92975 2013 Evaluation of the Solomon Islands Rural Development Program Ananta Neelim & Joseph Vecci Monash University 10/31/2013 Solomon Islands Ministry of Development Planning and Aid Coordination Rural Development Program Evaluation of the Solomon Islands Rural Development Program Prepared by: Ananta Neelim and Joe Vecci Ananta Neelim, School of Economics, Monash University, 3800, Victoria, Australia. Email: ananta.neelim@monash.edu. Joe Vecci, School of Economics, Monash University, 3800, Victoria, Australia. Email: joseph.vecci@monash.edu. 2 Table of Contents: Contents Executive Summary.................................................................................................................................. 6 Section 1: Introduction........................................................................................................................... 13 Background ........................................................................................................................................ 13 Section 2: Methodology.......................................................................................................................... 14 Field work. ....................................................................................................................................... 16 Data entry and cleaning ....................................................................................................................... 16 Section 3: Setting .................................................................................................................................... 18 Section 4: Access to Infrastructure ......................................................................................................... 19 4.1.1) Travel time to Access Water ..................................................................................................... 19 4.1.2) Factors that improved Access to Water..................................................................................... 23 Section 5) Access to Infrastructure ......................................................................................................... 29 5.1.1) Access to Primary Schools ........................................................................................................ 30 5.1.2) Access to Health Services ......................................................................................................... 32 5.1.3) Access to Roads........................................................................................................................ 33 5.1.4) Access to Markets ..................................................................................................................... 34 5.1.5) Access to Sanitation Facilities ................................................................................................... 35 5.1.6) Access to Electricity.................................................................................................................. 36 5.1.7) Construction of the variable to measure achievements. ............................................................. 40 5.2.1) Temotu Examined ........................................................................................................................ 41 5.3) RDP Processes ................................................................................................................................ 45 5.3.1) Sub-project Management .......................................................................................................... 45 5.3.2) SIC Management ...................................................................................................................... 46 5.3.3) Implementation challenges ....................................................................................................... 50 5.3.4) SIC Bank Account .................................................................................................................... 52 5.3.5) Ward Development Committee ................................................................................................ 54 5.3.6) Sub-Project Quality .................................................................................................................. 55 5.3.7) Sub-Project Maintenance .......................................................................................................... 58 5.4) Sub-project Participation ................................................................................................................. 62 5.4.1) Participation ............................................................................................................................. 62 5.5) Gender ............................................................................................................................................ 65 5.5.1) Engagement of Women in RDP ............................................................................................... 65 3 5.5.2) Effects of Female Participation in the RDP Process. ................................................................ 69 5.6) Economic Activity and Services:...................................................................................................... 73 5.6.1) Primary School Attendance....................................................................................................... 73 6.1) Component 2 ................................................................................................................................. 77 6.1.1) Agricultural Advice and Farming Practices................................................................................ 77 7.1) Intermediate Indicator................................................................................................................... 82 7.1.1) Agricultural Services ................................................................................................................. 82 7.2.2) Consultation ............................................................................................................................. 84 7.1.2) Changes in farming practices .................................................................................................... 86 7.2) MAL services .................................................................................................................................. 88 7.2.1) Type of Advice Sought ............................................................................................................. 88 7.3.1) Food Security ............................................................................................................................... 89 7.3.2) Food Security and Markets. ...................................................................................................... 92 Section 8: Recommendations ............................................................................................................... 95 8.1.1) Component 1: Infrastructure Services Recommandations. ........................................................ 95 8.2) Component 1 RDP Processes- Recommendations: ..................................................................... 96 8.3) Component 2 Recommendations: ............................................................................................... 97 8.4) Future Methodology: ................................................................................................................... 98 Appendix 1:.......................................................................................................................................... 101 Session Logs .................................................................................................................................... 101 Choiseul ....................................................................................................................................... 101 Malaita ......................................................................................................................................... 101 Temotu: ....................................................................................................................................... 101 Appendix 2 .......................................................................................................................................... 104 Annex 1. Distribution of Village Size ............................................................................................... 104 Annex 4 Percentage of villages with satisfactory access to roads (using 70% cut-off) ........................ 105 Appendix 3 .......................................................................................................................................... 107 Section A: General........................................................................................................................... 107 Section B: Economic Activities ........................................................................................................ 109 Section C: Participation .................................................................................................................... 110 Section D: Access to infrastructure................................................................................................... 112 Section E: Organisation of RDP Sub-project.................................................................................... 113 Annex 11: Household Survey Results. .................................................................................................. 115 Section A: General ........................................................................................................................... 115 4 Section B: Access to Water, Electricity and Sanitation. ..................................................................... 117 Section C: Children .......................................................................................................................... 119 Section E: RDP Sub-project ............................................................................................................. 120 5 Executive Summary The purpose of this report is to analyse the impacts of the Rural Development Programme (RDP) of the Solomon Islands. RDP is an initiative to facilitate development in rural Solomon Islands. It focuses on participatory development, demand-responsive provision of Government services and creation of a supportive economic environment for small-scale rural development to foster employment and income growth. As a part of this evaluation process, a follow up survey was conducted in eighty villages, equally distributed across four provinces (Choiseul, Malaita, Temotu and Western) between June and July 2013. In each village we utilised two survey instruments, a) a household survey which was administered on ten randomly chosen households and b) a second questionnaire administered on community leaders (which also included members of the RDP’s subproject implementation committee). The data collected in the baseline survey, conducted between 2010 and 2011, was the main basis for comparison and subsequent impact evaluation. A total of 80 villages were surveyed in the follow up an overlap of 63 with the baseline survey. A number of caveats must be taken into account when interpreting these results: 1) The surveying in both the baseline and follow up do not follow a randomised controlled trial methodology making it difficult for causal inferences. 