general surveys, but for case studies and field experiments. One part of this work should be a study of farm forums, which have existed for over 40 years, despite wide scepticism about their value, and of which there is apparently only a single cost study. 6.18 There is scope for a guide to practice in the use of mass media for rural education. Existing documents, such as The educational use of mass media (World Bank Staff Working Paper No. 491), do not meet this demand. 6.19 This paper has concentrated on basic education and agriculture. There is considerable, but scattered, third world experience on the use of similar methods for health, nutrition and family planning and a similar survey could usefully be undertaken. 6.20 If basic education for adults is to be effective, it needs a curriculum, based on their interests and experience, rather than being a shadow of children's primary school curriculum. Further research is needed on the ways of devising and implementing such a curriculum. -128- APPENDIX A : MASS MEDIA PROJECTS Al The literature on mass media projects varies in the detail with which it describes them. In this Appendix we report evidence which is important for the argument of the paper but at too great a level of detail to be included in its main body. A. Radio school students A2 The literature on radio schools provides us with some detailed informa- tion about their students. Accounts of the work of ACPO in Colombia, ACPH in Honduras and Radio Santa Maria in the Dominican Republic show the number of students over a period of years, as in table Al. Table Al: Student enrolments at 3 radiophonic schools Year ACPO ACPH RSM Colombia Honduras Dominican Republic 1948 45 a/ not yet not yet establsihed established 1949 1 339 " if " " of 1950 3 636 it " " it 1951 5 436 " " " I " " 1952 10 848 " " " " " " 1953 15 648 " " " " i t 1954 20 409 " " " " " " 1955 82 319 .. .. It .. it 1956 109 989 " " " " " " 1957 136 385 " " " " " " 1958 145 248 " it " I " I " I It 1959 112 576 " " " " " " 1960 114 186 " " " " " " 1961 129 681 306 d/ " " " 1962 215 309 7 520 " " " 1963 n/a 11 298 " " " 1964 n/a 14 624 n/a 1965 n/a 8 261 n/a 1966 n/a 9 745 n/a 1967 n/a 11 548 n/a 1968 n/a 16 564 n/a 1969 n/a 13 595 n/a 1970 n/a 14 254 n/a 1971 169 696 b/ n/a 12 238 e/ 1972 n/a n/a 22 375 1973 167 451 c/ n/a 20 171 Notes: a. ACPO figures for 1948-62 are from Primrose (1965) p.l59 b. Brumberg (1975) p.8 c. McAnany (1975) p.242 d. ACPH figures are from White (1976a) p.38 e. RSM figures are from White (1976b) pp.52-3 -129- A3 Information from the Dominican Republic and from Bolivia gives us more detail about the students of Radio Santa Maria and of ERBOL. For RSM, White looked at the amount of land owned by the fathers of RSM students, and then compared this with the amount held in farm units in the region generally, with the results shown in table A2. This showed that RSM reached some of the poorer families but also enabled him to reach the conclusion reported that RSM students 'have a somewhat higher socio-economic status within rural communities (though not within the larger national social status system)' (White 1976a, p.25). Table A2: Land ownership and RSM students RSM student Regional figures for % farm units Father owns no land 21.6 Father owns land 78.4 100 Amount held less than 0.5 hectares 22.2 20.0 0.5 - 4.9 hectares 40.7 56.1 more than 5 hectares 37.1 23.9 100 100 Source White 1976a, pp.18-25 A4 From Bolivia comes evidence, referred to in the text, showing that radio schools can reach the poorer members of communities, such as the non- Spanish speakers. Table A3 shows some details of two groups of ERBOL students in 1976. Table A3: Characteristics of ERBOL students, 1976 Literacy students 2nd and 3rd year students Sex Male 42.0 48.6 Female 58.0 51.4 Occupation Farmer 35.8 42.6 Housewife 45.7 39.8 Shopkeeper or trader 7.0 5.6 Other or not stated 11.5 12.0 Language Spanish 33.7 35.6 Quechua 32.9 47.7 Aymara 30.5 15.7 Other or not stated 2.9 1.0 Source : ERBOL - ACLO 1977 -130- B. Effects of farm forums A5 There is little firms researched, evidence on the effects of radio farm forums. The single study on forums in Ghana is therefore important, particularly as it compared villages with forums with others supplied just with a radio and others with neither forum nor radio. The results of a study, completed after the farm forum had been operating for five months, which show differences between the forum villages and the controls, are set out in Table A4. These suggest the effectiveness of forums although the report did not test the results for their statistical significance. Table A4: Effects of radio farm forums in Ghana Measure Percentage of respondence reporting positively Villages with Villages with Villages with Villages with 1 forum 2 forums no forum but no forum and (no=89) (no=82) with radio no radio supplied supplied Evidence of activity (n=84) (no=83) Action reported to in- crease agricultural productivity 60 53 42 35 Measures taken to im- prove marketing 25 23 21 9 Action or plans to im- prove harvesting and processing crops 27 24 8 13 Joined a co-operative 16 19 5 6 Evidence of knowledge or attitude a/ Knowledge of government policy on import controls 25 30 14 6 Knowledge of government loan schemes 40 33 31 28 Favouring school sup- port for preservation of Ghanaian culture 21 22 12 10 Supporting government policy on self-help projects 81 78 81 62 Use of protein in babies' diets 51 46 41 36 Claim to be saving for emergencies 70 60 57 50 Belief in joint husband-wife decision making 55 56 51 47 Attitudes to citizenship no significant difference a/ Source : Abell (1968) pp.36-40 Note: a. Certain questions on attitude to government omitted because of difficulty in interpretting them. -131- C. Effects of radio campaigns A6 The Tanzanian campaign 'Man is health' (Mtu ni afva) was evaluated with an assessment of its effect on participants' health practices. Hall investi- gated changes in health practices, and the building of latrines, which could be attributed to the campaign, in a sample eight villages, with the results set out in tables A5 and A6. Table A5: Changes in health practices in sample eight villages - Tanzania 1973 Health practice Number of Number of Percentage houses, pre- houses, post change campaign campaign 1. Elimination of. vegetation growing near house 286 1 916 +570 2. Elimination of stagnant water near house 357 548 + 54 3. Mosquito netting in bedroom(s) 136 147 + 8 4. Mosquito netting on windows 84 111 + 32 5. Latrine meeting Mtu ni Afya standard 494 335 - 32* 6. Cover on latrine 328 685 +109 7. Latrine in use 421 939 +123 8. Elimination of rubbish around house 421 1 248 +182 9. Elimination of animal feces near house 1-399 1 223 - 13 10. Absence of rats or other vermin in or around house 503 773 + 46 11. Windows in house 245 375 + 53 12. Absence of "a lot" of flies in and around house 818 1 572 + 92 * Because the criteria for latrines were relaxed at the time of the post-survey, some latrines counted in the pre-survey were, apparently, later not thought to be of standard quality. Source : Hall (1978), p.58 Table A6: Latrines built or rebuilt in sample eight villages during campaign - Tanzania 1973 Region Number of Number of new or Percentage of houses with houses rebuilt latrines new or rebuilt latrines* Dodoma 1 720 558 32 Coast 364 143 39 Total 2 084 701 34 * Only latrines that met campaign standards were counted. Source : Hall (1978), p.59 -132- D. The cost of radio schools A7 The fullest accounts of the cost of radio schools are for ACPO. Brumberg (1975) and Tuckman and Nias (1980) provide figures from which it is possible to calculate costs, including costs for capital, and produce a cost function for ACPO. Brumberg showed that the running costs for ACPO were as in Table A7. Table A7: ACPO costs million constant (thousands) pesos 1972 1978 $$ a/ radio production 12 948 cultural/teaching 14 1 106 printing and publishing 20 1 580 newspaper 7 553 regional activities 15 1 185 administration and other charges 18 1 422 86 6-794 Source: Brumberg (1975) p.51 Note: a. We have multiplied the 1978 figure by 1.58 and divided by 20 for this conversaion. A8 His figures do not include an estimate for the cost of buildings and equipment, but Tuckman and Nias (1980), p.159 suggested that, at a 10% dis- count rate, the annualised cost in 1976 was $1 307 067. We can assume that the costs of buildings and equipment in 1972 were of the same order of mag- nitude in 1976 as in 1972. Inflating the 1976 figure by 1.16 we therefore have, in 1978 $$: '000 $$ Running costs 6 794 Capital costs a/ 1 516.2 9 310.2 A9 Brumberg found that 61% of ACPO's expenditure was on education, as opposed to their commercial activity, so that their total expenditure on education was (61/100 x 9 310) or $5 679 200. AIO We can also distinguish between expenditure on ACPO's newspaper el campesino and on its teaching activities. El campesino cost 8.1% of the total running costs. Assuming that 8.1% of the capital costs can properly be attributed to it, we get the following allocation: -133- 1978 $$ Cost of el campesino (8.1/100 x 9 310 200) 754 130 Other educational costs 4 925 070 Total educational expenditure 5 679 200 All At the time of his study, Brumberg found ACPO was teaching some 170 000 students a year so that the basic annual cost per student in 1978 $$ was 4 925 070 or $28.97. 170 000 El campesino had a press run of 70 000 a week so that the cost per copy was 754 130 or $0.21. 70 000 x 52 A12 Students on ACPO's complementary course (25 500 students) used a monthly supplement of el campesino as their basic text; this amounts to 8 pp, and el campesino is normally 16 pp in length. We should therefore add to the cost of teaching students the value of (12 x 25 500 x i) copies of the news- paper or $32 130 which would raise the average cost per student to 4 925 070 or $29.16. 170 000 A13 We can therefore calculate the following cost function for ACPO. The total cost for ACPO's students is: 4 925 070 + 32 130 or $4 957 200. We can regard as fixed the costs of capital, radio production and administra- tion. Taking into account the allocation of 39% of costs to ACPO's non- educational activities and a further 8.1% to publishing el campesino we there- fore have as fixed costs: 52.9% of capital costs 902 070 61% of radio production 578 280 61% administration etc. 867 420 $2 247 770 This gives us a variable cost of $2 709 430 and a cost function TC = AC(N) + F $4 957 200 = 15.94 (170 000) + 2 247 770. A14 White (1976a) gave some comparable figures for RSM, including details of the budget for 1975, which are set out in table A8. -134- Table A8: RSM costs 1975 $ 1978 $ Central administration, teaching and records 95 550 117 527 Depreciation 7 500 9 225 Building, lights, telephone 8 400 10 332 Broadcasting 64 700 79 581 Preparation of masters for printing 4 800 5 904 Printing - supplies and distribution 73 100 89 913 Field teachers and supervisors 183 600 225 828 Public relations 25 000 30 750 Office costs 23 600 29 028 486 250 598 088 Source: R White (1976a) table 9.2 If we regard the first five items in table A8 as fixed and add to them a notional 20% to allow for capital costs we get, in 1978 $$: Fixed costs : capital (20% of 598 088) 119 618 recurrent 222 569 $342 187 Variable costs: 375 519 The cost function is then: TC = AC(N) + F $717 706 = 18.78 (20 000) + 342 187 A15 White went on to compare RSM costs with those for conventional classes. He looked at the costs both for the 20 000 students which RSM had and for a projected 40 000. Figures for conventional classes were available only for 40 000 students. The comparison is shown in table A9. Table A9: RSM cost comparisons Enrolment 20 000 40 000 1975 $$ 1978 $$ 1975 $$ 1978 $$ Cost per student enrolled conventional classes n/a 38.91 47.86 RSM 24.31 29.90 19.96 24.55 Cost per student promoted conventional classes n/a 62.44 76.80 RSM 40.52 49.84 33.27 40.92 Source: R White (1976a) Table 9.2 -135- E. The cost of radio campaigns A16 The extent of the difference between reports of the radio campaigns Tanzania and Botswana makes it necessary to discuss their costs in more detail. As we saw, reports, which include some cost information, have been published for five campaigns. Hall and Dodds (1974) produced an account of Tanzania's campaigns up to 1973 and included costs of 'A time to rejoice'. Hall (1978) then produced a separate study of the 1973 health campaign 'Man is health' with figures for the externally financed costs of that campaign which claimed an audience of two million. Meanwhile Grenholm (1975) had described the same campaigns and published the budget for the next campaign, run in 1975, 'Food is life' on nutrition and food production. So we have figures for three Tanzanian campaigns, two large and one small. They have in common that they include direct but not indirect costs, omitting, for example, the cost of staff time. A17 The 1976 Botswana campaign on land policy 'Our land' was described in detail by Ministry of Local Government and Lands (1977). Unusually the report describes costs in detail, and separates direct from indirect costs. This information is amplified in a technical note by the Evaluation Unit, Botswana Extension College (1977). Costs for the 1977 regional campaign of political education, 'Learning about government' organised in the western Kalahari were included in the report by Boipelego Education Project (n.d.? 1980). A18 The fullest analysis of costs appears in the account of 'Our land'. MLGL (1977) para 5.31 says in presenting the costs: 'The total cost of the campaign between 1st November 1975 and 31st March 1977 is estimated at P416 000 (US $483 000). This can be divided into direct and indirect costs. Direct costs are costs which arose expressly because of the campaign and would not otherwise have occurred. Indirect costs include the costs of Botswana's resources which were diverted from other activities to the campaign. An RLG leader training course, for example, would have the direct costs of food, accommodation and transport for the leaders, and the indirect costs of extension workers' time spent organising and running the course.' Etherington expands on this (Evaluation Unit, BEC 1977, p.2) to point out that: no estimation of rent or its opportunity cost (what a government department's buildings could be leased for) are included no allowance is made for the tax paid on salaries (and the net cost to Government of its officers working on the campaign is their income after tax) the costs cover only the seventeen month period between 1 November 1975 and 31 March 1977; thus they exclude, for example, the tour by HE the President and the Cabinet before the RLG campaign and the recent report writing period -136- - the opportunity costs of the time the estimated 3 200 RLG leaders and 52 100 RLG members spent in attending meetings for 532 900 per hours are also excluded.' They also exclude the supplementation and gratuity for expatriate staff working on the campaign, so that the staff costs are those that would have been met had the campaign staff been entirely localised - as, in the future, one might expect them to be. A19 With those exceptions, we have clear and detailed figures for the costs of the campaign, as shown in Table A10. The direct costs proved to be only 38.6% of the total. The figures enable us to produce a cost function for this campaign. If we regard as fixed the items for co-ordination, evaluation, senior staff support and overheads to senior staff we get a total fixed cost of P110 800. Omitting the cost of radios, for reasons discussed in the next paragraph, we then have variable costs of P251 500. (The costs of 1978 $$ are $149 451 and $339 233 respectively.) With 55 300 participants in the campaign, we then have a cost function: TC = AC(N) + F $488 684 = 6.13 (55 300) + 149 451. This function omits any allowance for capital costs. Table A10: Direct and indirect costs of 'Our land' campaign Direct costs Pula Percentage of total costs Campaign coordination 9 000 2.2 Fieldwork and training 19 000 4.7 Teaching materials 45 500 10.9 Hardware (radio and tape recorders) 54 100 13.0 Consultation 17 300 4.2 Evaluation 8 800 2.1 Transport 5 000 1.2 Postage (official free) 1 200 0.3 Total direct costs 160 600 38.6 Indirect costs Field support (extension workers' time, etc.) 137 700 33.1 Senior staff support 41 600 10.0 Overheads to senior staff 51 400 12.3 Radio air-time 4 700 1.1 Transport 20 400 4.9 Total indirect costs 255 800 61.4 TOTAL CAMPAIGN COSTS 416 400 100.0 Source: MLGL (1977) para. 5.31 A20 If we want to compare these figures with those of any other campaign we are faced with three difficulties. First, in the absence of a similar breakdown between direct and indirect costs for any other campaign, we do not know how far the Botswana experience is typical. Second, the Botswana TABLE All:The costs of five campaigns 1 2 3 4 5 6 7 8 Title A time to Man is Food is Our land Learning about rejoice health life government Country Tanzania Tanzania Tanzania Botswana Botswana Date 1971 1973 1975(budget) 1976 1979 Currency 000/- 000/- 000/- P 000 p No. of people 20 000 2 m 2.5 m 55 300 3 500 Direct costs Total costs Direct costs Total costs only omitting omitting equipment equipment 7. 7. 7. 7. 7.