2) A number of villages are not re surveyed from the baseline making it difficult to draw inferences from before and after comparisons. Component 1 Compared to the baseline we find significant improvements in access to roads, markets, water, sanitation and electricity, ranging from 11.3 to 1250 %. However, at the same time there was a decline in access to primary schools and health centres. We find in general that RDP villages fared better than non-RDP villages in terms of access to most of these services. For example, the rate of improvements in access to roads, markets and water was higher in villages which had related RDP projects. Similarly, the rate of deterioration in access to primary schools and health centres were lower in villages that had related RDP projects. This suggests that the RDP has had an impact in improving or providing stability in the access of basic services in the villages that they serve. Improving access to water was one of the main targets of component 1 of the RDP. We found that on average there was a decrease in the time taken to fetch water in both the wet and dry 6 seasons. This was driven mostly by improvements in villages that decided to choose water related projects. In the baseline survey these villages generally reported longer times in fetching water in both wet and dry seasons compared to villages that did not choose water related projects. This is evidence that the RDP processes with regards to community driven selection performed well. In the follow up, we found that the time required to fetch water in the wet season was lower in RDP villages that selected a water project compared to villages that choose not to do water projects. In the dry season, fetching water is still quicker in the villages that did not choose water projects. However, the difference in time taken across these two types of villages in the dry season has been drastically reduced. This is evidence that RDP projects have generally performed well improving access to, as well as the time taken, to fetch water in both dry and wet seasons. Moving forward, the RDP needs to address the issue of sustainability of improved services. Evidence from access to primary schools, health services as well as water suggest that if coordinated efforts are not put into maintenance of resources there could be deterioration in the future. Our analysis suggests that communities lack knowledge as well as experience on how to maintain existing infrastructure projects. Alarmingly, a significant portion of the existing project recipients have no plans for future maintenance of their projects. Given that these are community projects (as opposed to private ones) the incentives at the individual level, to provide efforts unilaterally, are also very little. Thus, it is imperative that RDP designs strategies to raise awareness about the importance of timely maintenance of existing infrastructures. In designing these strategies, consulting existing research on inducing cooperation in common property resources management would be very helpful. RDP Processes: Selection to implementation One of the most important features that make RDP unique is its objective to foster community participation in every step from project selection to implementation. One of the rationales for this approach is that community participation induces ownership and as a result reduces the inherent perverse incentives associated with consumption of benefits from common property resources. The first strategy that RDP utilised to achieve community participation was by allowing villagers to select their own projects. In our survey we find that the majority of the households actively took part in the project selection process. We also find that people who did not attend selection meetings could not attend because of prior commitments. 7 To enhance community participation during the implementation stage, RDP required each village to elect a sub-project implementation committee (SIC). The main purposes of this committee were (a) to be a liaison between the community and RDP and (b) to ensure necessary steps were taken for timely completion of sub-projects. We find that villagers were relatively satisfied with the SIC and believed it was an effective method of coordinating the sub-project. However, households also believed that if the representation of women and youth increased then these committees would perform even better. Households also believed that there is scope for improvements in SICs financial and project management as well as SICs ability to coordinate activities within the community. It was found that the general quality of RDP projects were good. Households also believed that the quality of RDP projects were better than similar projects undertaken by other agencies. An important reason for this high level of satisfaction amongst villagers is related to RDPs processes to engage the community. However, communities also identified challenges and impediments that they faced. The leaders in the community reported that the level of meaningful community participation (in terms of providing raw materials) was either low or slow. They also identified the RDP process of funding limits as well as delays in the disbursement of funds as cumbersome and detrimental to the timely completion of projects. In remote villages distance from suppliers provided an additional challenge. We find that women generally participated less in the selection of RDP projects and when they did participate they were more passive compared to men. However, the various strategies undertaken by RDP did have impacts over time. Most households believed the RDP process encouraged greater female engagement. In fact households report that female members that join the SIC increasingly participated within and outside the community. Female memberships and engagement was also found to have positive correlations with project outcomes. Households believe that female participation enhancing cooperation between groups. Most communities seem to understand this and as a result suggest increased women representation as a main suggestion in terms of improving SICs. Progress across Provinces The improvements in access to basic services were not similar across provinces. There was a systematic pattern: we found that Temotu did not perform as well as other provinces. In fact we found that access to most infrastructures drastically deteriorated in Temotu. This in part can be 8 explained by a small number of villages which suffered floods. In terms of component two indicators, the disparity across provinces were less pronounced and not systematic. Component 2 The main goal of component 2 was to increase the take up of improved farming practices, which RDP believes will lead to better livelihood and development. The main instrument that was developed to achieve this goal was village level dissemination of agricultural advice/information and related services through the Ministry of Agriculture and Livestock (MAL). The rationale was that improved access to better information would lead to improved agricultural practices. Results from the follow up survey show that compared to the baseline, households were more likely to have received agricultural advice from MAL. Generally, a majority of households that received advice were satisfied with it. This meant that in a significant number of cases households changed farming practices upon receiving advice from MAL. This led to increased agricultural output for these households. However, the intensity of this process was low: at follow up three quarters of the households did not receive advice from the MAL and when they did about half of them changed practices. This alludes to the important need to improve both the scope of dissemination of agricultural advice as well as the take up rate upon receiving advice. One of the reasons we suggest the aforementioned improvements is because we find a statistically significant relationship between receiving advice and food security. In the follow up, we find that 1 % increase in advice received is correlated with a 6 % decrease in the probability of being food insecure. The non- existence of this relationship during the baseline further justifies the rationale for having this strategy as one of the core objectives of RDP. It also alludes that increasing the rate at which advice is disseminated will lead to higher level of food security. Targets: Achievements and Failures The RDP had set itself targets against which their performances were to be evaluated. With regards to improvements in access to basic services the target set was to increase the proportion of villages with satisfactory access to basic services by 10 % compared to the baseline. For water the target was more specific, which was to decrease the travel time to access improved water by 10 %. In terms of meeting these targets RDP has shown good progress. On both the indicators RDP has met their objectives. (Component 1 Indicators). 9 Table 1: Component 1 Indicator One & Two Results PDO Level Results D=Dropped Unit of Cumulative Target Indicators C=Continue Measure Values Percentage Evaluation N=Revised Baseline Change R=Revised Yr Yr Yr Yr Score 1 3 5 7 Indicator One: N Percent Zero 13.54 NA 10 Percentage decrease in travel time to access improved water supply sources Indicator Two1: R Percent 34.21 65.82 92.4 10 Percentage of villages with satisfactory access to infrastructure Targets were also set with regards to intermediate achievements of RDP processes. From the table below we can also see that the performance have been satisfactory. Intermediate indicator six has already been achieved, whereas indicator one is very close to being achieved. However, RDP is off target with indicator three. Table 2: Component 1 Intermediate Results Indicators PDO Level Results D=Dropped Unit of Cumulative Target Values Indicators C=Continue Measure Evaluation N=Revised Baseline R=Revised Yr Yr Yr 5 Yr 7 Score 1 3 Intermediate Result R Percent Zero 47.9 80 indicator Three: Percentage % of households in target planning units participating in RDP Community Development Sub-project decision making Intermediate Results N Percent Zero 93.92 80 Indicator six: Percentage of households in communities receiving project funding that feels project investments reflect their needs Intermediate Results N Percent Zero 93.4 80 Indicator six b):…of which are women 1 The calculation of this indicator is explained in the Section 5. 10 Finally, RDP has failed to achieve its targets with regards to component 2. Although improvements were made compared to the baseline, the rate of progress has been slow. Table 3: Component 2 Results Indicator Three & Intermediate Results Indicator One PDO Level Results D=Dropped Unit of Cumulative Target Values Indicators C=Continue Measure Evaluation N=Revised Baseline Core R=Revised Yr 1 Yr 3 Yr 5 Yr 7 Indicator Three: N Percent 2 11.20% 35 Percentage of Households who have changed agricultural practices including varieties, as a result of agricultural advice. Intermediate Result R Percent Zero 49% 20 50 indicator One: Percentage of villages in participating provinces to whom agricultural services have been provided. Recommendations Component 1: Infrastructure  A clear definition of what constitutes clean water needs to be identified.  We recommend RDP introduces targets based on overall coverage rate in addition to percentage improvements between baseline and follow up surveys.  We recommend that the RDP creates a results framework for the province level.  