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Coordination, planning, staff - - - 9 5.6 102 24.5 9 8.5 102 28.2 - Fieldwork and training 13.5 16.5 463 23.9 1 138 24.5 19.7 12.2 157.4 37.8 19.7 18.5 157.4 43.4 5 204 37.6 Materials, radio, publicity 54.8 17.1 1204.5 62.1 3 150.8 67.8 45.5 28.2 50.2 12.0 45.5 42.7 50.2 13.9 5 211 37.7 Distribution, tra- vel, transport 11.7 14.3 257.6 13.2 335 a/ 7.2 5 3.1 25.4 6.1 5 4.7 25.4 7.0 535 3.9 Hardware - - - 54.1 33.7 54.1 13.0 omitted omitted - Research, evalu- ation,follow-up 1.7 2.1 16.5 0.8 15 0.3 26.1 16.3 26.1 6.3 26.1 24.5 26.1 7.2 2 695 19.5 Other - - 8.8 0.2 1.2 0.7 1.2 0.3 1.2 1.1 1.2 0.3 182 1.3 TOTAL 81.7 100 1941.6 100 4 647.6 100 160.6 100 416.4 100 106.5 100 362.3 100 13827 100 Estimated cost in 1978 $$ 19 224 405 180 810 857 216 623 561 656 143 651 488 683 14993 Cost per partici- pant in 1978 $$ 0.96 0.20 0.32 3.92 10.60 2.60 8.84 4.28 Note: Source : see text a. Some transport costs are included in the figure for materials in this campaign. -138- costs include a figure of P54 100 for radios and cassette tape recorders which were supplied to group leaders or used for group training. This rep- resented 13% of the total costs of the campaign. But other campaigns assumed that groups had access to a radio and did not provide them so that, to com- pare like with like, we need to omit the costs of the radios and tape recorders from the Botswana figures. The effect of this is to reduce the direct costs of the Botswana campaign to 29.4% of the total. Third, the costs are set out in slightly different forms for different campaigns: even where they are similar, we cannot assume that different researchers would put similar items under the same heading. A21 But despite these reservations, there is enough similarity between the reports for us to be able to set out the figures, which appear in Table All. Here we show, for 'Our land', both direct and indirect costs, and, in columns 6 and 7, examine the effect of omitting the expenditure on equipment. At the foot of the table, total costs and costs per participant have.been converted into constant 1978 $$. It is reasonable to assume that, for all campaigns except 'Our land', the costs shown are just the direct costs and omit most, if not all, the indirect costs. In order to facilitate comparisons between campaigns, Table A12 shows the direct cost per participant in 1978 $$, omitting costs of equipment. Table A12: Direct costs of campaigns per participant (omitting equipment) Currency : 1978 US $$ A time to Man is Food is Our Learning rejoice health Life Land about government Coordination, planning staff - - - 0.22 - Fieldwork and training 0.16 0.05 0.08 0.48 1.61 Materials, radio, publicity 0.64 0.12 0.22 1.11 1.61 Distribution, travel, transport 0.14 0.03 0.02 0.12 0.17 Research, evaluation, follow-up 0.02 - - 0.64 0.83 Other - - - 0.03 0.06 TOTAL 0.96 0.20 0.32 2.60 4.28 Source : see text A22 Three conclusions follow. First, the Botswana costs for research, evaluation and consultation are far higher than those for Tanzania. This fits with the descriptions of the campaigns whose aims included consulting the public about central and local government policy. Costs of this kind are incurred if there is a deliberate choice to spend money on such follow-up work. A23 Second, if we assume from the Botswana experience that direct costs account for less than a third of the total cost of a campaign, then the five campaigns fall into two groups, with significant economies of scale in the large campaigns. -139- A24 But, third, we are unwise to take the figures at their face value. We know that some of the Tanzania figures are of externally financed costs only, and the contrast between the Eigures for different campaigns casts doubt on some of the very low estimates. If, however, we look at the costs for 'A time to rejoice' and 'Our land', both within the range of 20 000 to 55 000 participants, we find reasonable consistency. The figures suggest that, for a small campaign like this, one might expect the key costs per participant to fall in the range: $ fieldwork and training 0.16 - 0.48 materials, radio, publicity 0.64 - 1.11 travel 0.12 - 0.14 0.92 - 1.73 The costs for 'Learning about government' are-higher than this, but one would expect this as that campaign had only 3 500 participants. (As some materials had to be prepared for fieldwork and training, one would expect economies of scale even under this heading as one moved from 3 500 to 20 000.) Thus, if we can assume that co-ordination is usually hidden among the indirect costs, we can suggest that the direct cost per participant of a small campaign is in the range of about $1.00 - 2.00. To this must be added a figure for any evaluation and follow-up. Assuming that the Botswana experience is typical, we need to multiply this by 3.33 in order to allow for indirect costs, giving a total cost per participant in the range of about $3.00 - 5.85. A25 If these figures are right, then the figures for the large Tanzanian campaigns seem to under-report expenditure on fieldwork and training, and on travel, where one would expect few economies of scale in moving from an audience in the tens of thousands to one in millions. On the other hand, one would expect a reduction in the unit cost for teaching materials and radio, though even there the Tanzanian figures may be on the low side. To make a new estimate of the cost of a large campaign, we may therefore apply the figures from smaller campaigns for fieldwork and training and for travel, but take the figures for the larger campaigns for the cost of materials, radio and publicity - items that were externally financed for the large Tanzanian campaigns. If we do this, and allow for any under-reporting of the materials costs by multiplying these by a factor of 3 for our higher estimate, then we get the following range of costs for large campaigns: $ fieldwork and training 0.16 - 0.48 materials, radio, publicity 0.12 - 0.66 travel 0.12 - 0.14 0.40 - 1.28 -140- Multiplying this by the same factor of 3.33 to allow for indirect costs we get a total cost per participant in the range of about $1.30 - 4.30. -141- REFERENCES Helen C Abell (1968) 'Assessment of the project' in intr W F Coleman and A A Opoku (1968) An African experiment in radio forums for rural development (Unesco, Paris) 'Abolition of adult literacy in the USSR 1917-40' Fundamental and adult education C A Amtmann M., F Fernandez M. and J. Moraga R., (1977) La tecnica de "radio-vision" en la capatacion campesina (Centro de sociologia del desarrollo rural, Universidad Austral de Chile, Valdivia) N A Ansari and K B Rege (1964) 'Per capita cost of adult literacy in India' in ed. A Bordia, J R Kidd and J A Draper (1973) Adult education in India (Nackiketa, Bombay) George H Axinn and Sudhakar Thorat (1972) Modernizing world agriculture (Praeger, New York) C E Beeby (1966) The quality of education in developing countries (Harvard University Press, Cambridge Mass.) L R Beltran (1974) 'Communication research in Latin America: the blindfolded enquiry' (Conference paper - Leipzig 17-20 September 1974) L R Beltran (1976) 'Social structure and rural development communication in Latin America: the "radiophonic schools" of Colombia' in ed. G C Chu, S A Rahim and D L Kincaid (1976) Communication for group transformation in development (East West Center Honolulu) D Benor and J Q Harrison (1977) Agricultural extension: the training and visit system (World Bank, Washington DC) H S Bhola (1980)'Reading materials for the new reading public: a policy brief' Literacy Review 1, pp.1-45 M Blaug (1966) 'Literacy and economic development' School Review 74 pp.393-418 BEC Evaluation Unit (1978) Re Botseng (Botswana Extension College, Gaborone) (mimeo) BEC Evaluation Unit (1975) The growing fruit trees campaign for 4B club members (unpublished report, Botswana Extension College, Gaborone) Boipelego Education Project (n.d. ? 1980) A radio learning group campaign in western Botswana (Ministry of Education, n.p. ? Gaborone) M J Bowman 'Education and economic growth: an overview' - in ed. T King Education and income (World Bank 1980) D D Brown Agricultural development in India's districts (Harvard University Press, Cambridge, Mass. 1971) S F Brumberg (1975) 'Colombia: a multi-media rural education program' in M Ahmed and P Coombe (1975) Education for rural development (Praeger New York) J Bruner, R Olver and P Greenfield Studies in cognitive growth (Wiley 1966) -142- Henry Cassirer (1973) Mass media in an African context (Unesco, Paris) R Chambers (1974) Managing rural development: Ideas and experience from East Africa (Scandanavian Institute of African Studies, Uppsala) D P Chaudri (1979) Education, innovations and agricultural development (Croom Helm, London) M Colclough and D Crowley (1974) Setshaba la togamaano: the people and the plan (Division of Extramural Services, University of Botswana, Lesotho and Swaziland, Gaborone) M Cole, J Gay, J A Glick and D W Sharp (1971) The cultural context of learning and thinking, (Methuen, London) Philip H Coombs (1980) Meeting the basic needs of the rural poor (Pergamon, New York) P Coombs with M Ahmed, Attacking rural poverty (Johns Hopkins University Press, Baltimore, 1974) The development of the rural press 'Literacy work 7:4, 1978 pp.29-30 D D Dicko (1967) La pratique de la communication en milieu rural (Rapport du seminaire) (Agence de co-operation, Paris) ed. D Dan Dicko (1977) Rapport du seminaire: la pratique de la communication en milieu rural (Paris, 1977) Kiranmani A Dikshit et al. (1979) Rural radio programme formats (Unesco, Paris) Clifford Donkor (1979) 'The rural-radio forum' in Dikshit et al. (1979) A B Edington et al. (1974) New media in education in the Commonwealth (Commonwealth Secretariat, London) ERBOL/ACLO (1977) Educacion Radiofonica en Bolivia (Escuelas Radiofonicas de Bolivia - Accion Cultural Loyola, Sucre) Luis Espina Cepeda (1980) 'Relacion coste-eficacia en Radio ECCA' (Conference paper, Buitrago 28 January 1980) mimeo Evaluation Unit, Botswana Extension College (1977) Technical note no. 4: The costs of the campaign (Botswana Extension College, Gaborone) P J Foster (1971) 'Problems of literacy in sub-Saharan Africa' in T A Seboek et.al (1971) Current trends in linguistics vol. 7 pp.588-617 D Giltrow and J Potts (1978) Agricultural communication: the role of the media in extension and training (British Council, London) Ely D Gomez (1976) 'Radio programs and their impact on multiple cropping: a preliminary evaluation' Agricultural economics and development 6:2 Sven Grabe (1975) 'Tanzania: an educational program for co-operatives' in ed. Manzoo Ahmed and Philip H Coombs (1979) Education for rural development: case studies for planners (Praeger, New. York) L G Grenholm (1975) Radio study campaigns in the United Republic of Tanzania (International Bureau of Education/Unesco Press, Paris) Jonathan Gunter and James Theroux (1977) 'Developing mass audiences for educational broadcasting: two approaches' Prospects 7:2 -143- B L Hall (1978) Mtu ni afya: Tanzania's health campaign (Clearinghouse on Development Communication, Washington DC) B L Hall (1972) Wakati wa Furaha: an evaluation of a radio study group campaign (Institute of Adult Education, Dar es Salaam) B L Hall and Tony Dodds (1974) Voices for development: the Tanzanian national radio study campaigns (International Extension College, Cambridge) ed. Charles T Hein and Keith K Kanyogonya Rural press for village post- literature (Afrolit Society, Nairobi, 1979) INADES (n.d.) Qu'est-ce que inades-formation? (INADES-formation, Abidjan) A Inkles and D H Smith (1974) Becoming modern: individual change in six developing countries (Harvard University Press, Cambridge, Mass.) J Jenkins and H Perraton (1981) Training farmers by correspondence in Cameroon (International Extension College, Cambridge, forthcoming) (mimeo) F Kahnert 'Radio Togo's educational programme' in ed. W Schramm New educa- tional media in action: case studies for planners vol. 2 (Unesco/IIEP, Paris) Elihu Katz and George Wedell (1978) Broadcasting in the third world (Macmillan, Luton) A W Khan (1977) 'All India Radio's non-formal education broadcasts for rural development' (International Extension College seminar paper) ed. Sufia Khanam (1976/77) An evaluation study on effectiveness of radio as a means of communication in rural Bangladesh (Radio Bangladesh) Kenneth King (1979) 'Research on literacy and work among the rural poor' Convergence 12:3 N Kuhanga (1976-7) 'Adult education and development: a Tanzanian case study' Literacy Discussion Winter 1976-77 R LeFranc (1967) 'Radio clubs in Niger' in ed. W Schramm (1967) New educational media in action: case studies for planners vol. 3 (Unesco/IIEP, Paris) U Lele (1975) The design of rural development: Lessons from Africa (Johns Hopkins University Press, Baltimore) D K Leonard (1977) Reaching the peasant farmer (University of Chicago Press, Chicago) Lesotho Distance Teaching Centre (1978) Learning from a booklet (LDTC, Maseru) B G Lever (1970) Agricultural extension in Botswana (Department of Agricultural Economics, University of Reading, Reading) M E Lockheed, D T Jamison and L J Lau (1978) Farmer education and farm effic- iency: a review of the literature (World Bank, Washington DC) (mimeo) C P Loomis and J A Beegle (1957) Rural sociology: the strategy of change (Prentice-Hall, Englewood Cliffs) E G McAnany (1972a) Rural radio in Dahomey (mimeo) E G McAnany (1972b) Radio clubs of Niger (mimeo) -144- E C McAnany (1975) 'Radio schools in nonformal education: an evaluation per- spective in T J La Belle (1975) Educational alternatives in Latin America (UCLA, Los Angeles) Sean MacBride et al. (1980) Many voices, one world (Unesco, Paris) L Maliyamkono, A Ishami, S Klees and S Wells (1978) Health education for development: a case study of Tanzania (EDUTEL, Palo Alto) Richard K Manoff et al (1977) Radio nutrition education - using the advert- ising technique to reach rural families: Philippines and Nicaragua (Manoff International, Washington DC) J Matiko (1976) 'Radio campaigns for adult education in Tanzania' (Conference paper, Stanford University) D Mbunda and A B Edington (1979) Educational media in Tanzania (n.p., ? British Council, London) J W Mellor quoted in D G Green 'Non-formal education for agricultural develop- ment: a system perspective' in P Foster and J R Sheffield Education and rural development (World Year Book of Education 1974) (Evans, London, 1973) Ministry of Local Government and Lands (1977) Lefatshe la rona: our land Ministry of Local Government and Lands, Gaborone) S Musto (1971) Los medios de communicacion social al servicio del desarrollo rural (Editorial Andes, Bogota) F Orivel (1981) The impact of agricultural extension services: a review of the literature (World Bank Discussion Paper) H Perraton (1977) Starting the Botswana Extension College (International Extension College, Cambridge) H Perraton, D Jamison and F Orivel (1981) Mass media for agricultural exten- sion in Malawi (World Bank Discussion Paper) T A Phillips and M P Collinson, (1976) 'The organisation and development of small-holder schemes in the programmes of the Commonwealth Development Corporation' G Hunter, A H Bunting and A Bottrall Policy and practice in rural development (Croom Helm, London, 1976) R Postgate et al. (1979) Low cost communication systems for educational and development purposes in third world countries (ITDG, London/Division of structures, content, methods and techniques of education, Unesco, Paris) V M Primrose (1965) A study of the effectiveness of the educational program of radiophonic schools of .Sutatenza on the life of the Colombian peasant farmer (St Louis University PhD thesis) Thomas a Rich et al. (1978) Basic village education Guatemala Evaluation Report: Evaluation Summary no. 3 (Human Resources Institute, University of South Florida) Felipe Risopatron and Peter T Spain (1979) An evaluation of Dialogo: Human sexuality education through radio in Costa Rica (Institute for Communi- cation Research, Stanford) (mimeo) R Rondeau (1973) Un exempled'education permanente Par les mass media: 1'experience des radio-clubs du Niger (Universite Rene Descartes, Paris) (mimeo) -145- P Roy, F B Waisanen and E M Rogers (1969) The impact of communication on rural development: an investigation in Costa Rica and India (Unesco, Paris) D M Savicevic (1973) 'A review of research on functional literacy in Yugoslavia' Literacy discussion 4:3 W Schramm (1967) 'Ten years of the radio rural forum in India' in ed. W Schramm (1967) New educational media in action: case studies for planners Vol. 1 (Unesco/IIEP 1967) S Scribner and M Cole (1978) 'Literacy without schooling: testing for intel- lectual effects' in ed. K King (1978) Literacy research in developing countries: final report (International Development Research Centre, Ottawa and German Foundation for International Development, Bonn) Larry Shore (1980) 'Mass media for development: a re-examination of access, exposure and impact' in Emile G McAnany (1980) Communications in the rural third world (Praeger, New York) 'Sudan: cost problems of literacy campaigns' Literacy work 1:1 (1974) M K Trabelsi (1972) 'Les couts de la radio educative rurale du Senegal' (IIEP, Paris) (mimeo) H P Tuckman and T F Nas (1980) 'Accion Cultural Popular (Colombia): alternative methods of determining project costs' in Unesco (1980) The economics of new educational media vol. 2 (Unesco, Paris) Unesco (1976) The experimental world literacy programme: a critical assessment (Unesco Press/UNDP, Paris) Unesco (1977) The economics of the new educational media (Unesco, Paris) Unesco Regional Office for education in Latin America and the Caribbean (1975) MOBRAL - the Brazilian adult literacy experiment (Unesco, Paris) Academy for Educational Development (1977) The Basic Village Education Project Guatemala: Oriente Region - Combined Report (AED, New York) Unesco/UNDP (1976) United Republic of Tanzania Functional Literacy Campaign and Materials Development Project Findings and Recommendations (Unesco, Paris) Unesco/UNICEF (1978) Basic services for children: a continuing search for learning priorities (Unesco, Paris) UNICEF (1978) Basic Education for nomads (Report of seminar in Mogadishu) (UNICEF Nairobi) R White (1976a) An alternative pattern of basic education: Radio Santa Maria (Unesco, Paris) R White (1976b) Mass communications and the popular promotion strategy of rural development in Honduras (Institute for Communication Research, Stanford) I Yazidu (n.d.) 'The study of radio as a means of communicating agricultural information to farmers in northern states of Nigeria' (unpublished MSc thesis Ahmadu Bello University, Zaria) M Young, H Perraton, J Jenkins and T Dodds (1980) Distance teaching for the third world (Routledge, London) PART III MASS MEDIA FOR AGRICULTURAL EXTENSION IN MALAWI Hilary Perraton Dean T. Jamison Francois Orivel Table of Contents page Acknowledgements ............................................ 149 Abbreviations 149 Currency ......... .................. 149 SUMMARY 150 BACKGROUND ........................................... 151 AGRICULTURAL EXTENSION AND TRAINING IN MALAWI .151 Agricultural extension .151 Training Centres .155 EXTENSION AIDS BRANCH .156 Publications section .157 Technical section .159 Cine section 160 Radio section .160 Evaluation and action research unit .161 RELATIONSHIPS .162 CONSTRAINTS 162 RESULTS .164 Measures of output .164 The audience .166 Did they learn anything? .168 A COMPARISON WITH THE WORK OF EXTENSION OFFICERS .171 COSTS .173 The cost of radio .174 The cost of mobile units .174 The cost of print .177 The cost of evaluation and administration ............... 178 Comparison with extension services .180 CONCLUSIONS .182 APPENDIX 1 : Main print jobs 1978 ............ 184 2 : Mobile vans schedule - northern region 1980 ....... 185 3 : Source of agricultural information for farmers ... 186 ANNEX A ...................................................... 187 ANNEX B 9.....................................................6 REFERENCES . 200 -149- ACKNOWLEDGEMENTS This report is based on a visit to Malawi by Dean T. Jamison (World Bank) and Hilary Perraton (International Extension College) in May 1980. We are grateful to the staff of Extension Aids Branch, and of other parts of the Ministry of Agriculture, who were most generous of their time, their hospitality and their information. It is a pleasure to acknowledge their help here, and that of Unesco who funded Mr. Perraton's work on the study. The views expressed in the report are our own and not necessarily those of the World Bank or of the Malawii Ministry of Agriculture. ABBREVIATIONS ADMARC Agricultural Development and Marketing Corporation EAB Extension Aids Branch EAR Evaluation and Action Research Unit IADD Lilongwe Agricultural Development Division LLDP Lilongwe Land Development Project MBC Malawi Broadcasting Corporation CURRENCY Currency Unit Kwacha (K) 1980 US$1.23 = K1.00 -150- SUMMARY This paper reports the results of research concerning the role of mass media in providing information to Malawian farmers. To set the context, the paper begins with a brief description of the face-to-face information services provided by the agricultural extension system and by the farmer training centers; it then describes in more detail the operations of the Extension Aids Branch (EAB) of the Ministry of Agriculture. EAB prepares and delivers messages to farmers using radio, print, and mobile vans (carrying films for night-time use and puppet shows for daytime use). Of the approximately $6 400 000 per year that agricultural information services cost Malawi (or $6.80 per farm household), about 78% is for agricultural extension, about 12% is for farmer training centers, and about 10% is for the media provided by EAB. After describing agricultural information services, the paper assembles available information, concerning media effectiveness and costs. A number of studies were reviewed concerning sources of farmers' information and how much had been learned from media exposure. In one study, 67% of respondents listed extension workers as a source of agricultural information, 8% listed training course attendance, 27% listed radio, 9% listed films, and 2% listed puppets. Other studies have indicated that, even though radio and films are carrying a lot of information already known to farmers, they also communicate effectively; farmers learn from both these media and they are popular. Finally, even though results of all the effectiveness analyses must be received with some caution, there is evidence that the information services can pay off in improved productivity. In Annex A, for example, several agricultural production functions are estimated from a Malawian data set of 300 farmers, from which it appears that the more educated farmers tend to have more contacts with extension agents and that farmers with more contacts with extension agents tend to have better yields and are more likely to produce cash crops (but not at the expense of subsistence crops). The paper's analysis of costs found, not surprisingly, dramatic differences in the costs of different ways of reaching farmers. Agricultural extension agents were estimated to cost about $21 per farmer contact; residential training centers cost about $30 per farmer for a five-day session; and day training centers cost about $4 per farmer for a one-day session. The EAB's films cost about $0.17 per farmer contact-hour (a mobile van will typically show 2 or 3 hours of films on a given night in a given place); the puppet shows (lasting about an hour) cost about $0.08 per viewer. Radio's costs, at about $0.004 per listener-hour, are by far the lowest of any medium. Put slightly differently, the paper estimates that mobile vans (film and puppet shows) cost about 44 times as much per contact-hour as radio; and extension agents cost 55 to 80 times as much as mobile vans. Economies from mass outreach are evident. What is less clear, in Malawi and elsewhere, is how best to make the mass media complementary to face-to-face methods. Nonetheless, even if there were severe limits on the amount of information carried by the radio programmes, it is hard to resist the conclusion that the broadcasts are highly cost effective. This conclusion has implications for Malawi; and it has perhaps more important conclusions for other countries, which, unlike Malawi, at present make almost no use of radio for reaching farmers. -151- BACKGROUND When Malawi was nearing independence, it was described in Britain as having no assets except its scenery and an educational system so patchy that there were only a dozen graduates in the whole country. Life President Banda, at one time feared by the British as a revolutionary, has argued consistently and successfully that Malawi's gold lies in its soil. With a firm government, extensive international aid, and fair soil and rainfall, Malawi has concentrated on rural development with the result that it is now a significant exporter of ground-nuts, maize, and tobacco. Until recently, Malawi has pursued a dual strategy to raise agricultural productivity. In four parts of the country (Lilongwe, Shire Valley, Karonga and Lakeshore), integrated rural development projects were set up between 1968 and 1974 and resources concentrated in these. For the rest of the country, an agricultural extension service was developed; its staff in 1977 was 1 776. At the same time, a single marketing corporation, the Agricultural Development and Marketing Corporation (ADMARC) was established to buy all the cash crops throughout the country and to sell agricultural supplies. More recently, the four integrated rural development projects have been incorporated with the regular work of the Ministry of Agriculture and Natural Resources which now operates in eight agricultural development districts. In the long run, this will reduce the disparity between the project areas, where the supply of resources and staffing ratios for agricultural extension have been relatively generous, and the rest of the country. In an attempt to increase the effectiveness of its agricultural extension service, Malawi has for more than twenty years run an 'Extension Aids Branch' (EAB) which uses printed pamphlets, radio and mobile cinema vans. This report assesses the significance of the Extension Aids Branch. AGRICULTURAL EXTENSION AND TRAINING IN MALAWI Agricultural extension in Malawi has developed three techniques for reaching farmers: through individual or group visits to farmers by field assistants; by teaching farmers at day or residential training centres; and through the Extension Aids Branch. Table 1 sets out the basic facts about each. The two subsections of this section describe agricultural extension and the training centres; the next section describes the EAB. Agricultural Extension In the country as a whole, there were in 1977 1 388 field assistants, and 15 residential centres (World Bank 1980, pp. 30 and 35); in 1980, there were said to be about 75 day training centres. There are more training centres and a better ratio of field assistants to farmers in the four former project areas than in the country as a whole. Nationally, the ratio varies between 1:500 and 1:1500 or even more. Within the former Lilongwe Land Development Project (LLDP) it *is about 1:600; we can take this area as an example to illustrate the structure in which the field assistant is working (see Figure 1). For agriculture, the country is divided into eight agricultural development divisions, of which the Lilongwe Agricultural Development Division, for example, is one. This division is then divided into six project areas, whose project managers report to a programme manager. Each project manager is usually a graduate in agriculture, having followed a degree course, generally at: the University of Malawi. The development officers have a diploma in agriculture, which is awarded after a three-year course. Entry to this course, as for a degree course, is the Malawi Table 1: Agricultural Information Services in Malawi -----ApproximateCostse/---- Service Outreach Total Annual Per Unit Agricultural Extension Agent There are approximately 1,400 The cost is about The cost is about field agents providing about $4.97 million per $20.1 per farmer 240,000 contacts with farmers year. contact. per year. Farmer Training Centers There are 30 residential centers The cost of residen- The unit cost is (Residential) providing for 12,000-17,000 men tial training about $30.4 per and women farmers and 3,000-4,000 centers is about person per week staff each ya7r for 85,000 $0.515 million per (5 days). student days- . year. Farmer Training Centers There are 75 day centers The cost of day The cost is about (Day) providing about 70,000byeraon/ training is about $4.1 per farmer days teaching annually- . $0.286 million per day. per year. Extension Aids Branch About 4-1/2 hours per week are The total cost is The cost is about (Radio) broadcast by Malavi Broadcas- about $0.271 $1,160 per broad- ting Corporation, reaching million per year. cast hour or about about 30S of the populAtion. $0.0038 per listener hour. Extension Aids Branch There are 17 mobile vans The total cost is The coat is about tn (Mobile vsns - films) providing 2.5 hour cinema 0.322 million $0.17 per viewer show, reaching an annual per year. hour. adult audience of 765,000s'. Extension Aids Branch The same 17 mobile vans The total cost is The cost per viewer (Mobile vsae - Puppet shows) provide puppet shows for an *0.089 million is gbout $0.08. adult audience of 1,060,000 per year. per year. Extension Aids Branch There are 30,000 literate The total cost is The cost is $2.33 (Print) farmers who receive printed about $0.067 Rl- per farmer. material, lion per year- a/ Malawi Basic Econonic Report Annex: Agriculture Smallholder Sector, para. 3.35. b/ ibid, para. 3.34. c/ The figures are extrapolated from EAB's experience with 12 mobile vans in 1972-78. d/ The cost of print support for agricultural extension agents, 0.070 million, is included in the cost of the extension agents, rather than in EAB print costs. e/ Data on EAB costs come from the cost section of the main body of this paper; the data on other costs come from Annex B. f/ There may, of course, be more literate farmers than this, and each copy may be read by more than one person. -153- Perrmanent Secretary MLinistry of Agriculture and Natural Resources Under-Secretary Deputy Secretary Deputy Secretary (administration) (natural resources) (agriculture) rLE I Animal Agricultural Extension Aids Chief Agricultural health research Branch Development Officer I~~~~~~~~~~~ Lilongwe Agricultural 7 other National Rural Development Technical Development Division ADDs Programme Co-ordinator posts - Project Manager I - - I I Project Manager 5 other Specialist project managers staff Development 39 other officer development officers I I - ' 4 field assistants 2-3 credit Farm Home assistants Instructress Figure 1: Organisation of extension staff a/ a/ We understand that the organisational structure has been reviewed and this chart is now out of date. -154- Certificate of Education, which is taken after four years of secondary education. At the next level down, field assistants are trained for two years at the Colby College of Agriculture and the entry qualification for their course is Junior Certificate taken after two years of secondary education. Education brings better transport as well as better jobs. Project managers are each provided with a vehicle and development officers with a motorbike. Field assistants are expected to travel by bicycle and are paid a bicycle allowance of K 3.50 per month. The cars are said to average 2 400 kilometres a month and the bicycles 1 280. The field assistant's key activity is to demonstrate better agricultural practices to individual farmers or to groups of farmers. In theory, he does this partly by cultivating demonstration plots himself, partly by actual demonstrations on his farmers' lands, and partly by reiterating the advice passed down to him through the hierarchy of agricultural extension. In the past, extension workers have been encouraged to concentrate their efforts on progressive farmers, achikumbi. There seems to have been a move away from this approach and towards group demonstrations and advice. The 1978/79 Guide to Agricultural Production says: 'The achikumbi programme has been given top priority by Extension staff since it was introduced. Field staff are aware that farmers must be of the highest agricultural performance if they are to receive the title "Mchikumbi" and that they should be real leaders in the field of agriculture. Achikumbi are established farmers who have shown over a period that they are fully entitled to the honour. All agricultural staff should encourage as many farmers as possible to reach the high standards required to qualify as Mchikumbi. In recent years, groups of farmers have got together in order to organise themselves in various ways that can help them to achieve better standards of farming. Many of these groups have been formed to facilitate the purchase of fertilizer and improved seed, and some well established groups have been able to obtain credit for this purpose. Such groups or farmers' clubs have also often organised concerted plans for agricultural development in their local area. The Government welcomes this development. A strong local group of farmers interested in improving their standards of farming gives the staff of the Extension Service a good chance of helping more farmers effectively. ................ Selection of achikumbi is a continuous process. 380 achikumbi have so far received certificates, and His Excellency the Life President is Malawi's Mchikumbi No. 1.' But field assistants have other jobs. They are the main contact between the farmer and specialist services of the Ministry of Agriculture, such as advice on forestry. The field assistant also has the job of encouraging farmers to attend one day or residential courses. He accompanies the yellow mobile cinema and puppet van when it visits his district and gets a spot of publicity for himself and his messages by taking the microphone after each puppet show. And he submits weekly returns of activity to his development officer. He does not, however, deal with credit--the function of a cadre of untrained credit assistants-- nor does he supply seed or fertilizer, which are sold through ADMARC. And -155- he does not teach women; despite the relatively high number of female heads of households, role segregation is seen as a reason why male field assistants should not teach women farmers. There is a small cadre of Farm Home Instructresses, but their work is mainly, though not solely, in teaching nutrition, cookery, needlework and health, rather than agriculture. It is not possible t precise about the field assistant's work. In Malawi, as in many other countries, field assistants are criticised for concentrating too much on wealthy farmers. The system of reporting is inadequate; one development officer we asked did not know which activity should go under which heading on the report form which he submits regularly month by month. As far as we could discover, no studies appear to have been made of the day-to-day activities of field assistants. At the next step up the hierarchy, the development officer may control a staff of about a dozen. At Nathenje in LLDP, for example, there are four field assistants, three credit assistants, one farm home instructress, two labourers who work on the demonstration farm, a messenger and a clerk. The office is housed in a substantialbuilding on the main road. In the development officer's office, there is a desk with a set of wooden trays, a notice board, two cash boxes, one for regular sales and one for credit, half a dozen chairs and unusually a telephone. He splits his time between supervising his staff and planning their work, visiting villages, teaching on one-day courses, and keeping in touch with the higher levels of the extension service. He attends regular meetings every quarter/month with his project manager, at which his work is guided and he is told of the subjects which Extension wants stressed to farmers over that period. We will return later to consider the effectiveness of this quite orthodox, approach to agricultural extension. Training centres The second main way of teaching farmers is through day and residential training centres. Throughout the country, there were, in 1974, 59 centres: there are, of course, more in the areas of the former land development projects. In the former LLDP area, for example, there are four residential training centres and six day training centres. In the year 1979/80, for an estimated 100 000 farm families they ran courses as in Table 2. Table 2: Use of residential and day centres in LLDP area 1979/80 ------ Staff --------- ---------Public--- Men Women Total Men Women Total Residential centres 7 437 1 317 8 754 9 877 9 411 19 288 Day centres - - - 31 258 37 760 69 018 The subjects of residential courses included home economics, credit procedure, forestry, fruit trees, land allocation, ox training, and general agriculture. The courses offered at one day training centres in Lilongwe division for the period February-July 1980 are in Table 3. Three courses--on crop storage, family health and horticulture--were for both men and women: -156- Table 3: Courses at Nathenje training centre February-June 1980 No of courses No of courses Course subject for men for women Animal husbandry 3 - Child care - 2 Credit 1 - Crop husbandry 5 Crop storage 1 1 Family health 2 3 Farm management 3 - Forestry management 2 - Home improvement - 5 Horticulture 5 2 Land husbandry 3 - Laundry - 3 Needlework and handcraft - 8 Nutrition and cookery - 6 Poultry keeping - 1 Preparation for display - 1 Revision 5 1 30 33 Courses for farmers are free, as