The distinction between private type and public type usage projects needs to be made and based on these distinctions best practice strategies for improving outcomes in both private projects and public projects will have to be formulated.  It is recommended that the RDP trains communities on best practices in management and allocation of water resources to improve maintenance. Component 1: RDP Processes  The RDP should devise strategies to improve participation, such as coordinating with churches and other community groups to provide advance notices about meetings. 11  Raise awareness about the importance of timely maintenance of infrastructure to ensure project sustainability. Particular focus should be on people who are given responsibility for operations and maintenance.  Improve monitoring of CH’s especially those further away from (in terms of distance) from PSUs. A possible suggestion would be for communities to maintain log books to record CH attendance in meetings.  We recommend that the RDP make it mandatory for women to be included in future project implementation and maintenance committees. Component 2:  MAL officers must increase participation at MAL events including training and consultation.  We recommend undertaking research into costs involved in changing farming practices to identify why take up of advice is limited and how it can be improved.  MAL should encourage communication and learning between those households that received advice and changed practices and those that did not.  To improve outreach MAL should provide passive ways through which farmers can reach out to them. For example, a help exchange could be set up where farmers from any village are able to telephone MAL officers seeking advice. 12 Section 1: Introduction Background The Rural Development Project (RDP) is based in the Solomon Islands Ministry of Development Planning and Aid Coordination. RDP is an initiative to facilitate development in rural Solomon Islands. It focuses on participatory development, demand-responsive provision of Government services and creation of a supportive economic environment for small-scale rural development to foster employment and income growth. The Project comprises three components: Component One: Delivery of Local infrastructure and services Component Two: Improvement of agricultural services Component Three: Rural enterprise development This report is concerned with an evaluation of components one and two based on data from the baseline survey carried out between June and July 2009 (and subsequently April/ May 2010) and a follow up survey carried out in between June-July 2013. Format of this Report This report comprises of seven main sections. Section 2 outlines the research methodology, providing an account of various processes that were utilised to ensure an unbiased evaluation. Section 3 reports the basic sample characteristics at both the village as well as household level. In section 4 and 5 we present the results from component 1 and in section 6 and 7 we provide the results from component 2. In the final section we provide specific recommendations, in light of the findings of this report, to improve the various components of RDP. Section 2: Methodology Sample Design The baseline study sample consisted of four provinces - Choiseul, Malaita, Temotu and Western provinces. It comprised a mix of villages that had an RDP sub-project approved but not yet commenced, and villages that had applied for a sub-project but were still awaiting the outcome of the project selection process. Up to 20 households were selected for interview in each village. Age/gender discussion groups were also formed from eligible villagers who were available and willing to participate at that time. The follow up involved surveying 80 villages equally distributed across the same four provinces, RDP villagers were chosen using a three state sampling strategy: 1. All 76 villages were taken from the baseline sample. 13 villages were dropped, 5 because they did not have a sub project and 2 due to distance. 4 villages were dropped 4 in Malaita and 2 in Western due to unforeseen circumstances such as the weather. 2. Eleven Cycle 3 communities (3 in Choiseul, 5 in Malaita, and 3 in Western, were randomly chosen to ensure 20 communities in each). 3. Communities that were unreachable were replaced with randomly chosen cycle 3 villages in the same ward. In nearly all of the 80 villages RDP subprojects were completed or were near completion. Up to 10 households from each village was selected for interview. In addition, in each village a community leader’s questionnaire was administered. In this discussion typically the village chief, a female representative and members of RDP's sub project implementation committee's were present. Scope of the study The scope of the study was determined by the RDP monitoring and evaluation framework. All questions in all survey instruments (both baseline and follow up) were closely linked to the project’s results monitoring table to provide indicators to capture the primary areas of impact 14 and outcome defined in project documentation. The baseline survey provided a benchmark against which impacts were to be determined. The main topics that this report looks at (i) changes in poverty including changes in household access to and quality of infrastructure, income generation, (ii) participation in development process, particularly RDP processes for both component 1 and 2, (iii) review of implementation and challenges of the RDP processes for both component 1 and 2 and (iv) changes in farming practices and subsequent impact on livelihood which comes under component 2. In addition the report provides recommendations on how to improve RDPs functions based on these findings. Survey Instrument The core survey instruments from the baseline survey that were used in preparing this report were household surveys, and surveys from discussions with men, women, young men and young women focus groups. In the follow up survey we used data from the household survey and community leader’s survey. We also used general observations (from sessions logs) based on field visits to supplement our analysis. All the core survey instruments were designed in consultation with RDP officials and were pilot tested in four villages in Guadalcanal and Malaita provinces. The complete set of questionnaires and other survey instruments are attached with the report. Recruitment and Training For both the baseline and follow up surveys, the interviewers were experienced fieldworkers. Initially interviewers were given a week of intensive training in the use of the data collection instruments prior to the pilots. During the pilots, their performances were monitored and further sessions were carried out to fine tune their skills to ensure that they were properly trained. The training process included teaching interviewers strategies of approaching villagers, introducing survey teams and explaining the purpose of the surveys, consent procedure, detailed explanation of each question to be asked and practice or mock interviews3. Field work for the 3 The baseline and follow up survey enumerators were given training to ensure randomisation in the collection of household data. 15 baseline survey was conducted by the Australian National University while the follow up survey was conducted by the University of New South Wales. The follow up survey was implemented by four province teams. Teams included 2 post-graduate students from either the University of San Francisco or the University of New South Wales, 1 RDP Employee (Community Helper) and 2 interviewers. It was ensured that RDP employee were not present during any of the household surveys or the community leaders discussion to guarantee the integrity of the instrument. Field work. The baseline field work was completed in 2 phases. The first phase was June/July 2009. Because of the extreme remoteness of some villages and seasonal rough seas, teams were unable to reach all target villages in Choiseul, Western and Temotu in the required timeframe with the funds available. It was therefore necessary to carry out a second round of fieldwork in April /May 2010. The follow up survey field work was implemented in June/July 2013. Data entry and cleaning For the baseline survey a team of experienced keyboard users with competence in Microsoft Excel were recruited to undertake data entry in Honiara. All numeric data were entered into prepared templates and any open ended answers were transcribed verbatim. On-the-spot checking and cleaning was implemented during data entry. When data entry was complete frequency tabulations were run with computer software (PASW Statistics 18) and out-of-range values were checked and cleaned. In the follow up survey all datum were entered by the post-graduate students in the field. Entered data was checked for consistency and subsequent mistakes if found were corrected. Upon completion of data complete frequency tabulations were computed using Stata 11 and irregular data entries were checked and cleaned. 16 Data Analysis and limitation of data analysis As mentioned earlier the basis of analysis in this report is the baseline and the follow up surveys. Given the different characteristics of the various datasets multiple techniques were utilised to analyse the data collected. The statistical software ‘Stata’ was used to conduct most of the analysis. Multivariate analysis was undertaken to answer questions on food security and schooling. Interpretations of these models are discussed in their respective sections. The limitations of our analysis are presented below. 1. The main limitation of this evaluation is the fact that it does not follow a randomised controlled trial (RCT) methodology. As a result it is not possible to make causal claims about the impacts of RDP. However, given the aims and objectives and the working processes of RDP it would have been difficult to conduct an RCT to ascertain impact evaluation. However, there are multiple second best strategies that are available all of which rely on having a control and a treatment group. Given the nature of sampling where all of the villages chosen for follow up survey were villages which had RDP involvement thus making it difficult to identify causal effects directly. 2. Between the follow up and baseline 65 villages were surveyed in both. This fact along with the non-random selection of villages for both surveys, create additional complications with analysis. In this report we compared a set of instruments from 80 non randomly surveyed villages (follow- up) with a set of instruments from 76 non randomly surveyed villages (baseline). Thus the main findings of the study should not be extrapolated beyond that of sampled villages without exercising further caution. The non-random nature of sampling also prohibits us from undertaking any meaningful tests of formal hypothesis. 3. In the baseline survey, (i) the selection of households were not entirely random, (ii) participants for the four age/gender discussion groups were selected on the basis of availability and willingness to participate and not randomly. 4. In the follow up survey, (i) the number of people in the community leaders questionnaire were not always the same (as the number of SIC members varied based on availability). Also, given that one of the aims of this instrument was to critically analyse the RDP processes including the SIC, which most of the respondents were a part of, this information gathered may be biased. 17 Nevertheless, these respondents were the most involved in the RDP process thus they had the best knowledge on how to improve it. Finally, response rates for certain questions both in the community and household questionnaire were very low which made it difficult to interpret some results at the province level. In these cases as a result we present country level comparisons. Section 3: Setting Most villages in Solomon Islands are small. On average each village in the follow up survey had 78 households. The largest village had 500 households and smallest had 8 households. The distribution of village sizes is presented in the appendix (Annex 1. Distribution of Village Size). Most of the villages are governed by traditional village chiefs. In some cases elected leaders and church leaders also help with governance. Most villages have churches (in some villages multiple churches) which also act as the community hall for meetings. On average every village has at least one church service a day. In most villages the three most important sources of livelihood come from selling produce (fish, crops, livestock), cocoa/copra and other cash crops or from logging royalties. In around 80 % of the villages more than three quarters of all villagers depend on a subsistence (or semi-subsistence) economy. In 34 and 17 % of the sampled villages, there was at least one person within the community involved with logging or mining. In 53 % of the villages, the national MP had relatives. The average household size in our sample was 5.54 people4 equally split across gender. Around 70 % of the households reported having at least one child attending school or kindergarten. Additionally, 35 % of households reported having at least one child below the age of 5. In 58 (42) % of cases household survey respondents were male (female). The average age of household responders was 43.7 years. The majority of households reported having access to rainwater catchments for drinking water in the wet season. 82 % of households used solar lamps for lighting their households. However, 72 % of households reported that they defecate in the Sea or River and 3% in the bush. In terms of access most surveyed areas were located remotely: only 32.5 % of the villages reported that the time required to reach their provincial capital from their village is less than two hours. On average the travel time between villages and their respective provincial capital was 8.6 hrs. On the other hand, the average travel time to Honiara was two and a half days. People from 68 % of the villages use ship or OBM canoe to travel between their 4In the 2009 census this value was 5.51. Most of the statistics we provide are consistent with the findings of the 2009 census. 18 village and their provincial capital. Most of the surveyed villages were located on the coast with very little access to proper roads. Boat transport is the main mode of transport in most provinces. Section 4: Access to Infrastructure This section will examine access to infrastructure. It will be divided into three main parts. The first part will look at water access including the first results indicator. The second section will examine access to other infrastructure including; schools, health centres, roads, markets, water, sanitation and electricity. The last section will examine RDP processes including; implementation, construction, participation and gender. 4.1.1) Travel time to Access Water Table 4: Indicator One: Percentage decrease in travel time to access improved water supply sources PDO Level Results D=Dropped Unit of Cumulative Target Values Indicators C=Continue Measure Evaluation N=Revised Baseline R=Revised Yr 1 Yr 3 Yr 5 Yr 7 Score Indicator One: N Percent Zero 13.5% 10 Percentage decrease in travel time to access improved water supply sources Indicator one examines the percentage decrease in travel time to access improved water supply. Improved water supply sources are defined in question D9 of the household survey which asks: has your household’s access to clean drinking water improved during the past few years? Results are shown in Table 4. Overall, households experienced a 13% reduction in access to improved water sources. Results differ between provinces (Table 5). Choiseul and Temotu experienced a 23.5% and 76% increase in travel time to improved water sources respectively, while Malaita and Western provinces had a 40% and 15% decrease in travel time, respectively, to access improved 19 water sources. In conclusion, this report finds that RDP has exceeded its target in reducing travel time to access improved water sources. These results are analysed further below. Table 5: Travel time to access improved water sources Improved Not improved Difference Province Choiseul 7.7 6.3 23.4% Malaita 9.8 16.5 -40.6% Temotu 29.6 16.8 76.2% Western 10.0 11.7 -14.9% All 12.5 14.4 -13.5% Source: Improved access is defined in D9 of the Household survey. Yearly travel time is reported here by averaging travel time between wet and dry seasons (B6 and B9 of Household). Results are denoted in minutes. Figure 1 : Time it takes to fetch water in wet and dry season Wet Season 20 15 10 5 0 Choiseul Malaita Temotu Western All All (ex. Temotu) Dry Season 40 30 20 10 0 Choiseul Malaita Temotu Western All All (ex. Temotu) Baseline Wave 1 Source: Wave 1: Household- B5 and B6. Baseline: Household- B2 and B3. Yearly travel time is reported here by averaging travel time between wet and dry seasons (B6 and B9 of Household). 20 From Error! Reference source not found.5, we observe a significant decrease in travel time to fetch drinking water in the wet season. Malaita (73 %) and Choiseul (63 %), show the biggest improvement in travel time over the past 5 years. On the other hand, Temotu had a 9 % increase in travel time to fetch water. Overall in the wet season, villagers experienced a 49 % decrease in travel time to collect drinking water. The dry season also saw significant improvements in some provinces. In Choiseul and the Western province, travel time decreased by 46 % and 32 % respectively. However in Temotu travel time increased by 127 %. Overall, the travel time to fetch water increased on average by 0.5 minutes per day. This increase was mainly driven by Temotu. Another way of ascertaining access is to analyse the proportion of households or villages with satisfactory access to water6. In the baseline, only 34.2 % of the villages reported satisfactory access to water. This figure improved to 72 % during the follow-up survey. This translates to a 110.5 % increase between the two surveys. In terms of performance, the Choisuel province exhibited the highest levels of improvements followed by the Western province. In terms of the overall status of access to water, the Choisuel also had the highest rate. Although Temotu registered a 53.3 % improvement, the overall status of access to water services is still low at 40 % (Figure 2). Figure 2: Percentage of villages with satisfactory access to water 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western Full Full w/o Temotu Baseline Wave 1 Source: Wave 1: Household- D2. Baseline: Houshold-B21. There has been a marked improvement in access to water services in all of the four study provinces. On average, 71 % of the households reported that they have satisfactory access to water. In the baseline, the corresponding number was 33.3 %. There is progress in terms of 5 Data here is used from two points of time, baseline and follow-up. 6 In this metric we use the definition outlined in the RDP evaluation framework. Details presented in section 5. 21 quality as well. Over 54 % of the households at follow up indicated good access to water. This is a 280 % increase compared to the baseline. In terms of performance, Choiseul province rated the best with almost 95 % of the households reporting having satisfactory access to water. In Temotu, on the other hand, the majority of the people still report having poor access to water (Table 6). Table 6: Access to Water (Household Data) Good Satisfactory Bad Province B W1 D B W1 D B W1 D Choiseul 14.8% 79.9%440.0% 18.4% 14.8% -19.3% 67.1% 5.3% -92.1% Malaita 23.9% 56.3%135.4% 18.2% 10.1% -44.7% 57.9% 33.7% -41.9% Temotu 9.3% 17.7% 91.0% 15.0% 26.3% 75.1% 74.8% 56.1% -25.0% Western 18.4% 69.9%280.5% 16.8% 13.7% -18.4% 64.5% 16.3% -74.7% Full 15.8% 54.8%246.1% 17.0% 16.4% -3.5% 66.7% 28.8% -56.8% Full w/o Temotu 18.7% 68.4%266.1% 17.8% 12.8% -28.2% 63.4% 18.9% -70.2% Source: Wave 1: Household- D2 & Baseline: Houshold-B21 Greater access to water is important for development outcomes7, however unless this access is accompanied by basic sanitation and improved water quality, health improvements will be marginal. Table 7 reports the main source of drinking water for a household between the baseline and wave 1. The main source of drinking water in the wet season 8 varied markedly across the sampled villages and provinces. However, a number of trends appear: the majority of households, 56%, use rainwater as their main water source. This is up from 24 %, an increase of 128 %, from the baseline. Households report that rainwater is stored in tanks provided by projects such as CSP and RDP. This water source is protected and is less likely to be contaminated given the low level of industrial activity in the study provinces. Figure 3: Improvements in household access to clean drinking water improved during the past few years 7 Todaro,M and Smith S (1997). Development Economics. New York: Addison-Wesley. 8 Dry season not available 22 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Source: Household-D9 The movement towards rainwater tanks is associated with a decrease in the usage of springs, boreholes and rivers throughout the provinces. It was found that households decreased their usage of springs and rivers by 52%, and boreholes and wells by 38%. These water sources are more likely to be polluted, particularly as a number of streams are prone to flooding. In summary, these results show a significant improvement in the access to and quality of water sources throughout surveyed villages. These results are confirmed in Figure 3. Overall, 60% of households believe their access to clean drinking water has improved over the past few years. Such improvements are recorded to be highest in Choiseul where 91% of households experienced greater access to clean water. The lowest findings are in Temotu and Western provinces, where 42% of households believed their access improved. 4.1.2) Factors that improved Access to Water. Since the baseline in 2009/10, households have notably had better access to quality water sources. Improvements in quality and access are often funded by a number of different organisations including the local community, the RDP and the Solomon Islands’ Government. Error! Reference source not found. lists the organisations responsible for funding the improvements to the water system in each province. On average, for all villages, the RDP was the largest contributor to the improvement of water supply, funding 25% of all cases. The community funded an average of 20% of the water improvement projects, while the Solomon Islands’ government contributed to 8% of the projects. However, large differences exist between provinces. Foreign governments and NGOs were the largest contributor to water projects in Temotu (20% and 25% respectively) and in Western province, the majority (40%) of projects were funded by unspecified other parties. 23 Figure 4: List of Organisations funding improvements to water. 