is accommodation and food at residential training centres. Each residential centre has a demonstration farm so that costs can be kept down by growing food for courses--though horticulture demands expensive watering in many parts of the country. The centres try to provide nutritionally well-balanced meals, seeing the teaching of nutrition along with agriculture as one of their proper functions. Courses at day training centres are taught mainly by field extension staff while each residential centre has a principal and resident teaching staff. EXTENSION AIDS BRANCH Beyond these two traditional ways of instructing farmers, Malawi has, since 1958-59, used mass media organised through its Extension Aids Branch. It began as a one-man public relations unit from the Ministry of Agriculture and Natural Resources to provide publicity material to support the agricultural extension services (Extension Aids Branch 1978, p. 1). The unit soon grew to have a public relations officer, a senior agricultural instructor and a dark-room assistant. In 1960, it began producing leaflets and two weekly radio programmes. For two years, it worked under the control of the Department of Information before moving back to the Ministry of Agriculture in 1962 and being expanded. Further exppnsion came in 1966-67 when British and American aid funds paid for printing equipment (still largely in use today). In 1968, the Branch set up a film unit so that it was capable of producing and distributing film as well as print and radio programmes. In order to raise the effectiveness of its work, the Branch started in 1976 Evaluation and Action Research Unit, whose function was to pre-test material produced by the Branch. Today the Branch has an establishment of 76, to whom some 62 are actually in post, to- gether with some 30 non-established staff, and consists of eight sections: publications, editorial, radio, technical (which runs the mobile vans), photographic, cine, evaluation and central administration. Its functions are defined as 'to provide educational teaching material for the Ministry -157- of Agriculture and Natural Resources, plus other Ministries to some extent, to create an interest in rural people in modern farming methods and to encourage them both to increase production thereby improving rural living standards. It is also concerned with the health and proper nutrition of rural people and works closely with other extension services.' More specifically, '- It provides mass coverage through the radio; - It compiles extension literature; - It provides visual material; - It provides maintenance and mobile cinema services; - It provides movie films; - It provides media evaluation services; and - It undertakes printing services for written media.' (Extension Aids Branch 1979, p. 1) It is housed in a standard office block three storeys high, about 40 metres by 10, indistinguishable at first glance from others in the capital. This has been its home since 1973. At the back is a line of mouldering yellow Land Rovers, half hidden under the trees. EAB has wisely got permission to keep its vehicles when they are scrapped in order to use the spare parts. Inside too, it looks at first sight much like any other office, apart from a display in a glass cabinet of the Branch's publications on sale. Its activities are quite different. Its staff are set out in Table 4. Publications section The largest number of staff is concerned with print. The publications department is divided into four sections--editorial, typesetting, graphics, and printing. Tlhe editorial section has the job of preparing the text of everything to be printed by the Branch, and of writing the puppet plays which are to be performed in the mobile vans. Its three staff thus divide their time between translating material drafted in English into Chichewa, preparing text for farmers' magazines, checking and proofreading and writing the puppet plays. This last activity takes about one man/week in each month, in which time the texts of eight plays are produced. Otherwise, by far the largest job of the section is the preparation of the bi-monthly farmers' magazine Za Achikumbi. It runs to 16 pages (255 x 190 mm) and is now printed in an edition of 32 000 copies and circulates to extension staff and to literate farmers. The text consists of advice to farmers appropriate to the time of year when it is published. To select topics for inclusion, the editorial staff may refer to the cropping guide, to the annual radio programme guide, which sets out the subjects for broadcasts, and make a request for guidance to project managers. While articles can come from a variety of sources, it in fact falls to the editorial staff to generate a considerable amount of the copy themselves. The Branch also publishes four magazines in English on behalf of the Land Development Projects. Beyond that, it prints posters, leaflets and handbooks on the production of different crops. A handbook on cotton, originally produced in 1971 and revised in 1976 was, for example, being reprinted at the time of this report, and a handbook on coffee was to follow it. Appendix 1 shows the main production jobs for 1978. Drafts produced by the editorial section are then typeset; an IBM composer has just been added to the range of typewriters used by the Branch, which will enormously aid the production of material with the -158- Table 4: Extension Aids Branch Staff Section Establishment In post - May 1980 Administration 15 1 head of branch 1 assistant head of branch 1 executive officer 1 senior clerical officer 2 clerical officers (+ I away on training) 2 copy typists 1 stenographer 1 head messenger 3 messengers 2 watchmen Editorial 4 1 head of section 2 (+1 away on training) Publications 32 1 head of section 1 assistant press supervisor 1 senior typesetter 3 typesetters I graphic artist (+1 away on training) 6 graphic assistants 8 printers 7 collators 2 press workers Photographic 3 1 photographer 1 photographic assistant Cine 6 1 head of section/script- writer 1 cameraman 1 assistant cameraman 1 sound recorder 1 librarian/commentator Technical 34 1 head of section 1 maintenance officer 2 maintenance assistants 2 carpenters 15 driver/operators 13 puppeteers Radio 10 1 head of section 4 recording assistants 5 radio correspondents in field Evaluation and action 8 1 counterpart head (+ coun- research terpart away on training) 1 technical officer 6 enumerators -159- right hand margin justified; at present this is achieved by laborious dual typing of manuscripts. Titles are added with the aid of a headliner and the copy made up by the graphics section who pass plates on to the printing section. Here a team of eight printers operate three ageing and small presses (capable of taking A3 size paper at a maximum); the greater part of their time, too, is spent in producing the various farmers' magazines. A team of eight collators put printed material together by hand. The head of the publications section controls the flow of work through its various stages. While the photographic section is organisationally separate from publications, much of its work feeds into the publications section; still photographs are commissioned by the section, and photographs from their collection are selected to illustrate magazines and reports. Two photographers use 35mm (Pentax SLR) and 6 x 6 cm (120 dual lens reflex) formats for both black and white and colour slide still photography. Technical section The technical section operates a fleet of mobile vans--Land Rovers converted to show cinema programmes and puppet plays throughout the country. There are 17 mobile vans in all, of which four are attached to the former Land Development Projects so that the Technical Section in Lilongwe employs 13 drivers and 13 puppeteers. Its staff at Lilongwe convert and maintain the vans and their equipment. They are then sent out for a series of two to three week campaigns in which they show the same programme of films and' puppet shows in villages chosen according to a programme agreed by the local field extension staff. Thus, while there is a centrally produced programme showing the area in which each van will operate for each campaign, the details of each van's day-to-day schedule are arranged by local field staff. The vans are ordered from British Leyland in England and four adaptations are made by the factory: an oil cooling system and a governor to control engine speed are added for use when the vehicle is stationary but with the engine running to drive a generator; a heavier back axle is fitted; and a power take-off unit fitted. A contractor in Lilongwe fits a generator so that the cine projector can be run from the vehicle. The technicians and carpenters at EAB then instal a tape recorder, to carry music and the dialogue of puppet plays, amplifier and loudspeakers, and a puppet stage at the back of the vehicle. A cine projector and folding screen are also carried. The cost of the conversion is kept down by doing as much work as possible at EAB. The vans are operated by a team of two--a driver/operator and a puppeteer. Each van goes out for a campaign of two to three weeks, carrying films and puppet plays on a set of topics appropriate to the agricultural season. Appendix 2 shows, for example, the schedule for the four vans operating in the northern region for 1980. At the beginning of the campaign, the van drives from Lilongwe to the selected area and reports to the local agricultural supervisor. He, or a field assistant, then accompanies the vehicle. It drives to a village using its loudspeakers for music and announcements to advertise the puppet plays which are shown in four or five different villages during the day time. The plays usually consist of dialogue and agricultural practice between a farmer and his wife, or a farmer and a field assistant. The puppeteer operates the glove puppets for which he has had a two or three week training, but the dialogue is given from a tape recording and broadcast on the loudspeakers. Little attempt has been made to introduce a story line into the plays. At the end of each performance, the field assistant uses the loudspeaker in order to reinforce the advice given in the play. -160- As the vans drive round, they announce over the loudspeaker where they will show films that evening. After sunset, a screen is erected and a two to two and a half hour programme of films is shown. The majority of the films are those made by the cine department of the Branch; single concept films on improved practice for a particular crop, or on the use of fertilizer, or marketing, for example. Along with them, some advertising films may be shown together with old newsreels, occasional documentaries often from embassy film libraries and the like. And, in many programmes, a film from the Malawi Department of Information is shown. Usually these are appreciably longer than the Branch's own films. At the end of the campaign, the vehicles return to Lilongwe for servicing and preparation for the next campaign some ten days to three weeks later. The vehicles remain at headquarters during the heaviest period of agricultural work, from December to February. Cine section 'Before the (cine) Section was established, there were very few films on Agriculture, and worse still, the films were in English, which made it difficult for the farmer to understand.' (Extension Aids Branch 1978, p. 8). In 1968, EAB therefore established its own film unit which today produces up to fifty films a year. The staff were trained on the job by a small team from the University of Missouri. Having experimented with a variety of formats in the past, the unit now produces mainly single concept films, which generally run for about ten minutes; occasional films are longer, and a few are made on commission for other government departments. The unit is equipped with three Bolex 16 mm cameras and two Auricon (sound) cameras. The complexities of recording sound at the same time as filming are such that the Auricons are no longer in use. In 1979, the Branch bought a Steenbeck console editor; all editing of films and the preparation and dubbing of sound tracks is done at the Branch. Films are sent to a commercial laboratory for processing--Malawi is now again sending them to Zimbabwe for this--and once they have been edited and dubbed, the master is then also sent for prints to be made. 22 copies of each film are usually made, and each copy is then used for about five years, with perhaps eighty showings during this time. Films are generally produced in colour with a soundtrack in Chichewa. The Branch does not employ actors, but gets farmers to demonstrate agricultural practices which adds conviction and realism to the films. It manages to achieve a shooting ratio of 3:1. Many films are on particular agricultural techniques - 'Tobacco barn building and repairs' or 'Cotton weeding' or 'Rice management' or 'Visoso' (slash and burn cultivation), for example; others are more general. And some go beyond agriculture: 'Nkhokwe construction' describes methods of grain storage - 'Use clean water' and 'Eat more eggs' are self-explanatory. Radio section The Branch produces four-and-a-half hours of broadcasting each week for six radio programmes: Farm forum is the key agricultural programme, dealing each week with agricultural topics felt to be of particular importance for that week in the agricultural year. Its content is a guide to the material prepared for the other programmes as well. (Broadcast on Monday, -161- repeated on Wednesday, from 1312 to 1400). Modern farming is a general programme, making considerable use of interviews with farmers. The Branch employs two full-time and one part-time field worker who collect recordings for this and other programmes. (Broadcast on Tuesday, repeated Friday, 1320-1400). Cotton is an eight minute insert within Modern farming consisting of a talk about cotton production and marketing. Farmers request programme is a music request programme, whose principal functions are to provide entertainment and associate EAB in the mind of listeners with a popular programme. (Broadcast on Thursdays, 1310-1400). Farmers' notebook is a title for a series of short (two to three minute) radio news spots dealing with current activities of farmers, credit, cattle market. arrangements and so on. (Broadcast between 0500 and 0600 daily). Ophiri is a farming family serial produced by the Malawi Broadcasting Corporation (MBC) using information from both EAB and from the Ministry of Health. (Broadcast on Friday, 1145-1200, repeated Saturday, 1645-1700). With the exception of Ophiri, all programmes are both scripted and produced by the radio section's staff and recorded in the Lilongwe studio of NBC. The Branch has its own portable Uher tape recorders and two Ferrograph recorders. Programmes are edited at EAB. All the broadcasters are agricultural extension workers by background, who have been trained in radio techniques and attend occasional seminars at EAB as part of their in-service training. There is one woman producer who, as well as dealing with some agricultural topics, prepares all the material for broadcasts aimed particularly at women. Evaluation and action research unit (EAR) The aim of the evaluation unit is not to measure or test the performance of EAB, but to pre-test materials as they are being produced. The unit is the most recently established section of EAB. The idea for it first took shape at the Agricultural Extension and Rural Development Centre at Reading University, England, when two of its staff were invited by Extension Aids to advise on methods of evaluating its work. After visits and discussions, the Reading team put forward its suggestions. It did not propose the traditional form of evaluation exercise in which a visiting expert studies the institution, makes an assessment and produces a report with general suggestions for improvement. Instead, the team put forward the idea of a small permanent evaluation unit charged with the task of helping to improve the effectiveness of each media project or campaign during the production stage. (Warr 1978, p. 122). The unit cannot test all materials produced by the Branch, but selects from those being drafted for radio, film or print. Its team of six enumerators test materials on a sample of the potential audience and recommend any necessary changes to producers within the Branch. While -162- this adds to the lead time needed for preparing any agricultural campaign, it is intended to increase the effectiveness of all the media used. Evaluation is often seen as a threat: the EAR unit makes a strenuous effort to gear its research to the needs of producers so that they see its work as helpful and its staff as supportive rather than externally critical. While its role is limited to formative evaluation and to advising on particular topics within EAB's work, it has inevitably gathered a great deal of information about the effectiveness of the Branch's work more generally: we make considerable use of this information below. Relationships Strangely, the Extension Aids Branch is no longer a part of the organisational structure of agricultural extension. The various Agricultural Development Divisions (which have taken over from the integrated land development projects) report to the Chief Agricultural Development Officer, as do the technical specialists, and he, in turn, reports to the Deputy Secretary (Agriculture). The Extension Aids Branch also reports direct to the same Deputy Secretary (see Figure 2). In practice, there is, of course, informal cooperation with agricultural extension staff, but cooperation seems neither as close nor as regular as it might be. EAB's other key relations are with Malawi Broadcasting Corporation, where it works with a producer in Lilongwe although MBC's headquarters are in Blantyre, and with farmers. Appendix 2 sets these out. As we have seen, EAB communicates with farmers through three channels: by radio and yellow vans it does so directly, although the yellow van puppet plays are introduced and followed by announcements from the local field assistant. Its communications through print are necessarily indirect: with a literacy rate estimated at 25%, few farmers could read Za Achikumbi even if they were to receive it. The magazine's major value lies in supporting and stimulating the work of field assistants. Feedback from farmers to EAB presents more problems. In theory, farmers are in close contact with field assistants and comment to them on the work of the Branch. Field assistants meet regularly with their supervisors, and can pass on comments about farmer reaction. In practice, the meetings at which this might happen take place once a quarter, and there is much else on the agenda. The drivers of the yellow vans report on audience size and can give the cine section an impressionistic view of how particular films are received. The EAR unit was established specifically to guide the Branch-- but only on the effectiveness of its materials and not on teaching arrangements generally, or on the content of its work. Constraints Before attempting any assessment of the work of the Extension Aids Branch, it is important to note the constraints within which it is working. The first is the constant pressure to maintain the flow of production. Week by week and month by month, the Branch produces its quota of radio programmes and the regular issues of periodicals. Its establishment is below strength and any proposal for it to undertake new tasks must be considered in the light of its existing commitments. Second, we formed the impression that the Branch suffers from its location. It is physically in a different building from other parts of ag- ricultural extension and, as we saw, now organisationally separate from it. In practice, it seems it has had closer links with LLDP than with the other development projects or other parts of the country, simply because the LLDP |ADeputy SecretaryR [Agriculture Chief Agricultural A FDevelopmenst Officn A B Rlin p Agriculnfouial D vlp Extension Aids informal T cncal | Malawi |ment Divisions | contactBranch contact Sevcs|Broadcasting I - + ' -fozrai and informal contact--- . z intu | ~~~~Yellow vans] ai < | ~~quarterl Lmeetings o Z > / t .