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All RDP Community SIG Foreign Gov. NGO's Source: Household-D4 As mentioned in Figure 4 the RDP plays an important role in the access to water through water projects. To clearly identify if these projects are associated with improved access, we report the mean time to fetch water for those villages with an RDP water project. The same is done for RDP villages not choosing water projects, in other words RDP villages which did not select a water project. A greater difference in the time taken to fetch water between villages with an RDP water project and RDP non-water project villages9 would demonstrate that RDP water projects are crucial in improvements in access to water. Figure 5 Minutes to fetch household water in the wet season in a RDP water project village 35 30 25 20 15 10 5 0 Choiseul Malaita Temotu Western All Baseline Wave 1 Source: This variable is created using B5 in Wave 1 and B2 in the baseline together with information as to whether a community built a RDP water project. An RDP water project is defined as a sub-project that provides water, including water tanks and dams. Figure 6: Minutes to fetch household water in the dry season in a RDP water project village. 9 An RDP water project here refers to a village which selected an RDP water project. Non- RDP water projects refers to villages that select an RDP project other than water. 24 70 60 50 40 30 20 10 0 Choiseul Malaita Temotu Western All Baseline Wave 1 Source: This variable is created using B6 in Wave 1 and B3 in the baseline together with information as to whether a community built an RDP water project. An RDP water project is defined as a sub-project that provides water, including water tanks and dams. The results in Error! Reference source not found. through to Figure 8 show that in the baseline, villagers who selected an RDP water project took on average 14.5 minutes to fetch water in the wet season and 42 minutes in the dry season. In comparison villages which selected projects (other than a RDP water project) took 7 minutes to fetch water in the wet season and 19 minutes in the dry. This result infers that at the baseline RDP water project villages had considerably lower access to fresh water compared to other project villagers. This means RDP project selection is correctly providing water projects to those villages which need it most. Furthermore, as mentioned, at the time of the baseline survey RDP water project villagers had much lower access to water sources, however, by the completion of the follow up survey RDP water project villagers travelled less to access water compared to other project villages. Figure 5 shows that, on average, villages with a water project had large decreases in time to fetch water, on average 79% in the wet season and 40% in the dry (Error! Reference source not found.). All provinces saw a decrease in travel time except Temotu where travel time increased by 63% in the wet season and 41% in the dry. These results are in direct contrast with all non-RDP water project villages in which travel time decreased by 39% in the wet season (Figure 7) but increased by 1% in the dry season (Error! Reference source not found.). Figure 7: Minutes to fetch household water in the wet season in a RDP non-water project village. 25 14 12 10 8 6 4 2 0 Choiseul Malaita Temotu Western All Baseline Wave 1 Source: This variable is created using B5 in Wave 1 and B2 in the baseline together with information as to whether a community built a RDP water project. An RDP water project is defined as a sub-project that provides water including water tanks and dams. Figure 8: Minutes to fetch household water in the dry season in a RDP non-water project village. 45 40 35 30 25 20 15 10 5 0 Choiseul Malaita Temotu Western All Baseline Wave 1 Source: This variable is created using B6 in Wave 1 and B3 in the baseline together with if information as to whether a community built an RDP water project. An RDP water project is defined as a sub-project that provides water, including water tanks and dams. The above results thus provide evidence to support four important conclusions: 1) Villages that are most in need of improved access to water sources are, on average, selecting these projects. This provides confirmation that the method of selecting RDP 26 projects is correctly matching projects with needs and thus improving returns from investments. 2) Villages with RDP water projects saw significant improvements in access to water, much higher than the general trend of improvements seen in non-water project villages. Results are especially salient in the dry season. Non-water projects saw, a 1% increase in the time taken to fetch water. This was largely driven by Temotu and Choiseul which saw increases in travel time by 55% and 450%, respectively. While in villagers with RDP water projects, travel time declined by 72% (Choiseul) while Temotu saw a 10 times smaller increase in travel time compared to non-RDP water projects. 3) Travel time to collect water in villages with RDP water projects are now lower than non RDP project villages, which is a significant achievement and will not only have health benefits but will also maximise labour for productive activities. 4) Access to water has noticeably declined in the Temotu province. This is a worrying trend and is difficult to explain, based on survey responses. We conjecture that a number of factors may be driving these poor findings: 1) natural disasters that cause damage to infrastructure such as tsunamis and earthquakes often occur in the province (4 villages report either an earthquake or tsunami effect their village in the last year); 2) the Reef Islands may be driving this result. The Reef Islands are a significant distance from both Santa Cruz and Honiara, and have relatively poor access to most infrastructures including water. Furthermore, the small and highly dispersed population also make up 35% of the Temotu sample. Temotu is examined further in section 5. 27 Table 7: The main source of drinking water for households in the wet season. Rainwater/ Catchment Spring Borehole/Well River/Dam/Stream Water Tanker Differences Differences Difference Difference Difference Baseline Baseline Baseline Baseline Baseline Wave 1 Wave 1 Wave 1 Wave 1 Wave 1 Province Choiseul 41.5% 45.2% 8.9% 0.1% 0.0% -100.0% 0.0% 0.5% 0.0% 25.9% 48.9% 88.7% 0.9% 0.5% -39.7% Malaita 5.7% 50.0% 771% 17.6% 2.2% -87.7% 10.5% 0.0% 100.0% 66.0% 46.2% -29.9% 1.1% 1.6% 51.9% Temotu 33.5% 67.2% 100% 7.4% 11.6% 56.7% 8.4% 14.6% 72.9% 60.1% 9.09% -84.9% 0.0% 0.5% 0.0% Western 47.7% 66.5% 39.4% 2.7% 3.5% 30.9% 1.3% 0.0% -100.0% 6.2% 5.20% -16.1% 2.5% 0.0% -100.0% All 33.1% 56.4% 70.5% 9.3% 4.4% -52.2% 6.4% 4.0% -37.1% 58.0% 27.5% -52.6% 6.6% 1.8% -71.7% Source: Wave 1: Household-B1. Baseline: Houshold-B1. Note: The baseline asks households “what is the main source of drinking water for this household”. They do not specify season. Since the baseline was conducted in the wet season, we compare the baseline to the wet season in wave 1. 28 Section 5) Access to Infrastructure One of the key goals of the RDP, highlighted in component objectives, is to improve access to infrastructure in rural Solomon Islands. In this subsection, we report data on the level of satisfactory access to primary schools, health centres, roads, water supply, sanitation and markets for both the baseline and follow-up surveys. Table 8: Proportion of Villages with Satisfactory Access to Infrastructure PDO Level Results Indicators Baseline Evaluation Percentage Change Percentage of villages with satisfactory access to 88.0% 82.3% -6.5% primary schools Percentage of villages with satisfactory access to 62.0% 58.0% -6.5% health services Percentage of villages with satisfactory access to 32.0% 73.0% 128.1% roads Percentage of villages with satisfactory access to 53.0% 59.0% 11.3% markets Percentage of villages with satisfactory access to 1.0% 9.0% 800.0% sanitation Percentage of villages with satisfactory access to 4.0% 54.0% 1250.0% electricity The key indicator outlined in the results framework for evaluating the impact of component 1 of an RDP is the proportion of villages with satisfactory access to infrastructure. It is important to note that data for both baseline and follow-up surveys were collected at the household level and thus an indicator variable for the villages had to be computed using household level data. The way we created the indicator variable for each infrastructure is described below. Each household in both the baseline and wave 1 survey was asked whether their access to particular infrastructures was good, satisfactory or bad. For each village, we tallied the total number of responses which indicated that their access was at least satisfactory. If this tallied number was more than 50 % of the total number of respondents in that village, then the village was classified as having satisfactory access to that infrastructure. However, one can argue that our choice of 50% as the cut-off point is arbitrary. Thus, we conducted the same exercise with a more stringent cut-off point criterion of 70 %. Under this criterion, for a village to be classified as having satisfactory access to a particular infrastructure, at least 70 % of the households in that 29 village would have had to indicate so. In our analysis, we also present results using this stricter criterion as a robustness check. 5.1.1) Access to Primary Schools In general, the level of access to primary schools in the surveyed provinces is good. On average, around 83 % of the sampled households believe that their access to primary schools is satisfactory. From Figure 9, we can observe that there has been an overall decrease in the number of villages with satisfactory access to primary education from 88 % to 82.3 % between the baseline and follow-up survey. However, this is mainly driven by data from the Temotu province. For the Temotu province, this drop was 28.8 %. If we exclude data from Temotu, we actually observe a 12.8 % increase in the percentage of villages with satisfactory access to primary schools. In terms of performance, the Western province has shown the most progress with a 7.3 % increase in the indicator variable. Results using the 70 % cut-off (Appendix 2) validate our main observations of incremental improvements. Using this second criterion, we see improvements in Malaita and Choiseul and deterioration in the Temotu and Western provinces. Figure 9: Percentage of villages with satisfactory access to primary education 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western Full Full w/o Temotu Baseline Wave 1 Source: Wave 1: Household- D2. Baseline: Houshold-B21. In terms of the quality of access, there have been improvements in all of the four study provinces with an average of 62.2 % of people expressing that their household has good access to primary schools. In terms of improvement, Choiseul province performed the best with 93.8 % reporting good access. This was followed by Malaita and Western province (Table 9). The overall level of (at least) satisfactory access to education is relatively high (79 % on average) in the study areas. Also, all provinces reported 70 % (at least) satisfactory access to primary education. However at the same time, dissatisfaction with access to education also increased in 30 all of the provinces. This implies that the improvements in infrastructure have only improved the access of people who already had prior access to primary schools. Even though this is significant, it would be worthwhile to look into why 20 % of the population still have poor access to primary education. Table 9: Access to Primary Education (household data) Good Satisfactory Poor Province B W1 D B W1 D B W1 D Choiseul 32.1% 62.1% 93.8% 50.7% 12.6% -75.1% 16.7% 25.3% 51.1% Malaita 45.2% 70.0% 54.8% 21.9% 14.5% -33.8% 14.2% 15.5% 8.8% Temotu 37.8% 47.0% 24.4% 50.3% 24.8% -50.7% 11.3% 28.3% 151.4% Western 47.4% 70.6% 49.1% 42.4% 15.8% -62.7% 9.7% 13.6% 40.4% Full 42.3% 62.2% 47.0% 43.5% 17.0% -61.0% 13.7% 20.8% 52.1% Full w/o Temotu 44.2% 67.6% 53.0% 40.8% 14.3% -64.9% 14.6% 18.2% 24.0% Source: Wave 1: Household- D2. Baseline: Houshold-B21. One of the important features of RDP is that it is based on the needs of the community. Our data finds evidence of this. From the baseline survey, we can observe that the villages that chose to undertake education-related sub-projects had almost 27 percentage points lower levels of satisfactory access to primary education compared to villagers who chose other projects to participate in. Thus, in terms of targeting, the RDP performed well according to its mandate. In terms of delivery, however the performance has been mixed. Although there has been only a marginal increase of 1.6 % satisfactory access, the overall quality of access has improved with almost 50 % of the population stating that they have good access compared to 32 % in the baseline. Furthermore, there has been an increase of 8 percentage point in dissatisfactory access to primary education in villages choosing not to undertake education-related project. This highlights the success of the RDP. However, approximately 36 % of people have not benefitted from RDP improvements (Table 10). Table 10: Impact of RDP on access to primary education At least Satisfactory Access Poor Access B W1 D B W1 D Non RDP 90.4% 82.4% -8.9% 9.6% 17.6% 84.2% RDP 62.8% 63.8% 1.6% 37.2% 36.2% -2.7% Good Satisfactory Non RDP 52.6% 64.9% 23.4% 37.8% 17.5% -53.8% RDP 31.9% 49.2% 54.5% 31.0% 14.6% -52.8% Source: Wave 1: Household- D2 & Community Leader Survey. Baseline: Houshold-B21 31 5.1.2) Access to Health Services There have been improvements in the satisfactory access to health care services between baseline and follow-up surveys in two of the four study provinces. The degree of improvement was highest in Malaita (41.5 %) and lowest in Temotu, which registered a 51 % decrease in the indicator (Figure 10). Choiseu province saw an 11.3 % decline whereas the Western province experienced a 4.2 % improvement in access to health services. The results from the 70 % cut-off indicator depict a better overall picture. There are improvements in all of the study provinces barring Temotu. Indeed, if we restrict our sample to omit Temotu province, the magnitude of improvement is around 30 % (Appendix 2).This report finds an overall decrease of 6.5 % in the proportion of villages with satisfactory access to health services. Figure 10: Percentage of villages with satisfactory access to health services 80% 70% 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western Full Full w/o Temotu Baseline Wave 1 Source: Wave 1: Household- D2. Baseline: Houshold-B21. There has been a decrease in the satisfactory access to health services in two of the four provinces. This drop was the highest in Temotu where almost 68 % of the households reported having poor access to health facilities, an increase of 72.6 % compared to the baseline. There have been improvements in both the quality and quantity of access to health care in Malaita and Western provinces. In Malaita, 47.5 % and 22 % of the people report having good and satisfactory access to health services. In Western province, the corresponding figures are 37.9 and 33.3 %, respectively. In fact, aside from Temotu, all the other three provinces recorded improvements in the quality of access to health care ( 32 Table 11). Table 11: Access to Health Care (Household data) Good Satisfactory Poor Province B W1 D B W1 D B W1 D Choiseul 25.5% 40.5% 59.1% 37.3% 15.8% -57.6% 37.5% 43.7% 16.4% Malaita 15.1% 47.5% 215.2% 25.2% 22.0% -12.7% 41.1% 30.5% -25.8% Temotu 21.3% 18.2% -14.4% 39.0% 13.6% -65.0% 39.5% 68.2% 72.6% Western 29.6% 37.9% 28.2% 40.5% 33.3% -17.8% 29.6% 28.7% -2.9% Full 23.7% 36.0% 51.6% 37.2% 20.9% -43.8% 39.0% 43.2% 10.6% Full w/o Temotu 24.7% 42.2% 70.7% 36.4% 23.4% -35.7% 38.9% 34.4% -11.5% Source: Wave 1: Household- D2 & Baseline: Houshold-B21 In terms of targeting as well as performance, the RDP villages did well. On average, villages that chose to undertake health projects had poor access to health services (56.5 %) compared to villages that chose to undertake other projects (31.2 %). In terms of performance, the villages that chose RDP health projects registered an improvement of 36 % in satisfactory access. This increase is quite remarkable given that the villages that chose not to undertake health projects registered a deterioration of 18 %. There were improvements in terms of quality as well. RDP villages, on average, improved good access by 205 % and at the time of the follow-up survey, almost 50 % of households in these villages reported that they had good access to healthcare. The corresponding figure for non-RDP villages is 15.4 percentage points lower, at 34.1 % (Table 12). Table 12: Impact of RDP on access to health centre At least Satisfactory Access Poor Access B W1 D B W1 D Non RDP 68.8% 56.5% -17.9% 31.2% 43.5% 39.4% RDP 43.5% 59.1% 35.9% 56.5% 40.9% -27.7% Good Satisfactory Non RDP 32.1% 34.1% 6.0% 36.6% 22.4% -38.8% RDP 16.2% 49.5% 204.7% 27.3% 9.7% -64.5% Source: Wave 1: Household- D2 & Community Leader Survey. Baseline: Houshold-B21 33 5.1.3) Access to Roads There have been substantial improvements in the satisfactory access to roads in all study provinces. On average, 49.4 % of villages indicated that they have satisfactory access to roads in the follow-up survey compared to a 32 % in the baseline survey. This translates to a 54 % increase. The degree of improvement is the highest in Choiseul where 25 % villages have satisfactory access to roads compared to 0 % in the baseline. The improvement is slowest, not surprisingly, in Malaita as the province originally had good levels of access to roads shown in the baseline (Figure 11). Figure 11: Percentage of villages with satisfactory access to roads 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western Full Full w/o Temotu Baseline Wave 1 Source: Wave 1: Household- D2. Baseline: Houshold-B21. 5.1.4) Access to Markets Figure 12 presents the proportion of villages with satisfactory access to markets at both baseline and follow up. From the table, we can observe that access to markets has improved in all provinces except Temotu, where there was a 72 % decrease in the proportion of villages with satisfactory access to markets. The level of improvement was the highest in Choiseul, followed by Western and Malaita. However, it is important to note that in the Western province, 37 % of the villages still do not have satisfactory access to markets. On average, there has been a 40 % increase in the proportion of villages with satisfactory access to markets in these three provinces. Our robustness results verify our main observations that access has improved in all provinces except Temotu (Appendix 2) Figure 12: Percentage of villages with satisfactory access to markets 34 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western Full Full w/o Temotu Baseline Wave 1 Source: Wave 1: Household- D2. Baseline: Houshold-B21. Access to markets has shown mixed improvements. On one hand, satisfactory access increased in Choiseul and Western provinces whereas, on the other, it decreased in Malaita and Temotu provinces. In terms of quality, households in Choiseul reported having the highest level of satisfaction (68.3 %) followed Malaita (47.2). Less than a quarter of households in Temotu and Western province reported having good access to markets. In particular, it is alarming that around 85 % of households in Temotu province reported having poor access to markets (Table 13Error! Reference source not found.). Table 13: Access to Markets (Household Data) Good Satisfactory Poor Province B W1 D B W1 D B W1 D Choiseul 7.4% 68.3% 822.6% 37.3% 11.6% -68.8% 54.8% 20.1% -63.3% Malaita 28.5% 47.2% 65.7% 33.4% 27.9% -16.5% 19.5% 24.9% 27.9% Temotu 18.8% 4.7% -74.7% 18.8% 10.5% -43.8% 61.0% 84.7% 38.9% Western 15.9% 22.8% 43.2% 43.0% 39.7% -7.6% 40.2% 37.5% -6.7% Full 18.6% 36.7% 97.4% 34.1% 21.3% -37.3% 46.6% 42.0% -9.8% Full w/o Temotu 18.5% 48.2% 160.3% 40.3% 25.2% -37.3% 40.7% 26.6% -34.7% Source: Wave 1: Household- D2 & Baseline: Houshold-B21 5.1.5) Access to Sanitation Facilities The overall level of access to sanitation is very poor in the survey areas. Only 9 % of villages across the four study areas have satisfactory access to sanitation facilities (Figure 13). There is disparity in access across location as well, with no villages in Malaita and Temotu registering satisfactory access to sanitation. Western province registers that 26% of its villages with having satisfactory access to safe sanitation facilities. The results from the robustness indicator also present the same dire conditions. These improvements however have very little meaning because of the aforementioned poor general conditions that still exist. 35 Figure 13: Percentage of villages with satisfactory access to sanitation facilities 30% 20% 10% 0% Choiseul Malaita Temotu Western Full Full w/o Baseline Wave 1 Temotu Source: Wave 1: Household- D2. Baseline: Houshold-B21. The level of access to sanitation facilities in the Solomon Islands is very poor. At the time of the follow-up survey, 92 % of households across all four provinces reported having poor access to sanitation facilities. The level of access is highest in Western province with 33.5 of the households reporting satisfactory access to sanitation facilities. Improvements have also been very slow with a 10 % decrease in poor access between baseline and the follow-up surveys (Table 14). Table 14: Access to sanitation facilities (Household Data) Good Satisfactory Poor Province B W1 D B W1 D B W1 D Choiseul 0.3% 13.2% 4703.4% 5.2% 7.9% 51.6% 94.8% 79.0% -16.7% Malaita 1.1% 1.5% 38.7% 1.9% 9.6% 400.6% 78.4% 88.9% 13.4% Temotu 0.8% 2.5% 237.3% 5.5% 2.0% -63.3% 93.8% 95.5% 1.8% Western 7.8% 22.6% 190.2% 10.0% 11.9% 19.0% 82.6% 65.5% -20.6% Full 2.4% 9.6% 301.1% 5.8% 7.7% 33.6% 92.0% 82.7% -10.1% Full w/o Temotu 3.1% 12.0% 294.5% 5.9% 9.7% 64.9% 91.3% 78.2% -14.3% Source: Wave 1: Household- D2. Baseline: Houshold-B21. Table 15 reports the type of sanitation facilities used by different households. The statistics from the table reinforce the fact that sanitation needs to be identified as a priority area for future development. During the follow-up survey, almost three quarters of the respondents reported defecating in open spaces. In fact, only 20 % of the people on average use toilets with flush or pour flush. There has been very slow progress across time as well with very little change in access to safe sanitation facilities (6 percentage point improvements) between baseline and follow-up surveys. 