+ Evaluation and c _ _ | . , , ~~~~~~~~~Action Research to a few to extension a _ _a__ Unit farmers lagents Icxpn | ~~~~~~~~F A R Di E R S Figure 2: Extension Aids Branch Relationships -164- offices are close by. It is physically separated from those doing agricul- tural research and teaching agricultural extension, who remain at Zomba. There is a third constraint which is more subtle but potentially more serious. In discussions there, the dominant word in many conversations was 'messages'. The Branch seems to be working under the assumption that its role is to send messages to farmers, although agricultural extension thinking generally has moved away from this view of its function. This has consequences for the way EAB's work is planned. It tends to lead to the assumption that all their energies should go into the perfecting of the form of those messages whose content can be determined by the agricultural specialists. In practice, we were told more than once that farmers in Malawi are sophisticated and skilled and, over the last ten years, have moved a very long way from needing simply to be told what to do. And this assumption about messages means that relatively little effort has been made to build systems of feedback into the organization of EAB. RESULTS In order to make a full assessment of the work of EAB, one would want to compare the different methods used by agricultural extension to contact farmers, and then look at the effects on those farmers and on their farming practices. Ideally one would then be able to look at the cost of reaching a farmer through radio or through a residential course, for example; look at the effects of that contact on the farmer's production of food and compare the costs. In practice, there are major difficulties in the way of doing so. The most serious is that farm practices and production are affected by many factors other than contact with the extension service. Marketing and pricing policies of ADMARC have a profound effect, for example, on farmers' practices. Furthermore, one of the aims of agricultural extension is to achieve long term changes in farming practices and in the attitudes of farmers; these would not be discovered by an analysis that was based only on increased productivity. In any case, there is in practice not quite enough data to do this comparatively simple comparison. Malawi is, however, fortunate in having a wealth of information about agricultural productivity and about the work of extension services, and the effects of education and other variables on agriculture from the socioeconomic survey of the Ministry of Agriculture, from the evaluation units of the Land Development Projects, and from the Evaluation and'Action Research Unit of the Extension Aids Branch. It is on the basis of this kind of information that we can make some estimates of the effectiveness of both EAB and regular agricultural extension; it would -also be possible to pursue this analysis very much further and we return to this point in our final section. Measures of output We can start by measuring the output of the Extension Aids Branch. While this says nothing about effectiveness, it does give us some key information about the productivity of the tranch. Table 5 sets out the amount of work done by their print department 1972-1978. The printing equipment in the publications department is old, and both the 1976 and 1978 reports of the Branch refer to delays in production because of the breakdown of machines and difficulties in servicing them. The production of publications was also affected by the amount of work done by the Branch for other departments. Thus in 1976, for example, the editorial section was 'involved in editing and proofreading all Table 5: Production of print 1972-78 1972 1973 1974 1975 1976 1977 1978 Average 72-78 Job8 252 173 193 166 159 191 179 186 Impressions 2,313,878 1,354,705 1,322,914 1,719,640 1,630,911 2,035,920 2,364,992 1,820,423 ON lA -166- publication material requested by other departments of the Ministry of Agriculture and Natural Resources. The Section was engaged with production of materials for "Zikomo Kamuzu Show", the Trade Fair and the Malawi Congress Party National Convention.' (Extension Aids Branch 1976, P. 1). It is more difficult to produce a statistical summary on the work of the radio section. Their number of broadcasting hours has, however, remained constant over the last two years at about 3 hours per week (including Ophiri for which responsibility was shared), 1-3/4 hour repeats or about 156 hours of original broadcasting each year. Table 6 sets out the number of films produced over the same period 1972-78. Table 7 sets out the activity of the mobile units of the Branch. These figures show only the activity of the mobile units controlled directly at that time by Extension Aids Branch and not the work of the units controlled by Land Development Projects. No figures are available for the number of puppet shows or the audiences at these. It seems likely, however, that there were about three times as many puppet shows as film shows. The annual audience for puppet shows was probably between one and two million people a year; it is also almost certain that a larger proportion of the audience were relatively small children. At the first puppet show we saw, for example, we estimated the audience at 20 men, 20 women, and 40 children. The audience There is little information on the audience for EAB's printed materials. As we have said, the main value of these must lie in keeping the agricultural extension service informed rather than in communicating directly with the farmers. But it is important to stress that the magazine Za Achikumbi and the programme guides for radio programmes and for the mobile cinema vans are key planning documents for the extension officer's work. The work of regular face-to-face extension and the use of vans and radio would be seriously affected if such publications were not produced. Evidence about radio listenership is rather fuller. The EAR Unit carried out a series of studies of radio between 1977 and 1979, and found that one third of farmers claimed to listen regularly. They thought that this was probably an overstatement, and also knew that their samples were biased towards men and towards the more educated. In the areas in which they carried out their survey, 20-30% of households had radios, and a further 30-40% listened fairly often to somebody else's radio. From a wider set of EAR studies, it seems that 27% of farmers in their samples had learnt agricultural information from radio programmes. Gausi (1978) found that fewer people owned radios in nine villages of Lilongwe Land Development Programme where he carried out his research, and suggested that radio ownership there was only 10%. This contrasted with an earlier study by Phiri (1975) which suggested that 22.5% of people had radios and 32.5% listened to them. Discussion at MBC suggested that agricultural pro- grammes were very popular. There is of course, a danger that these figures may be distorted by the respondents' wish to be courteous to enumerators from the Ministry of Agriculture, but EAB try to guard against this. All in all it seems reasonable to conclude that in those parts of the countrywhere MBC's signals are clearly received, about 30% of farm families hear some of the Branch's broadcasts. The figures in Table 7 suggest that nearly a million people a year see films shown at one of the mobile units. Driver/operators keep a record of audience numbers, and a check by the EAR unit confirmed the Table 6: Production of films 1972-78 1972 1973 1974 1975 1976 1977 1978 1972-78 No. of films shot 57 58 42 22 45 50 60 48 Total minutes 162 160 169 100 150 167 197 158 Table 7: Mobile units 1972-78 1972 1973 1974 1975 1976 1977 1978 Average 1972-78 -~~~~~~~~~~~~~~~~~~~~~~~~- No. of units 10 11 12 12 12 12 12 a Film shows 1,084 1,200 1,117 1,089 1,833 1,191 1,738 1,322 Total audience 1,189,200 1,236,000 876,059 1,001,535 718,222 889,154 428,154 905,455 Audience per show 1,097 1,030 784 920 392 747 246 685 Audience per van 118,920 112,364 73,005 83,461 59,852 74,096 35,668 78,249 Distance travelled (km) 266,848 311,758 290,253 280,948 271,149 233,903 172,393 261,036 Distance per van (km) 22,685 28,342 24,188 23,412 22,596 19,492 14,366 22,559 -168- accuracy of their estimates. The sharp decline in audience between 1977 and 1978 is unexplained. EAR Unit studies have also shown that there is a considerable variation in the composition of the audience in different villages with the proportion of children varying between 19 and 30%. Their studies suggest that an overall average would be of the order of 32% men, 29% women, and 39% children. While an earlier study suggested that film vans were reaching about 10% of the population, an EAR study at Manyamula in the Mzimba district found that 'about half those respondents living at or near the film show site actually attended the films'. (EAR Unit, n.d., p. 10). A third of the audience here had not been to school and another third had left school at or before Standard 5. It seems reasonable to conclude that the film vans were successful in reaching a large proportion of the population and in teaching poorer as well as wealthier farming families. We can assume that there are advantages for children in learning something of agriculture from the film vans. And, in view of the fact that agricultural extension is directed almost exclusively towards men and not to women, the presence of women in the film van audience is important. Did they learn anything? Studies by the EAR Unit show that many farmers were already familiar with what was being taught by film or radio. In a pre-test of a film to be used in a campaign on growing maize, for example: Practically all the respondents who saw the film had grown maize last season. Comparatively few said that they had learned anything new from the film .... During our tour of the test areas, it became clear that maize stooking is commonly practised so that, to some extent, the film was preaching to the converted. (ibid. p. 4) Similarly, their radio study found that: In the two samples of audiences being urged to prepare their gardens early, 30% and 40% respectively said that they were already doing so. Similarly, 24% and 60% were already stooking their maize, 35% and 32% had used the recommended method of determining when to harvest cotton, 33% had used two bags when harvesting, and 15% had tried Actellic before. (EAR Unit, 1979a, p. 11) When a practice is already widespread, it is impossible to determine just why it has been adopted, or to compare the channels through which information about it has been passed. The EAR Unit drew the conclusion that 'A decision should be made jointly by EAB and the extension service as to whether it is appropriate to continue to preach only these messages when we know that about a limit of the target audience is already converted.' (ibid) There is a further difficulty in choosing the content of EAB's work: crops and agricultural conditions vary from one part of the country to another, so that advice which is appropriate in one part of the country may not be appropriate in another. Difficulties therefore arise for the Branch in running an effective service--and for outsiders in assessing its work--and do so very largely because of the didactic form of much of its work. As the EAR Unit says: -169- Most of the messages in the test programmes were aimed at persuading the target audiences to adopt some practice (e.g. prepare gardens early, pick cotton according to certain rules, apply Actellic). Often these messages were inappropriate. In some cases, they simply did not apply to large sections of the target audience (e.g. the time of the broadcast was wrong or farmers did not use the methods or equipment mentioned). In other cases, the recommendations had already been adopted, often by about a third of the test audience or more. Frequently, too, respondents said that the messages in the test programmes contained nothing they did not already know. (The messages were in these cases apparently redundant.) (EAR Unit 1979b, p. 2) The Unit's work has, however, shown that farmers can, and do, learn both from film and from radio. Tests of a film on an alternative way of planting millet to the traditional slash and burn technique (visoso) found, for example: Two-thirds of the respondents said that they had learned something from this film (a much higher proportion than for the maize film). Of these respondents, a large proportion (nearly two-thirds) said that they had learned about planting on ridges. A much smaller proportion said they had learned about the use of fertilizer. These responses were substantially confirmed by our other findings as will now be seen. An important set of messages in the film is concerned with the modern methods of planting finger millet. The messages are listed below together with the number of respondents who knew them before and then after seeing the film. No. of respondents giving correct message Response/message before film show after film show (a) plant finger millet on ridges 9 48 (b) plant in two drills 2 23 (c) make the drills 6 inches apart 0 3 (d) apply fertilizer 3-4 weeks after germination 1 31 (e) use sulphate ammonia 2 11 (f) apply at the rate of 200 lbs per acre 1 6 (g) rotate with other crops 3 0 (Total number of respondents 52) These figures suggest that the film did have a significant effect where some messages were concerned (although we could not prove this statistically in all cases). In particular, the recommendation that finger millet should be pLanted on ridges (the visoso method involves broadcasting seed on flat beds) was well communicated. Whereas only 9 out of 52 respondents know this before the film, the majority did so afterwards. (EAR Unit, n.d., p. 5) The film was less successful in communicating the more complicated ecological arguments for abandoning traditional methods of cultivation. -170- Similarly, studies of the recall of information in radio programmes showed that 65% of respondents were able to give correct responses to seven questions requiring single responses on a test programme. As one would expect, lower results were obtained on multiple response questions (e.g. 