36 5.1.6) Access to Electricity There have been improvements in access to electricity in three of the four study provinces (Error! Reference source not found.). During the baseline survey, no villages in Choiseul and the Temotu provinces had satisfactory access to electricity, while the corresponding numbers in the follow-up survey were 95 and 25 %, respectively. The Western province also improved its access from 12 % to 95 %. However, there has been a decline in access to electricity in Malaita from 7 % of villages having satisfactory access to a reduced 5 %. The robustness results also validate our main observations (Annex 8). Table 15: Type of Sanitation used by households Percentage of Households Using* Province Pour Flush Pit Bush River/sea Other flush Baseline Choiseul 0.8% 8.5% 5.8% 21.4% 85.5% 0.8% Malaita 2.0% 2.7% 33.0% 3.7% 64.3% 0.7% Temotu 0.0% 1.2% 4.0% 18.4% 80.8% 4.7% Western 8.7% 18.9% 3.7% 21.4% 72.4% 2.5% Full 2.6% 7.6% 10.6% 16.9% 78.9% 2.4% Full w/o Temotu 3.7% 10.2% 13.4% 16.3% 78.1% 1.4% Wave 1 Choiseul 5.4% 16.8% 0.5% 1.1% 76.2% 0.0% Malaita 1.0% 6.7% 10.3% 6.2% 73.8% 2.1% Temotu 1.0% 3.0% 0.5% 16.7% 78.3% 0.5% Western 8.5% 20.3% 2.3% 4.0% 65.0% 0.0% Full 3.8% 11.4% 3.4% 7.2% 73.5% 0.7% Full w/o Temotu 4.8% 14.4% 4.5% 3.8% 71.8% 0.7% * This is a multiple response question so all columns will not add up to 100 % Source: Wave 1: Household- B12 & Baseline: Houshold-B10 Figure 14: Percentage of villages with satisfactory access to electricity 37 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western Full Full w/o Baseline Wave 1 Temotu Source: Wave 1: Household- D2. Baseline: Houshold-B1. There have been drastic improvements in the access to electricity in Choiseul and Western provinces and modest improvements in Malaita and Temotu provinces. On average, there has been a 46.6 % reduction in poor access to electricity across the provinces in the Solomon Islands. There has been an increase in quality as well, with 36 % of households reporting having good access to electricity. Table 16: Access to electricity (Household Data) Good Satisfactory Poor Province B W1 D B W1 D B W1 D Choiseul 0.0% 83.0% 4.9% 9.0% 83.3% 94.8% 8.0% -91.6% Malaita 3.3% 8.2% 149.7% 1.1% 6.7% 508.6% 96.7% 85.1% -11.0% Temotu 2.8% 4.1% 47.6% 1.0% 22.8% 2184.0% 85.5% 73.1% -14.5% Western 5.6% 51.7% 821.8% 15.0% 30.9% 106.6% 77.6% 17.4% -77.5% Full 3.0% 35.9% 1110.3% 5.4% 17.2% 220.5% 88.0% 47.0% -46.6% Full w/o Temotu 3.1% 47.1% 1442.0% 7.1% 15.2% 112.7% 89.0% 37.8% -57.5% Source: Wave 1: Household- D2 & Baseline: Houshold-B21 This is perhaps reflected in the statistics from Annex 8. From the table, we can see that more than 80 % of households during the follow up used solar lamps to light their houses. This is a drastic improvement compared to 6.3 % utilisation of solar lamps during the baseline. It is not surprising, as a result, that the usage of kerosine lamps has decreased from 92.9 % to 17 % during this time period. Table 17: Type of electricity used by households for lighting purposes Percentage of households using* Province Solar Generator Torch Lamp Other Baseline 38 Choiseul 6.4% 1.6% 39.2% 99.5% 1.1% Malaita 5.7% 6.0% 65.4% 85.9% 1.7% Temotu 8.0% 2.4% 25.1% 94.5% 1.2% Western 5.3% 9.9% 43.4% 98.0% 1.3% Full 6.3% 5.0% 46.7% 92.9% 1.4% Wave 1 Choiseul 82.6% 3.2% 9.0% 13.7% 1.1% Malaita 86.5% 4.0% 10.5% 8.0% 4.0% Temotu 80.8% 0.5% 34.3% 13.1% 1.5% Western 76.4% 9.3% 21.4% 35.7% 1.1% Full 81.7% 4.2% 18.8% 17.3% 2.0% * This is a multiple response question so all columns will not add up to 100 % Source: Wave 1: Household- B10 & Baseline: Houshold-B11 The usage of solar energy for other electricity needs of the household use other than lighting has also improved. Compared to the baseline, 65.5 % point more households use solar for electricity needs. The usage rate of generators has also increased by 4.2 percentage points (Table 18). Table 18: Type of electricity used by households for other purposes Solar Generator Province B W1 D B W1 D Choiseul 5.8% 82.6% 1324.1% 5.8% 5.3% -9.3% Malaita 4.4% 77.5% 1661.4% 4.4% 15.0% 240.9% Temotu 7.4% 52.5% 609.5% 7.4% 1.5% -79.7% Western 5.3% 71.9% 1256.6% 5.3% 17.6% 232.1% Full 5.5% 71.0% 1190.9% 5.5% 9.7% 76.4% Source: Wave 1: Household- B11 & Baseline: Houshold-B12 In relation to the impact of the RDP, the evidence has been mixed, both in terms of targeting as well as performance. During the baseline, satisfactory access to electricity was 2 % in the villages that chose to undertake electricity projects compared to 5 % in the villages that did not. Given these low levels of satisfactory access to electricity in general, very little can be said about targeting. In terms of performance, the villages that did not choose to undertake RDP projects did better. On average, satisfactory access increased 49 percentage points in villages that chose not to pursue solar projects compared to 30 percentage point increases in villages that decided to pursue solar projects. In terms of quality, the villages which did not choose to undertake solar projects also performed better (Table 19). Table 19: Impact of RDP on access to electricity 39 At least Satisfactory Access Poor Access B W1 D B W1 D Non RDP 5% 56% 1051.8% 95% 44% -54.0% RDP 2% 32% 1553.3% 98% 68% -30.7% Good Satisfactory Non RDP 2% 40% 1712.9% 3% 16% 507.4% RDP 1% 9% 1295.0% 1% 23% 1682.5% Source: Wave 1: Household- D2 &Community Survey. Baseline: Houshold-B21 Finally we look into who have actually funded these projects. Figure 15 shows the breakdown of who funded projects when households reported that their access to electricity has improved. From the figure we can see that the main driving force behind the improvements in access to electricity has been the members of the parliaments. In almost all the provinces over the last few years MPs have been distributing solar lamps. Thus our finding is not surprising. Figure 15: Funding of Electricity improvements. 60% 50% 40% 30% 20% 10% 0% RDP Community MP SIG Self Other Source: Household Questionnaire D3 and D4 5.1.7) Construction of the variable to measure achievements. The criteria set in evaluation framework for RDPs progress to infrastructure is to achieve 10 percent improvements in proportion of villages with satisfactory access to infrastructure between the baseline and the follow-up survey. This poses a challenge because every village has at most had one RDP infrastructure project. Thus it should only be expected that a village's access should only increase in the particular type infrastructure that RDP has funded i.e. if a school is built in a village then that villages access to schooling should increase. To come up with one 40 variable which captures overall access to infrastructure is thus challenging. Moreover, the proportion of each type of project to that of the total project is also not equal for each project. For example, there has only been one RDP project related to improvements in road access where as there have been numerous projects aimed at improving water supply. In view of these facts, we created an index variable. In creating this index variable we only used information from access to water, health, education and electricity infrastructure. This is because these projects accounted for 90 percent of all RDP projects. We used a weighted average to calculate a villages infrastructure access. The highest value this indicator variable could take was 1, which would indicate that every individual that was surveyed in this village had at least satisfactory access to all infrastructure. The lowest value conversely was 0, which would indicate no one would have satisfactory access to any infrastructure. The formula for how the index was created is given below Here the is the weight given to that infrastructure . For our calculations we used weights of 0.215, 0.360, 0.159 and 0.167 for school , water, health and electricity. These values were computed on the basis of the intensity of type RDP projects in the sample villages. For example of the total number of villages that chose to do any one of these four types of projects 21.5 percent chose schooling project, 36 percent chose water projects, 15..9 percent chose health and 16.7 chose electricity projects. Based on this we created the target indicator. Improvements in this indicator is presented in Table 20. Table 20: Indicator 2 improvements Cumulative PDO Level Results Unit of Percentage Target Values Baseline Evaluation Indicators Measure Change Yr Yr Yr Yr 1 3 5 7 Indicator Two: Percent 34.2 65.8 92.4% 10 Percentage of villages with satisfactory access to infrastructure 41 From the table we see that RDP has achieved its targets. Proportion of villages with satisfactory access to infrastructure has increased from 34 percent to 65.8 percent, a 92.4 percent increase. Thus in summary, the performance of the RDP in improving access to infrastructure has been good. 5.2.1) Temotu Examined The Temotu province throughout this report has shown either no improvement or a decrease in access to infrastructure. This is a worrying trend and requires further analysis. This section will examine differences between the baseline and the follow up survey for the Temotu province to understand this decline. Of the 20 villages surveyed during the follow up in Temotu 19 were also surveyed in the baseline. This provides us with an important comparison sample. At the baseline 2779 household members are reported to reside in surveyed villages compared to 7917 in the follow up an increase of 180%. In terms of other comparable village characteristics, at the baseline each village had on average 3.5 businesses while at the time of the follow up Temotu had on average 11.45 businesses. This is a significant change especially in terms of population which may have put a strain on infrastructure access. Temotu’s poor performance may in part be due to the decline of a small number of villages, these outliers may not be a true reflection of outcomes throughout Temotu. To examine this we report access to primary school, health services, water supply and roads between villages. Satisfactory access to primary school declined in 8 villages. The decline was most significant in Malapu, Lipe, Bania and Taipe. Lipe for example at the baseline reported 70% satisfactory access compared to 20% at the time of the follow up. Satisfactory access to health services declined in 12 villages throughout Temotu. The decline was largest in Malapu, Nyimoa, Manoputi, Balo, Noipe, Bania and Manevi. Strikingly, 100% of Bania’s surveyed baseline households reported satisfactory access to health services compared to 0% at the follow up. Satisfactory access to water declined in 14 villages. Nearly all households in these villages reported access to water was poor. Finally, satisfactory access to roads decreased in half of all surveyed villages. However, the decline was less than 15 percentage points in nearly half of villages. The largest declines were seen in Malapu, Nyimoa, Laro, Mrnrg and Bania. These results provide evidence that poor access to primary schooling and roads can be explained by a small number of outlier villages (4 villages in the case of schooling and 5 villages in the case 42 of roads). A number of characteristics appear similar between these villages: 1) Nearly all villages experienced high levels of rain over the last year. This may have destroyed roads and reduced access to primary schools. 