'What reasons did the announcer give for preparing your gardens early?') (EAB, 1979a, p. 9). Few of the respondents could remember all the arguments stated in a programme - a finding which is consistent with other radio studies elsewhere. But these studies, designed to improve the effectiveness of the Branch's materials rather than as part of a media research programme, confirm that the methods can be successful in communicating with farmers. In a number of studies, the EAR Unit has asked farmers where they obtained information about agricultural extension. The results are shown in Appendix 3. A summary of this table suggests the results in Table 8. Table 8: Sources of information to farmers Source Percentage of farmers Extension worker 67 Radio 27 Film 9 Puppets 2 Za Achikumbi 1 Attendance at course 8 Other 14 (Total adds to more than 100 as more than one source mentioned by some respondents). These figures must be treated with caution. If asked a question about the source of information, especially by an enumerator from the Ministry of Agriculture, respondents are likely to mention one or other of its activities as the source. As we will see in the next section, many farmers have little direct contact with extension agents and the infuence of extension workers may well be over-stated by these figures. But, for what they are worth, they suggest that radio is used as a source of information on agriculture by about the same proportion of the population as it is reaching. The mobile cinema vans are reaching about 20% of the population a year, and something like half that number are learning something from the van's programme--a high proportion if we bear in mind that many months or years are likely to elapse before a film van revisits any one village. There is one final point to be made about learning. Generally, we learn most effectively if we enjoy doing so. The EAR Unit studies showed that both radio programmes and the Branch's films are enjoyed. Audiences watch the film shows--and indeed the puppet shows--with interest and concentration. Agricultural films are popular: in their Manyamula study, the unit found that 'the film on football was the best liked entertainment film, but it still took third place to the films on maize and visoso' (EAR Unit, n.d., p. 10). This matches closely with experience in Kenya where a film on the maize stalk borer proved the most popular in a mixed programme (Potts and Troughear, 1977). And it fits with our own observation: the audience at Wimbe seemed to enjoy most an old newsreel of football, to which the driver/operator added his own live commentary in Chichewa, turning down the English commentary on the film, and the agricultural films. -171- We can sum up then that, while films and radio programmes alike are carrying a lot of information which is already familiar to farmers, they can communicate effectively. Studies have shown that farmers do learn from both these media, that both are significant as sources of agricultural information, and that they are popular. A comparison with the work of extension officers We can make some rough comparisons between the work of EAB and the activities of the extension service. Much of the detailed evidence about the work of extension workers comes from the (former) LLDP where there have been in recent years 500 farmers to each extension officer; in other parts of the country the ratio was reported in 1975 as 1:1200-1300 (Lele 1975, p. 66), and more recently as 1:1400-2400 (World Bank 1980, p. 28). On average, for the country as a whole, the figure appears to lie between 600 and 700. We must assume that the proportion of farmers visited by extension agents outside the development project areas is substantially lower than the LLDP figures. Within LLDP, in 1978/79, only 30% of groundnut growers were given extension advice (though 43% had received advice in the previous year) (LADD Evaluation Unit 1980b, table 15), 30% of farmers had attended a day training centre and 3% both, while 61% had not attended one at all (LADD Evaluation Unnit, n.d., table 6). Total annual attendance at day centres has been estimated at 70 000 student days; 7 000-10 000 men and 5 000 to 7 000 women attend a residential centre in a year (World Bank 1980, p. 30). Despite the figures quoted in the previous section, from EAB's evaluation work, we can assume that most farmers have no contact with an extension agent even within LLDP. Outside LLDP, the proportion of farmers having contacts with extension agents is much lower. According to the data from Annex A, around 10% of farmers in the two samples have been visited for the past 12 months. We also have some evidence about the farmers who are in touch with extension agents. First, few women get help from extension agents although, as Uma Lele points out: In many traditional societies, it is the custom for the women to provide the major support for themselves and their children, either by supplying the family with home-grown food, or by obtaining cash through the sale of their own produce. Perhaps an even more important indication of the woman's role in agriculture is the fact that a large percentage of rural households is headed by women. Ag.ricultural extension programmes have frequently overlooked the importance of women, both as major contributors to the farm labour supply and as significant: family breadwinners. (Lele 1975, p. 77) She reports also that in Malawi 35% of rural households are headed by women (ibid). There is also some evidence from LLDP that farmers with larger holdings were more likely to receive credit from the project and to have contact with extension (LADD Evaluation Unit 1979, p. 3; and Annex A). Similarly, those attending courses tended to be farmers with larger holdings and with more education (ibid). As with many other extension services, that of Malawi tends to visit the wealthier rather than the poorer farmers. In the absence of local studies, we cannot estimate the proportion of the time which extension agents spend in touch with farmers, nor the number of farmers they contact. Studies in Kenya, where the -172- extension service was organised on lines very similar to that of Malawi, suggested that extension agents might spend only 2.9 days a week visiting farmers, and that each probably visited fewer than two farm families in a day (Leonard 1977, p. 32). If figures of this kind were applied to Malawi, they would confirm that very many of the country's 850 000 farm families have no chance of being visited by extension agents. And there is little chance of increasing dramatically the number of extension agents: it proved impossible to reach the planned ratios of 1:200 even within the development project areas. Concerning the effectiveness of agricultural extension, there are two studies. An LLDP study found a significant correlation between extension contact and the production of hybrid maize (+ 0.3, significant at the 0.05 level), but no significant correlation for either local maize or groundnuts (LADD Evaluation Unit 1979, p. 1). Interestingly, there was also a correlation between being a woman farmer and yield of both local and hybrid maize (+ 0.3 in both cases, significant at 0.02 and 0.05 levels respectively). This is unlikely to have been the result of extension contact; rather, it is suggested that families with a woman as the head of household are unlikely to have enough labour to cultivate tobacco, so that maize is a more important crop for them than for many households with male heads. In the case of groundnuts, where extension's advice concentrated on ridging, weeding, and the selection of the best variety for cultivation (plant population), Table 9 shows that hardly any farmers needed advice on the first two points; on plant population, slightly less than twice as many farmers who had received extension advice were following recommended practices as were those who had received none. But 86% had received no advice and some of the energy that went into giving advice about ridging and weeding might have been better spent on the question of plant population. Annex A shows that, if we hold constant the amount of land cultivated, contact with extension increases production significantly and increases the probablity that a farmer will grow cash crops (groundnuts, cotton and tobacco) without this being at the expense of subsistence crops. -173- Table 9: Interaction between groundnut farmers and extension agents, LLDP Practice LLDP 1977/78 LLDP 1978/79 Ridging OK, no advice 63 65 OK advice 32 35 Not OK, no advice 0 0 Not OK, advice 0 0 Weeding OK, no advice 89 52 OK, advice 4 47 Not OK, no advice 7 1 Not OK, advice 0 1 Plant population OK, no advice 2 16 OK, advice - 5 Not OK, no advice 90 70 Not OK, advice 8 9 Total observations (n) 728 504 Source: LADD Evaluation Unit (1980a), table 10. There is, therefore, reason to have some reservations about the efficacy of the day-to-day work of extension agents. In his review of the literature on the impact of extension services worldwide, Orivel (1981) shows that, according to most studies, extension agents have contact with a small proportion of farmers, and these farmers produce more only in half of the studies. That does not mean that investment in extension services are not cost-effective (rates of return estimates have rather high values), but that extension services have real difficulties in reaching the mass of farmers. We will look next at the costs, both of EAB and of extension staff, in order to see how far they can be compared in cost, and what policy conclusions one might draw about the use of both styles of extension work. COSTS The Extension Aids Branch organizes a variety of activities and its own accounts do not attribute expenditure activity by activity. Using their 1980 budget, we have therefore calculated the costs on a number of assumptions which we summarize here, before looking in detail at the costs for each kind of work. All costs are divided between current costs, such as staff salaries, vehicle fuel or paper; and capital costs, such as mobile vans or buildings. In order to work out a total annual cost, we need to add capital costs to recurrent. This is done, using standard cost-accounting methods, by calculating the equivalent of an annual rent on capital assets and adding it to the recurrent costs. (For a fuller discussion of this, see Jamison, Klees, and Wells (1978)). -174- We have seen that EAB's activities can usefully be considered under four headings: radio, mobile vans, print, and administration and evaluation. It is then possible to allocate the costs of administration and evaluation to the first three of these, and we do this in table 11 below. Costs have been calculated in Malawi Kwacha, equivalent to US $1.20 in 1980. The EAB budget does not break staff costs down into departments, although we do have the wages for drivers and puppeteers in the mobile units. Assuming that, otherwise, the distribution of grades of staff is much the same in each section, we can allocate total staff costs to the different sections. The total salary bill is K135 000 for established staff and K12 500 for non-established staff. We have allocated 8% of this to the radio section, 40% to the mobile unit, 32% to print, and 20% to administration and evaluation. 2 The headquarters building is 13 650 ft (1 288 m2). At the time of our visit, building costs were K22 per ft2. We have therefore valued the building at 1980 K300 000 and allocated this cost in the same proportions as we have allocated the staff costs. We have used three discount rates to amortise capital costs - 0%, 7.5%, and 15% - and have assumed that buildings have a life of 50 years; vehicles, electronic material and cameras, 4 years; furniture and printing equipment, 10 years. The Cost of radio EAB's broadcasts are written and produced by their own staff, but are then recorded in the Lilongwe studios of Malawi Broadcasting Corporation (MBC) and broadcast through their network of eight- transmitters. We have therefore calculated their costs by taking EAB's costs and adding to them a figure for MBC's transmission costs. MBC had in 1980 calculated that their costs were K840 per hour: this included amortisation of the capital cost of the broadcasting service, staffing and running costs, and notional losses of advertising revenue. It is thus a significantly higher figure than the marginal, or above the line, cost often used in calculating the costs of educational broadcasts. Knowing that EAB broadcasts 232 hours a year, we can now produce broadcasting costs in the form shown in table 10. The Cost of mobile units As we have seen there are three major activities involving the mobile vans: the making of films (and at much less cost the development of puppet plays), the showing of films, and the showing of puppet plays. We have included all these costs in Table 11, but then separated the costs for film shows from the costs of puppet plays on the necessarily arbitrary basis that three-quarters of the costs to be charged are for film shows and one-quarter for puppet shows. We can summarize Table 11 by saying that for a total cost of between K205 000 and 244 000 (US $250 000-300 000), the EAB can run some 12 vans, provide them with films, and show these films to half a million adults a year, while showing puppet plays to three-quarters of a million, and also attracting significant audiences of children. A single mobile van costs between K17 000 and K20 000 a year and can be expected to give some 100 film shows and 300 puppet shows a year. -175- Table 10: Cost of Radio Unit: 1980 Kwachas Recurrent: Personnel 12 000 Sound tapes 600 Narratorq 1 300 Travel a/ 4 200 Miscellaneous 500 MBC broadcasting costs b/ 194 900 Total recurrent cost 213 500 Capital: 0% 7.5% 15% Vehicles 3 470 4 150 5 050 Building 380 1 850 3 603 Recording material / 2 000 2 392 2 800 Total capital cost 5 850 8 392 11 453 Total Yearly Cost 219 350 221 892 224 953 Unit Cost: Per hour 945 956 970 Per farmer-listener per year d/ 0.731 0.740 0.750 Per farmer-listener per hour 0.0031 0.0032 0.0032 a/ Assuming 1/15 of vehicles cost in EAB. b/ Assuming 232 hours a year at K840 per hour. c/ Recording material has a purchase value of K8 000. d/ We have taken a conservative estimate of the audience of 300 000 persons, or about one person in each of 30% of the total number of farming households in Malawi. Thus, we can see that for a total cost of some K219 000 - 225 000 (about US $260 000), Malawi can reach at least 300 000 farmers by radio for 4 1/2 hours a week throughout the year. Of this figure, only about K20 000 - 25 000 appears in the budget of EAB. Table 11; Cost of Mobile Unit Unit: 1980 Kwachas Recurrent. Mobile Units For Films a/ For Puppets Personnel 60,000 45,000 15,000 Cine supplies and processing 30,000 30,000 - Vehicle maintenance 23,300 17,475 5,825 Fuel 35,500 26,625 8,875 Miscellaneous 2,700 2,025 675 Total Recurrent Cost 151,500 121,125 30,375 --------------------------------------------------------------__-------------__------------------------------------------------- Capital: 0% 7.5% 15% 0% 7.5% 15% 0% 7.5% 15% Vehicles 48,530 58,042 70,750 36,400 43,530 53,060 12,130 14,512 17,690 Buildings 2,400 9,249 18,017 1,800 6,937 13,513 600 2,312 4,504 Cameras 2,500 2,990 3,500 2,500 2,990 3,500 Total Capital Cost 53,430 70,281 92,267 40,700 53,457 70,073 12,730 16,824 22,194 Total Cost 204,930 221,781 243,767 161,825 174,582 191,197 43,105 47,199 52,569 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -_ _ - - - - - - -_ _ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Unit Costs; per van b/ 17,077 18,482 20,314 13,485 14,848 15,933 3,759 3,933 4,380 per show c/ 135 145 159 12 13 15 per adult viewer d/ 0.30 0.32 0.35 0.057 0.063 0.070 per viewer 0.18 0.19 0.21 0.029 0.031 0.035 per adult viewer per hour e/ 0.12 0.13 0.14 n.a n.a n.a Notes. a/ Where costs cannot readily be attributed to films or puppets, we lhave assumed that 3/4 of the costs should be attributed to films and 1/4 to puppets. b/ These figures are based on the experience of the 12 vans directly controlled by EAB, so the costs are divided by 12. c/ Assuming each van does 100 cinema shows and 300 puppet plays a year. At 400 plays a year, as more recent information from Malawi suggests, the cost at 7.5% falls to K1O per show. d/ This is based on an average of 900,000 viewers of films per year, of whom 540,000 are adults, and 1,500,000 watching puppet plays, of whom 750,000 are adults. e/ For an average film show of 2.5 hours. We have no figures for the average length of puppet shows. -177- The cost of print The printed materials produced by EAB are distributed both to farmers and to extension workers, which makes it more difficult to calculate a unit cost. Text books, bulletins, booklets and circulars go mainly to extension agents, but the magazine Achikumbi goes to 2 000 extension staff and to 30 000 farmers. If we assume that half the costs should be attributed to these 30 000 farmers and the other half to the extension services, then we need to add this latter half to the costs of the extension services (see Annex A). The unit cost for each literate farmer is then between K2.00 and K2.40 while the annual expenditure on printing is between K120 000 and 140 000 (US$140 000-170 000). The costs are set out in Table 12. Table 12: Cost of Print Unit: 1980 Kwachas Recurrent: Personnel 47 000 Paper 30 000 Print 12 000 Ink, chemicals 10 000 Stationery 12 000 Photo 5 000 Miscellaneous 2 200 Total Recurrent Cost 118 200 Less Revenue 12 000 Net Total Recurrent Cost 106 200 Capital: 0% 7.5% 15% Printing machines a/ 3 600 5 256 7 164 Building 1 920 7 399 14 413 Total Capital Cost 5 520 12 655 21 577 Total 111 720 118 855 127 777 attributed to: 30 000 farmers (1/2) 55 860 59 427 63 888 extension staff (1/2) 55 860 59 427 63 888 Unit cost for 30 000 farmers 1.86 1.98 2.13 a/ Replacement cost has been estimated at K12 000 per machine. -178- The cost of evaluation and administration The remaining costs are for administration and evaluation and are set out in Table 13. Table 13: Cost of Administration and Evaluation Unit: 1980 Kwachas Recurrent: Personnel 29 000 Miscellaneous 1 400 Capital: 0% 7.5% 15% Furniture 3 000 4 380 5 970 Building 1 200 4 624 9 008 Total 34 600 39 404 45 379 If we share these costs between the other activities of EAB, using the same proportions as we did for the staff costs, then we increase the costs for each activity as shown in Table 14. We can then summarise the figures by saying that for a total annual sum, including rent on the capital and a full share of the costs of braodcasting incurred by MBC of between K575 000 and 645 000 (US $680 000 - 775 000), the EAB: - can reach about 540 000 adults with films and 750 000 with puppet plays; - can provide 4 1/2 hours broadcasting a week which reaches nearly one farmer in three; - can provide printed support material to some 30 000 literate farmers, and to extension services throughout the country. -179- Table 14: Costs including Charge for Administration and Evaluation Unit: 1980 Kwachas Addition for Eva- luation & Rate Adminis- New Total New Unit % Raw tration a! Cost Cost Radio 0 219 350 3 460 222 810 0.0032Ž/ 7.5 221 892 3 940 225 832 0.0032 15 224 953 4 538 229 491 0.0033 Mobile Unit - 0 204 930 17 300 222 230 - Total 7.5 221 781 19 702 241 482 - 15 243 767 22 689 266 456 - Mobile Unit - 0 161 825 12 975 174 800 0.129-/ Films 7.5 174 582 14 776 189 358 0.140 15 191 198 17 017 208 215 0.154 Mobile Unit - 0 43 105 4 325 47 430 0.063_/ Puppets 7.5 47 199 4 926 52 125 0.069 15 52 569 5 672 58 241 0.078 Print 0 111 720 13 840 125 560 2.0929' 7.5 118 855 15 762 134 617 2.24 15 127 777 18 152 145 929 2.43 Total 0 536 000 34 600 570 600 (all media) 7.5 562 528 39 404 601 932 15 596 497 45 379 641 876 Notes: a/ 10% radio, 50% mobile unit, 40% print. b/ Per farmer per hour. c/ Per adult per hour. d/ Per adult member of audience. e/ Per literate farmer receiving the material. -180- Comparison with extension services If we take as an example of cost of extension services in Malawi those of the Lilongwe Agricultural Development Division, which counts 169 field assistants and 40 farm home instructresses, we can compare the costs of EAB activities with those of extension visits, as in Table 15. Table 15: Cost of extension services in LADD a/ Unit: 1980 Kwachas Recurrent: Personnel 320 000 Vehicles' operation 72 000 Building maintenance 27 000 Miscellaneous 17 400 Total Recurrent Cost 436 400 Capital: 0% 7.5% 15% Building 58 000 221 000 432 000 Vehicles 4 000 4 780 5 600 Total Capital Cost 62 000 225 780 437 600 TOTAL COST 498 400 662 180 874 000 Unit cost (per field agent) b/ 2 384 3 168 4 181 Unit cost/with printed material (on the basis of 2 000 agents receiving this material) 2 415 3 202 4 218 Notes: a/ Detailed explanations of costs appear in Annex B. b/ There are 209 field agents. As we have no reliable data on the number of visits paid to farmers by each agent during a year, we have done a cost simulation for the range of 1 hour visits between 50 and 500; and this is shown in table 16. Leonard (1977, p. 32) found that in the, broadly similar, agricultural service in Kenya, extension agents visited 20.3 farmers a month. Assuming the average visit took 1 1/2 hours, this suggests that a figure of between 300 and 400 contact hours would be realistic. -181- Table 16: Cost of Extension Visits Unit: 1980 Kwachas Discount Rate/ No. of Visits 50 75 100 200 300 400 500 0 47 31 24 12 8 6 5 7.5 63 42 32 16 11 8 6 15 84 63 42 21 14 10 8 We can summarise the costs of both EAB and extension services generally, which we do in table 17. Table 17: Agricultural Information Services: Cost Summary Unit: US $ Annual Cost Service Total Per Farm/ As % Family - of Total 1. Agricultural Extension b. 4 970 000 5.230 77.9 2. Farmer TraininV Centre (Residential) - 517 000 0.544 8.1 3. Farmer Training Centre (Day) bl 270 000 0.284 4.2 4. Radio (EAB) 265 000 0.279 4.2 5. Mobile vans (EAB) Movie 209 000 0.220 3.3 6. Mobile vans (EAB) Puppet 81 000 0.085 1.3 7. Printed material (EAB) 71 000 0.075 1.1 Total 6 383 000 6.719 100.0 Notes: a/ Assuming 950 000 farmers. b/ See Table 1 and Annex B. -182- CONCLUSIONS EAB is a well established part of the extension service of Malawi. It is no longer an experimental project. Thus, it has a wealth of experience which makes it possible to draw firm conclusions about its work and its significance for agriculture in Malawi. Further research on its effects would be possible and would be assisted by the variety and high quality of the other research work which is going on in the areas of agricultural extension and economics in Malawi. In the meantime, we can regard as robustly established three major conclusions: EAB succeeds in contacting farmers at a very low cost in comparison with orthodox extension methods; it has found that radio is especially cheap; farmers do benefit from EAB's activities. We have seen that the cost per contact hour is much lower for EAB's films, puppet plays or radio broadcasts than for contact with an extension officer. The audiences for EAB's mobile vans include a significant proportion of women and there is no reason to believe that the poorer farmers are excluded from the audience. We can also assume that there are significant numbers of women who hear their radio programmes, although we do not have figures for these. If we use a 7.5% interest rate, the costs per contact hour are: Radio KO.0032 ($0.00384) Mobile unit - films KO.140 ($0.168) Mobile unit - puppets KO.069 ($0.0828) Extension agents K8-Kll ($9.60-13.20) Thus, the mobile unit is 44 times as expensive as radio, while extension agents cost 55 to 80 times as much as the mobile unit. Extension services are 2 500 to 3 500 times as expensive as radio. Despite these dramatic differences in costs, there is evidence that farmers learn from radio and from mobile vans. While (cf. table 8) 67% of farmers may have used their extension agent as a source of information, 27% reported that they used the radio programme. Even if there were severe limits on the amount of information derived from radio programmes, it is impossible to resist the conclusion that the broadcasts were highly cost-effective. The importance attached to radio by farmers parallels findings elsewhere. Yazidu (n.d., pp. 47-8) found in 1971, for example, that nearly half the farmers he surveyed in northern Nigeria regarded radio as the most important source of agricultural information. It is important to qualify this conclusion. We would not argue for abandoning the use of mobile vans, simply because radio proved cheaper. There are theoretical and practical arguments for using a variety of media, rather than just one. While print, for example, takes a very large proportion of the resources of EAB, the background information it carries about radio programmes and about the activities of mobile vans contributes to their effectiveness. We believe, too, that there is scope for increasing the effectiveness of all media by EAB. The Branch has suffered from a degree of isolation and it seems that its work could be more tightly related to the work of extension agents in the field. This would entail the development of closer working relationships both at Lilongwe and in the field. We also sensed a degree of separation between EAB and the work of agricultural research. But the profoundest change which we would suggest to EAB is one of philosophy. As we saw, much of their output consists of direction to the farmers--instruction on what they should do. We were more than once assured that the Malawi farmers no longer needed instruction of this kind- -183- -if they ever did--and there is now scope for a more sophisticated approach by EAB which is concerned more broadly with the education of farmers and less narrowly with instructing them. For that change to be effective, closer links between EAB's mass media and the work of extension agents, or other ways of grouping farmers together, would be of particular value. -184- APPENDIX 1: Main print jobs 1978 Title No. of copies printed Text books: Guide to Agric. Production 1978/79 2 500 A Glossary of Research Management Terms 2 000 Department of Extension and Training 1973 Annual Report 500 The Livestock Husbandry Techniques 2 500 Bulletins: National Rural Development Programmes 4 500 Extension Aids Branch 1958-1978 5 000 Research Bulletin No. 2/72 3 000 Booklets: Tea Farmers Guide and Calendar 600 Malangizo kwa Oyang'anira Nkhalango 8 000 Shire Valley Agricultural Dev. Project 1 000 Colby College of Agriculture 200 Radio and Mobile Unit Programmes 1 100 Birds at Lilongwe Nature Sanctuary 1 000 Extension Aids Branch Work Calendar 50 Magazine: Za "Achikumbi" magazine (6 issues) 96 000 Inserts for "The Fields" Magazine 64 000 Posters: Art competition for Environmental Educ. Unit 250 "Bzalani Mitengo" for L.L.D.P. 500 "Mitengo ndi chuma" for L.L.D.P. 500 Reprint of Flannelgraph pictures 1 000 Cards and Certificates 24 000 Agricultural show kits 110 Circulars: Grow More Cassava 3 000 Growing Wheat in Malawi 3 000 Strawberries (Revised Edition) 3 000 Citrus (Revised Edition) 3 000 Soya Bean Recommendation 3 000 Interim Research Recommendation for Oriental Tobacco 2 000 Prevent Tick Diseases by Improved Dipping 3 500 Handouts and leaflets 151 000 Forms in general 300 226 Other (Maps, covers, lettering, etc.) 942 422 Source: EAB Annual Report 1978. -185- APPENDIX 2. Mobile -ans *chedule - northern rexion 1980 POGORAM SUBJECT MATTER DATE VEHICLE LOCATION BY DATE =NBMER VEHICLE I VEHICLE 2 VEHICLE 3 VEHICLE 4 I Oriental tobso-d harvesting, coring March 10th-14th Misba asitu/ Chitipa Karonga South and uprooting otalk. (Niaba and North Pereke Sooth Ruvphi Went). Cotton spraying (ifenga Valley). March 17th-21st Rumphi Euthini Chitipa Karonga Centrl. Cotton weeding and spraying (taronog). West Central Rice weeding, fertilizer application March 24th-28th Henga Bulala Chitipa Kaporo Sooth and birds caring (taronga). Valley North 2 Maise atnoking and harvesting. April 14th-18th Nkhata Euthini Chitip. Karonga South Greonadots harveetin.g Bay South/ South Li.phasa Schebe Birds Scaring and harvesting April 21st-25th Hkhata Maiba Chitipa Karonga Central rice (Nkhhata ay and X-ronga). Hay North South Central Weet Coffee weeding, fertilizer April 28th- Viphys Mzimba Chitipa Kaporo North application and poeat and May 2nd North South North dineane. control (Khkata Hay East North, Phola Hill., Misukh Hill. and Viphya North). Cotton picking and grading May 5th-9th Phoka aimba hMisukn Kaporo South (Henga Valley). Cotton spraying HiLls South Hill. (Karonga). Oriental tobacco cring and, uprooting and burning otalks (Ruophi Went and Mambo) 3 Oriental tobacc grading and Hay 19th-23rd Henga Maimba Chitipa Earonga Central Markoting (RHrphi Went and Maibo). Valley Central South Maine harvesting, elating and May 26th-30th Rumphi Kasitu/ Chitips Kaporo Sooth marketing, troand.dts shelling, West Perekazi Central grading and narketing. Cotton grading, aarketing and June 2nd-6th Maibs Bulala Chitipa karonga South otalkb apreoting (H.nga Valley). North North Cotton pitking, grading and etalka uprooting (taronga). 4 Livestock nanagenent. Ox-training. June 23rd-27th Phoka Viphys Misok. Kaporo North Fertilieer and improved seed buying. Hills North Hills Credit recovery. Coffee nursery preparation and Jlne 30th-July 4th Henga Nkhata Bay Chitipa Earonga Central sowing. Cotton stalks uprooting Valley North North and burning (Henga Valley). Cotton picking, grading and stalks uprooting (taronga). 5 Winter Rice weeding and fertilise Aug. lIlth-lSth Nkhata Bay Euthini Chitipa Saronga South application (Karonga and Nkhate Bay). South/ Sooth Limphasa Coffee harvesting, grading and sell- Aug. 18th-22nd Nkhata Nay Sulala Chitip Ka ronga Central ing, Coffee tonic spraying. North Central Cotton stalk4 uprooting and burning, Aug. 25th-29tb phoke Kteitu/ Hisuk. taporo Sooth cotton grading and marketing (Karonga). Hills Perekesi Hills Garden preparation. Livestock manage- ment. Credit recovery. Cashew nots - nurery preparation and sowing (Lakeshore). 6 Garden preparation. Manure application. Sept. 15th-19th Hahata Bay Viphys Chitipa Kaporo North Dangers of Vitoso. Livestock manage- South/ North South ment. Limphasa Rise nare..y prep.eatisn and swing Sept. 22nd-24th Nkhata Bey miumbs Hisoko Xaporo Snkth (Karonga and NkhataBay). North North Hills Coffee pruning, digging and filling SNpt. 29th-IXt.3rd Phoka Maibs Chitipa Ktarong Sooth coffee planting holes (Phoks Hills, Hills Gentral North Nisuks Hills, Nkhata Nay North snd Viphy. North). Cashew neta nursery preparation and seed sowing (Keronga and Nkhata Nay). 7 Maine mad groadndsta planting, weding Oct. 20th-24th Nkhata Bay Maimbe Chitipa keronga South and fertiliszr application. South/ South South Limphasa Orieatal tobacco naurery preparation Oct. 27th-31st HNnga Maimba Misuku Karong. Ceintral and seed cowing (Eamphi West and Valley South East Hills MaimSa). Winter rice harvesting, *vmr ricg Nov. 3rd - 7th Ramphi Maimba Chitipe Kaporo North and garden preparation (taronga and W et South Eact Central South Nkh.t. Na. Cotton planting and weeding (HNnga Valley). APPENDIX 3: Sources of a§ricultural information for farmers Study Sample Sex(%) Education (%) Literacy (%) Source of agricultural information (%) Total M F No. Std. Std. Std. 0th. reads with not at Ext. Radio Films Puppets ZA Course Other Sch. 1-3 4-5 6-8 well diff. all wkr. Michiru mountain park campaign KAP study (a) 260 41 59 31 17 18 23 13 29 16 54 33 54 4 0 0 11 Forestry campaign: KAP study, LLDP 229 76 24 52 - 40 - 8 0 N/A 40 22 - - - 5 33 Forestry campaign: leaflets pre- test, LLDP 309 49 51 60 25 9 6 0 16 17 67 79 6 - 5 - 11 15 Crop storage campaign: radio pre-test, SVACP 166 64 36 38 45 10 5 1 N/A 64 - 16 1 1 1 - Radio survey: Balaka (b) 166 71 28 49 -_36 - 14 0 N/A (43) (36) (12) (2) (1) (8) (11) Radio survey: LRDP (b) 162 68 32 52 -40 - 7 1 N/A (51) (18) (20) (4) (0) (1) (2) Radio study: Karonga 56 71 29 41 38 7 14 0 N/A 89 29 23 4 2 7 18 Film pre-test: Manyamula (c) 172 (73) (27) (15) (21) (15)(46) (2) N/A 74 26 25 6 6 12 17 (100) Training exercise: Nambumba 151 56 44 42 15 19 20 1 26 28 45 83 32 3 1 5 9 22 Notes ZA = Za Achikumbi Source: CAR Unlt. 1980 (a) sample included an urban area (Chilornoni) (14) % in brackets refer only to information about early garden preparation (c) h in brackets refer to sub-sample only; i. e. those who were interviewed both before and after the films -187- ANNEX A EXTENSION, EDUCATION AND FARMERS' PRODUCTIVITY IN MALAWI This paper examines the functioning, costs, and outreach of the Extension Aids Branch of the Ministry of Agriculture; it does not, however, assess the ultimate impact of EAB's activities on productivity. Unfortunately, data that would allow such an assessment are not available. To form some idea about EAB's potential impact and to reach a better sense of the efficacy of the extension service, we have, however, analysed available data. This Annex reports the results of that analysis. It begins by describing the data we use. It then analyses separately, for each of two samples, the relations between farmers' characteristics and their exposure to extension; then the relations between extension and education on the one hand, and crop adoption and productivity on the other, are analysed. Finally, we examine two key questions: are educated and/or visited farmers wealthier, and are non-traditional crops grown at the expense of subsistence crops? Data The data analysed in this Annex were collected by the Ministry of Agriculture's Agro-Economic Survey in 1979 and give information on the 1978 crop. Data on two samples of 150 farmers were randomly selected (by geographical region from the larger data sets). These farmers are smallholders. The majority of them are growing maize. In the first sample, a minority of farmers is growing groundnuts and/or cotton, and in the second one, ground-nuts and/or tobacco. For each farmer, we know the size of the household, the holding size, the number of years of education of the head of the household, the size of land cultivated for each crop, the output of each crop, and the number of visits of extension agents for each crop. Table Al below provides information on the characteristics of the farmers in the two samples. -188- Table Al: Main Characteristics of Farmers Sample 1 Sample 2 Standard Standard Mean Deviation Mean Deviation Holding sizel 134 84 134 76 Household s,ize 4.15 2.16 4.45 1.84 Education.'/ 0.54 0.82 0.94 0.96 Maize plot1/ 90 0.56 87 48 Maize crop-/ 1 238 1 100 1 072 928 Groundnuts plot!' 10 8.3 31 28 Groundnuts i op- 46 37 172 159 Cotton plot- 63 41 - - Cotton crop3 582 481 - - Tobacco plot - - 35 21 Tobacco crop- - 292 264 Extension (No. of visited farmers 22 15 (No. of visits 34 25 Regions 6 (25 farmers in each) 6 (25 farmers in each) 1/ In hundredths of ha. The national average is 166. 2/ Number of years of education. 3/ In kilos. I. SAMPLE I A. Extension, Education and Choice of Crop - Extension and Education Table A2: Extension and Education- / Unit: No. of farmers Extension - 0 Extension > 0 Total No. % No. % No. % Education 0 85 89 11 11 96 64 Education > 0 43 80 11 20 54 36 Total 128 100 22 100 150 100 1/ Education is measured by the number of years of education. Extension is measured by the number of visits of extension agents to the farmer. Both education and extension are considered as > 0 when the number of years or visits is 1 or more. -189- 15% of the farmers have been visited by an extension agent at least once. 36% of the farmers have one or more years of education; in practice, it is never more than 3 years, and usually only one or two. The educated farmers are twice as likely to be visited by an extension agent as those who are uneducated (20% against 11%). - Education and Non-traditional Crops Subsistence agriculture is characterised by the production of maize, the main source of food for a farmer's family. Farmers who produce groundnuts have usually two objectives: a diversification of the family diet, and the marketing of part of the crop for cash. We can therefore regard the cultivation of crops other than maize as evidence of innovative behaviour. Farmers who produce tobacco or cotton have a market objective. We have looked at whether innovative behaviour is positively connected with education and/or extension. Table A3: Education, Groundnuts and Cotton Unit: No. of farmers Education ---Yes--- ---No--- --Total-- No. % No. % No. X Grow groundnuts Yes 8 15 8 8 16 11 No 46 85 88 92 134 89 Total 54 100 96 100 150 100 Grow cotton Yes 25 46 25 26 50 33 No 29 54 71 74 100 67 Total 54 100 96 100 150 100 Grow groundnuts Yes 6 11 2 2 8 5 and cotton No 48 89 94 98 142 95 Total 54 100 96 100 150 100 11% of the farmers grow groundnuts, but this proportion reaches 15% among educated farmers and 8% among uneducated farmers (the probability is 2/3 higher). 33% of the farmers grow cotton, but here again, the educated farmer is twice as likely to grow cotton. Only 5% of the farmers grow both cotton and groundnuts, but the educated farmers are 6 times more likely to do so than are the uneducated. -190- - Extension and Non-traditional Crops Table A4: Extension, Groundnuts and Cotton Unit: No. of farmers Extension ---Yes--- ----No--- --Total-- No. % No. % No. % Grow groundnuts Yes 5 23 11 9 16 12 No 17 77 117 91 134 88 Total 22 100 128 100 150 100 Grow cotton Yes 22 100 28 22 50 33 No 0 0 100 78 100 67 Total 22 100 128 100 150 100 Grow groundnuts Yes 5 23 3 2 8 5 and cotton No 17 77 125 98 142 95 Total 22 100 128 100 150 100 23% of the farmers who have been visited by extension agents grow groundnuts, but only 9% of those not visited do so. All the farmers who have seen an extension agent grow cotton, whereas only 28% of the others do. Finally, 23% of the visited farmers grow both groundnuts and cotton, and only 2.3% (ten times less) of the non-visited do. On the whole, educated farmers grow non-traditional crops almost twice as often as non-educated farmers; farmers visited by extension agents cultivate non-traditional crops four times as often as non-visited farmers; and educated farmers are twice as likely to be visited by an extension agent as uneducated farmers. Although education is less efficient than extension in generating innovative behaviour among farmers, it has, in addition to a direct effect, the indirect effect of increasing the probability of meeting extension agents. The effect of education is more likely a causal one for extension, where, particularly given extension agents' instructions on whom to visit, it is plausible that much of the association between innovative behaviour and extension contacts results from the agents choosing to visit the better farmers. B. Production Function of Maize Given the absence of a good labour indicator, we could not examine substitution effects between land and labour, and in this case, the output tends to be linearly related to the size of the plot. For this reason, we have run linear regression, and not Cobb-Douglas functions (using the log of the variables). -191- Table A5: Determinants of Maize Production Regression Variable Coefficient t Land 14.02 11.75* Education -5.8 0.07 H/H size 36.2 1.15 Extensioy 313.2 3.45* Region 2_/ -91.1 0.42 Region 3 421.2 1.97* Region 4 -153.3 0.63 Region 5 658.5 2.70* Region 6 241.8 1.10 C2 nstant -390.0 R 0.60 1/ Each region is compared to region 1 which was taken as reference and deleted from the regression. * Significant at 5% level or better. Land: Land is, of course, by far the most significant variable in explaining maize production. On average, each hundredth of ha increases the production by 14 kg. Household size: Has a positive but not significant effect, although it is close to significance. This variable could have been considered as a proxy for labour. Region: Two regions have a highly significant advantage over the others (421 and 658 more kg). This variable can be considered as a proxy for environment (quality of soil, irrigation and water availability, climate). Extension: According to this regression, each visit of extension agents increases the production by 313 kg, in an undoubtedly significant way. The yearly benefit of a visit is then approximately K60 (for a KO.2 per kg price of maize), which we can compare with the cost of a visit: between K5 and K60 (see EAB Cost Study). But as we have seen above, this may simply be the result of the bias of extension agents in visiting the better farmers. Education: Education has an indirect effect rather than a direct one: it increases the probability of meeting extension agents (which may increase the productivity), and it increases the probability of growing crops. In this sample, the number of years of education is low (between 1 and 3 years, never more). This suggests that, to have a direct effect, education needs to be more than 1-3 years, perhaps a minimum of 4-5 years. 11. SAMPLE II A. Extension, Education and Choice of Crop Table A6 shows the relations, in sample II, between multiple crop choice and farmers' education and extension exposure. -192- Table A6: Extension, Education, and Number of Farmers Cropping Groundnuts and/or Tobacco Unit: No. of farmers Extension ---Education-- - No -Yes-- -Total- No - -Yes-- -Total- No. X No. % No. % No. % No. % No. % Grow Yes 88 65 11 79 99 66 37 57 62 73 99 66 ground-nuts No 48 35 3 21 51 34 28 43 23 27 51 34 Total 136 100 14 100 150 100 65 100 85 100 150 100 Grow Yes 37 27 8 57 45 30 16 25 29 34 45 30 tobacco No 99 73 6 43 105 70 49 75 56 66 105 70 Total 136 100 14 100 150 100 65 100 85 100 150 100 Grow Yes 25 18 7 50 32 21 10 15 22 26 32 21 ground-nuts No 111 82 7 50 118 79 55 85 63 74 118 79 and tobacco Total 136 100 14 100 150 100 65 100 85 100 150 100 Extension: Less than 10% of the farmers have been visited by extension agents. Nevertheless, those who have been visited tend to grow groundnuts slightly more often (79% versus 65%), to grow tobacco twice as often (57% versus 27%), to grow both groundnuts and tobacco three times as often (50% versus 18%). Education: A relatively high proportion of the farmers, 57%, have received some education (between 1 and 3 years). Educated farmers grow groundnuts more often (73%) than the uneducated (57%). Likewise, they grow tobacco (34% versus 24%), and both crops more often (26% versus 15%). The probability for the educated and visited farmers to grow non- traditional crops is higher, as in sample I, but the gap is slightly smaller in this case. But in comparison with sample I, the proportion of educated farmers is much higher, and so is the proportion of farmers growing groundnuts. The lower effect of education can only be the consequence of the law of decreasing returns, or statistically speaking, the fact that the curve of farmers growing groundnuts is asymptotically reaching the ceiling of 100%. B. Extension and Education Table A7: Extension and Education Unit: No. of farmers Extension - 0 Extension > 0 Total No. % No. X No. X Education - 0 62 46 3 21 65 43 Education > 0 74 54 11 79 85 57 Total 136 100 14 100 150 100 -193- Although the number of farmers visited by extension agents is low in this sample, educated farmers have a substantial advantage since 12.9% of them are actually visited, compared to 4.6% for the uneducated. C. Production Function for Maize We have run the same kind of regression used for sample I, with the same independent variables. Table A8: Determinants of Maize Production Regression Variable Coefficient t Land 11.7 8.94* Education 21.2 0.35 H/H Size 49.7 1.41 Extension 217.4 1.81 Region 2 325.1 1.66 Region 3 1047.3 5.65* Region 4 401.6 2.13* Region 5 433.6 2.14* Region 6 422.7 2.06* C2 nstant -648 R 0.54 * Significant at 5% level or better. The results are very similar to those of sample I. The significance of extension is slightly lower, but we must remember that few farmers were visited in that sample. Regions 2, 4, 5, and 6 have the same pattern. Region 1 produces significantly less than this group, and region 3 significantly more. D. Production Function for Groundnuts- Table A9: Determinants of Groundnuts Production Regression Variable Coefficient t Land 4.7 9.10* Education -3.9 0.17 Extension 50.5 0.66 H/H Size -0.24 1.25 Region 2 124.4 3.27* Region 3 179.3 5.03* Region 4 74.2 1.94* Region 5 59.8 1.52 Region 6 143.2 3.70* C2nstant -38.5 R 0.61 * Significant at 5% or better. -194- Here again, the most statistically significant variables are the size of the land cultivated for this specific cyop and the regions (here, all are more productive than region 1). The R is in the same range as in the first 2 regressions (0.61). Education has a negative but not significant effect, and extension has a strongly positive but not significant effect. Extension, in this sample, is scarce, and there are not enough observations to generate statistically significant effects. III. SPECIFIC QUESTIONS A. Are educated and/or visited farmers wealthier? Table A10 below indicates to what extent the educated and visited farmers have a holding size above the average (two samples together). Table A10: Education, Extension, Holding Size Unit: No. of farmers Holding Size 4/ < 1.34 ha > 1.34 ha Total Education = 0 111 50 161 Education > 0 71 68 139 Total 182 118 300 Extension = 0 174 90 264 Extension > 0 8 28 36 Total 182 118 300 1/ 1.34 ha is the average holding size in both samples. 51.1% of the educated farmers are below the average size and 48.9 above, but 68.9% of the uneducated are below. We can conclude that educated farmers are slightly better off, but we would qualify this difference as small. The difference is much greater among farmers exposed or not exposed to extension services. 77.8% of the farmers visited by extension agents belong to the group above average. Only 4.4% of the farmers belonging to the group below average have been visited, versus 23.7% of the others. Extension services are obviously more discriminatory than education services, and as we have seen earlier, the probability of visits increases with education.- Therefore, we can conclude that, in this context, education has a positive effect in reducing inequalities. It is also possible that the positive effect of extension agents visits on productivity is simply due to the fact that visited farmers are on average "better", and would have produced more, even without visits. Nevertheless, it remains true that education is not linked with wealth, and that it increases the probability of growing non-traditional crops and the probability of meeting extension agents, and therefore, of increasing -195- maize production. B. Are non-traditional crops grown at the expense of subsistence crops? It has often been observed that cash crops are developed at the expense of subsistence crops and can have negative food and nutritional effects on farmers' families. The data below do not support this view. On the contrary, the production of maize is constantly increasing when the farmer adds non- traditional crops, and reaches a maximum level when he grows everything. Nevertheless, the holding size follows the same pattern; it is lower among farmers growing only maize and larger among farmers growing non- traditional crops. But the maize plot is more or less similar in the four groups of farmers. Table All: Maize Production of Different Groups of Farmers Average Average Yield No. of Av. Crop H. Size H. Size per ha Farmers of Maize (Total) Maize Plot (in kg) Grow Maize alone 126 1 025 1.02 0.82 1 250 Maize + Groundnuts 109 1 318 1.63 0.67 1 970 Maize + Cotton/ Tobacco 91 1 456 1.77 0.83 1 750 Maize + G-nuts + Cot./Tobac. 40 1 905 1.99 0.94 2 026 -196- ANNEX B EXPENDITURE AND OOSTS FOR AGRICULTURAL EXTENSION AND FARMER CENTRES Table B1 provides a cost estimate of agricultural extension in a specific project, the Lilongwe Agricultural Development Division. We have calculated the unit cost of an extension agent in this project, and extrapolated this unit cost to the 1 388 field assistants working in the country. We have then allocated half the cost of the print division, in EAB, to the extension services. We have also allocated 100 of the 1 388 field assistants to the implementation of the day centres and of the residential centres. Table Bi: Costs of Agricultural Extension (From Lilongwe Agricultural Development Division) Unit: 1980 Kwachas Estimated costs of supporting 169 field assistants and 40 farm home instructresses on the ground (1980/81): 1. Operating Costs: (i) Personal Emoluments K320,000 (of which field assts. & FHIs are K148,350) (ii) Vehicle operations K 72,000 (of which bicycle allowances K6,960) (iii) Building maintenance K 27,000 (iv) Miscellaneous K 17,400 Total R(436,400 2. Capital Costs: 0% 7.5% 15% (a) Vehicle replacement and minor equipment 4,000 4,780 5,600 (b) Building costs: (i) Housing for field assistants and FHIs: 209 dwellings @ K7,000 - K1,463,000 (ii) Housing for area and unit supervisors: 46 DH8 or DH6 @ K22,100 = K1,016,000 (iii) Hoasing for project supervisor (CTO): 1 CH10 Q K29,300 (iv) Office buildings: 40 @ K9,000 - K360,000 Total building costs = K2,868,000 Annualized cost, assuming 50-year building life 58,000 221,000 432,000 TOTAL ANNUAL COST 498,400 662,180 874,000 TOTAL ANNUAL COST, per field worker (169 field assistants plus 40 FHIs = 209 total) 2,384 3,168 4,181 -198- Table B2: Total Cost of Extension Services 0% 7.5% 15% Unit cost of extension agent x 1288 3 506 000 4 080 000 5 385 000 Cost of EAB printed division x1/2 55 860 59 430 63 900 Total 3 561 860 4 139 430 5 448 900 Number of Visits to Farmers There are two categories of extension coverage in Malawi: projects like the Lilongwe Agricultural Development Division which have "intensive" extension activities, and the standard low level of coverage, as in the two random samples analysed in Annex A, where 11% of farmers have been visited in averaze one and a half times. We have estimated that 15% of farmers benefit from an intensive coverage and 85% a non-intensive coverage. If we take a total number of 950 000 farmers, [85% x 950 000] x 11% x 1.5 visits = 133 000 visits [15% x 950 0001 x 50% x 1.5 visits = 107 000 visits (considering that 50% of the farmers in regions with intensive coverage are visited). The total number of contacts is then approximately 240 000, for 1 288 field agents who make around 185 contacts a year. -199- Table B3 provides a cost estimate of both residential and day centres. Table B3: Cost of Farmers Centres (day and residential) 1. Residential - Recurrent (miscellaneous): K1.15 per student day x 100 days x 30 students = 3 450 x 30 centres = 103 500 0% 7.5% 15% - Personnel (30 full-time extension agents equiv.) 71 520 95 040 125 430 - Capital (30 centres at KIOO 000 each, amortised over 50 years) 60 000 231 000 450 000 Total Residential 235 020 429 540 678 930 2. Day Centres-i- - Recurrent de ISBNi 0-8213-0028-8. $3.00. of Agriculture and Rural Prices, Taxes, and Subsidies ISN02308.$0. Development Projects In Pakistan Agriculture, Dennis J. Casley 19601976 SocioculturalAspects of and Denis A. Lury Carl Gotsch and Gilbert Brown Developing Small-Scale This book provides a how-to tool for World Bank Staff Working Paper No. Fisheries: Delivering the design and implementation of 387. April 1980.108 pages. richr to Phelnac monitoring and evaluation systems in Stock No. WP-0387. $5.00. Richard B. Polnac rural development projects. Because World Bank Staff Working Paper No. rural development projects are com- 490. October 1981. iii + 61 pages plex. they seek to beneflt large num- (including references). bers of people in remote rural areas, Stock No. WP-0490. $3.00. Some Aspects of Wheat and Credit and Sharecropping In Agrarian Rice Price Policy In India Societies Raj Krishna and 0. S. Avishay Braverman and TN'1. Srinivasan Raychaudhuri World Bank Reprint Series: Number 216. Reprinted from Journal of Development World Bank Staff Working Paper No, Economics, vol. 9 (December 1981): 289-312. 381. April 1980. 62 pages (including Stock No. RP-0216. Free of charge. 2 appendixes, 6 tables, bibliography). Farm Size and the Diffusion of Green Stock Nlo. WP-0381. $3.00. reolution Technology On Information and Innovation Diffusion: A Bayesian Approach A System of 1Monitoring and Gershon Feder and Gerald T. O'Mara Evaluating Agricultural World Bank Reprint Series: Number 207. Extension Projects Reprinted from Economic Development and Michael M. Cernea and Cultural Change. vol. .30. no. I (October 1981):59- 76; and American Journai of Agricultural Benjamin J. Tepping Economics. vol. 64. no. I (February 1982J:145-47. World Bank Staff Working Paper No. Stock No. RP-0207. Free of charge. 272. December 1977. vi + 115 pages (including 9 annexes, bibliography). Sociological Dimensions of Extension Stock No. WP-02 72. $5.00. Organization: The Introduction of the T&V System In India Michael M. Cemea Thailand-Case Study of World Bank Reprint Series: Number 196. Reprinted from Extension Education and Rural Agricultural Input and Development, vol. 2. (1981):221-35, 281. Output Pricing Stock IYo. RP-0196. free of charge. Trent Bertrand World Bank Staff Working Paper Nio. 385. April 1980. ix + 134 pages iYEW (including 2 appendixes). Economic Analysis of Agricultural Projects Stock [No. WP-0385. $5.00. Second edition, completely revised and expanded REPRINTS J. Price Gittinger This entirely new edition of the Worid Bank's best-seiling Adoption of Interrelated Agricultural book sets out a careful and practical methodology for Innovations: Complementarity and the analyzing agricultural development projects and for using Impacts of Risk, Scale, and Credit these analyses to compare proposed investments. It Gershon Feder covers what constitutes a 'project" what must be con- World Bank Reprint Series: Number 206. sidered to identify possible agricultural projects, the life Reprinted from American Joumal of Agricultural cycle of a project, the strengths and pitfalls of project Economics. yol. 64, no. I (february 1982JS94-101. analysis, and the calculations required to obtain financial Stock IYo. RP-0206. free of charge. and economic project accounts. In the ten years since its publication, the first edition has Agrcultural Policies and Devejopment: been accepted widely as a standard reference and text. The A Socioeconomic Investigation Applied methodology reflects the best of contemporary practice in to Sri Lanka Martha 11. de Melo government agencies and international development institutions concerned with investing in agriculture and is World Bank Reprint Series: NumDer 191. accessible to a broad readership of agricultural planners, Repronted/ram The Journal of Policy Modeling, engineers, and analysts. vol. 1. no. 2 (May 1979):217-34. This revision adds a wealth of recent project data; Stock lYo. RP-0191. free ofcharge. expanded treatment of farmn budgets and the efficiency prices used to calculate the effects of an investment on Choice of Technique In Sahellan Rice national income; a glossary of technical terms; expanded croduction appendixes on preparing an agricultural project report and Charl(es e Humphreys and Scott R. Pearson using discounting tables; and an expanded, completely World Bank Reprint Series: rlumber 199. annotated bibliography. Reprinted from Food Research Studies. uol. 17, no. 3 (1979-80J:235-77. ED Series in Economic Deve10pment. Stock No. RP-0199. free of charge. The Johns Hopkins University Press. July 1982. 528 pages (including appendixes and glossary/index). LC 82-15262. ISBN 0-8018-2912-7, $37.50 (S22.50) hardcover; ISBN 0-8018-2913-5, $13.50 (48.75) paperback. WORLD BANK PUBLICATIONS ORDER FORM SEND TO: WORLD BANK PUBLICATIONS WORLD BANK PUBLICATIONS P.O. BOX 37525 or 66, AVENUE D'IENA WASHINGTON, D.C. 20013 75116 PARIS, FRANCE U.S.A. Name: Address: Stock or ISBN # Author, Title Qty. Price Total Sub-Total Cost: Postage & handling fee for more than two free items ($1.00 each): Total copies: _ _ Air mail surcharge ($2.00 each): TOTAL PAYMENT ENCLOSED: Make checks payable: WORLD BANK PUBLICATIONS Prepayment on orders from individuals is requested. Purchase orders are accepted from booksellers, library suppliers, libraries, and institutions. All prices include cost of postage by the least expensive means. The prices and publication dates quoted in this Catalog are subject to change without notice. No refunds will be given for items that cannot be filled. Credit will be applied towards future orders. No more than two free publications will be provided without charge. Requests for additional copies will be filled at a charge of US $1.00 per copy to cover handling and postage costs. Airmail delivery will require a prepayment of US $2.00 per copy. Mail-order payment to the World Bank need not be in U.S. dollars, but the amount remitted must be at the rate of exchange on the day the order is placed. The World Bank will also accept Unesco coupons. PUB HG3881.5 .W57 W67 no.564 Perratc.n, H. D. Basic education and agricultural extension costs, effects, and PUB HG3881.5.W57 W67 no.564 Perraton, H. D. Basic education and agricultural extension costs, effects, The World Bank Headquarters: U 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. Telephone: (202) 477-1234 Telex: WUI 64145 WORLDBANK RCA 248423 WORLDBK Cable address: INTBAFRAD WASHINGTONDC European Office: 66, avenue d'1ena 75116 Paris, France Telephone: (1) 723-54.21 Telex: 842-620628 Tokyo Office: Kokusai Building 1-1, Marunouchi 3-chome Chiyoda-ku, Tokyo 100, Japan Telephone: (03) 214-5001 Telex: 781-26838 ISSN 0253-2115/ISBN 0-8213-0176-4