2) The population of these villages increased by more than the province average. This increase in population could strain access to services such as schools. The finding above however, show that declines in access to health services and water occurred in over half of all villages and thus, poor service access is not driven by a small number of villages. Another explanation for this decline may be due to misreporting of access by households. This could occur for instance if households were not selected randomly, or differ in their definition of access. This issue can be examined further by comparing these household results with the community leader survey in which community leader groups were asked if services had improved in the village in the last four years. In terms of health services 35% of community groups in Temotu identify that access improved over the last four years. While, 15% of community groups in Temotu reported improvements in access to water supply.. These results are consistent with the household survey and provide evidence that community leader groups and households rank access to infrastructure in a similar manner. Table 21: Temotu’s Access to Services and Infrastructure by Village Village Primary School Health service Water supply Roads BL Follow Up BL Follow Up BL Follow Up BL Follow Up At least At least At least At least At least At least At least At least satisfactory satisfactory Satisfactory satisfactory satisfactory satisfactory satisfactory satisfactory Malapu 70% 30% 70% 10% 60% 6% 39% 0% Nifiloli 100% 100% 26% 0% 49% 0% 6% 0% Nyimoa 100% 90% 45% 0% 48% 0% 32% 0% Lipe 70% 20% 0% 0% 40% 0% 15% 23% Laro 74% 60% 5% 0% 10% 0% 21% 0% Otelo 100% 100% 85% 73% 45% 16% 47% 91% Nibana Nende 0% 0% 0% 0% 40% 0% 0% 0% Ngamubuluo 85% 90% 25% 30% 60% 0% 25% 10% Manoput 95% 100% 90% 12% 66% 0% 100% 100% Banua 100% 100% 100% 80% 30% 67% 89% 80% Mrnrg/Mrnar 100% 100% 100% 70% 20% 65% 100% 80% Balo 100% 80% 100% 40% 20% 90% 95% 100% 43 Nrrle 100% 100% 100% 90% 30% 85% 95% 90% Noipe 100% 100% 64% 0% 75% 0% 87% 75% Bania 100% 50% 100% 0% 60% 0% 100% 60% Taipe 86% 50% 29% 0% 70% 93% 13% 45% Nangu 100% 100% 84% 90% 10% 20% 26% 100% Neo-Malo 94% 100% 82% 100% 70% 0% 0% 85% Minevi/Maglalo 79% 50% 79% 10% 30% 0% 0% 66% Source: Baseline: Household- H21. Wave 1: Household- D2. Note: At least satisfactory is defined as satisfactory or better access to services. It is possible that the decline in access may be due to enumerator bias. To examine this issue, we report results for each enumerator in Temotu. Table 22 examines the difference in responses per village by enumerators. Results show significant differences between enumerators across villages. Differences are largest in Manoputi, Nifoli and Neo-Malo. These results imply, that the decline in Temotu may be partly driven by one enumerator who may have underestimated community access to infrastructure in a number of villages. Table 22: Differences Across Enumerators Primary School Health service Water supply Electricity Village Follow Up Follow Up Follow Up Follow Up Tina Roxon Tina Roxon Tina Roxon Tina Roxon Malapu 75% 0% 25% 0% 0% 100% 0% 17% Nifiloli 75% 100% 0% 0% 0% 100% 0% 33% Nyimoa 25% 25% 0% 0% 0% 0% 25% 75% Lipe 0% 33% 0% 0% 0% 66% 0% 17% Laro 0% 50% 0% 0% 0% 17% 0% 0% Otelo 60% 100% 60% 100% 0% 75% 60% 75% Nibana Nende 0% 0% 0% 0% 0% 58% 33% 0% Ngamubuluo 75% 17% 50% 18% 25% 84% 75% 50% Manoput 100% 100% 20% 100% 20% 100% 0% 25% Banua 63% na 70% na 25% na 12% Na Mrnrg/Mrnar 66% na 67% na 11% na 11% Na Balo 50% 50% 50% na 12% na 0% Na Nrrle 75% na 100% na 12% ma 12% Na Noipe 100% 80% 0% 0% 0% 80% 0% 0% 44 Bania 0% 44% 0% 0% 0% 86% 0% 0% Taipe 25% 17% 0% 0% 50% 84% 0% 50% Nangu 100% 84% 100% 84% 0% 17% 0% 0% Neo-Malo 100% 66% 100% 100% 25% 100% 0% 50% Minevi/Maglalo 50% 17% 0% 0% 0% 50% 0% 50% In summary, results are unclear as to the reason why access to infrastructure declined in Temotu. For some services such as schools and roads, the decline in access can be attributed to a small number of villages which suffered high rain and in some cases earthquakes and tsunamis. However, the decline in health services and water supply, occurs in most villages and cannot be explained by survey results. This area should be examined for further research. Comparing survey results with NGO’s such as World Vision which run water projects in Temotu would be a start and allow survey results to be checked for robustness. Consultation should also occur with the team leader and CH’s to gather other opinions as to causes of these poor results. 45 5.3) RDP Processes 5.3.1) Sub-project Management Table 23: Intermediate Results Indicator Six. PDO Level Results D=Dropped Unit of Cumulative Target Values Indicators C=Continue Measure Evaluation N=Revised Baseline R=Revised Yr Yr Yr 5 Yr 7 Score 1 3 Intermediate Results N Percent Zero 93.910 80 Indicator six: Percentage of households in communities receiving project funding that feels project investments reflect their needs Indermediate Results N Percent Zero 93.4 80 Indicator six b):…of which are women Intermediate results indicator six asks what percentage of households in the communities received project funding and felt the project investments reflected their needs. To answer this, we used household question F5- Overall, how satisfied were you with the way in which the was selected? Respondents often perceived this question as an evaluation of the sub-project rather than a question on the selection process11. Figure 16 shows that overall, 81.4% of the surveyed households were very satisfied with the sub-project while 12.5 % were a little satisfied and 6.1% were not satisfied. Malaita had the highest level of dissatisfaction with 8.7% of households not satisfied with the sub-project. In comparison, Choiseul had the highest level of satisfaction (93%). The report finds that RDP reached its target. On average 94% of households felt projects investments reflected their needs. 10The follow-up survey does not directly answer this question. Instead, we use household survey question F5- Overall, how satisfied were you with the way in which the was selected? Arguably, respondents tended to answer this question based on what they thought of the sub-project, rather than the process. 11 Evidence of this comes from the general observations by the field team 46 Figure 16: The proportion of households that were satisfied with sub-project selection methods. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Very Satisfied A little satisfied Not Satisfied Source: Household Questionnaire F5 Figure 17 shows the results for female respondents only. The outcome is very similar to all households shown above. On average, 80.7% of female respondents reported they were very satisfied with the sub-project, 12.7% were a little satisfied and 6.7% were not satisfied. At the province level, Temotu had the highest level of dissatisfaction with 10.7% of female respondents not satisfied with the sub-project. The report finds that the RDP reached its target. On average 94% of women feel projects reflect their needs. Figure 17: The proportion of female household (respondents) that were satisfied with the sub-project selection methods. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Very Satisfied A little satisfied Not Satisfied Source: Household Questionnaire F5 5.3.2) SIC Management An important component of the RDP process is the Sub-Project Implementation Committee (SIC). The SIC is in charge of organising and managing the sub-project from design through to organizing maintenance/operation of infrastructure/service after subproject completion. SIC 47 membership is minimally composed of a (i) Chairman, (ii) Secretary, (iii) Treasurer; the Technical Officer Representatives of different interest groups within the community (women, youth, clans etc.) Figure 18: The proportion of households satisfied with the SIC 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Very satisfied Little satisfied Not Satisfied. Source: Household-G6 Figure 18 shows, on average, 66 % and 20 % of households were either very or partly satisfied with the SIC committee respectively. On the other hand, 13% of households reported that they were not satisfied with SIC. Large differences existed between provinces: Temotu had the highest level of dissatisfaction (18%), while Choiseul had the lowest (8%). Figure 19: SIC proposed changes 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All No Change More Women More Men More Youth Source: Household-G7 Households were subsequently asked if they thought the composition of the SIC should be changed and if so, how. Nearly half (46.5%) of the households reported the composition of the committee should not change. Of those that reported that the composition should change, 27% indicated that the committee should consist of more women members. This is especially prevalent in Malaita, where 43% reported the need to increase female participation. In contrast, 6% of households believe more men are needed on the SIC committee. It should also be noted 48 that Temotu differs in this respect: 40% of households reported that the SIC should have more men. Finally, on average, 14% of households reported more youths are needed on the SIC committee. The involvement of all members of the community, including women and youths, is vital to ensure community ownership and sustainability. These results infer that the composition of the SIC is lacking both women and youths. Interestingly, men are often the strongest promoter of women’s involvement. On average, 25% of males reported the need to increase women’s involvement. Participation and its associated benefits will be discussed further in section 5. Questions about what the SIC committee did well and could have done better were also asked during the survey. Error! Reference source not found. reports household opinions on what tasks the SIC performed well in. Nearly 20% of households reported the SIC performed well in tasks related to financial and project management. Nearly 65% of all households reported satisfaction with the SIC’s ability to motivate and coordinate the community while 46% believed the SIC was able to communicate progress to the community. Results were similar across provinces. Figure 20: What the SIC did well. 80% 70% 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All Financial/Project Management Motivate/Coordinate Community Communicating progress to community. Source: Household-G8 Figure 21 reports households’ suggestions on how the SIC could have done better. In nearly 40% of cases, households reported dissatisfaction with the SIC’s financial and project management 12 . This issue was more widespread in Temotu (45.3%) and Choiseul (60%). A number of households also reported that motivating and coordinating the community could also have been improved (28%). Furthermore, households also reported the need to increase communication and cooperation between SIC and the community. Interestingly, a number of 12 4.7% of households report both satisfaction and dissatisfaction with SIC’s financial and project management. 49 households reported both dissatisfaction and satisfaction with the same issues (see Figure 20 and Figure 21). This could indicate that the categories may have been too broad as the SIC may have performed well in some sub-areas but less well in others. Figure 21: Improvement suggestions for the SIC. 70% 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All Financial/Project Management Motivate/Coordinate Community Communicating progress to community. Source: Household-G9 Community leader discussion groups were also asked a number of questions based on the SIC. The groups were firstly asked if having an SIC was an effective way of coordinating the sub- project implementation. Overall, 90% of groups reported that the SIC was an effective way of coordinating sub-projects. The results were consistent across provinces (see Table 24). Table 24: The proportion of community leaders that reported the SIC is an effective way of coordinating the sub-project. Province SIC effective Choiseul 90% Malaita 90% Temotu 94% Western 88% All 90% Source: Community-E6 The groups were further asked if they could recommend a better option than having SIC (Table 25). The responses to this question indicate that respondents were confused: majority of comments that were made suggest methods of improving the SIC rather than providing alternatives. Nevertheless, we categorised all valid comments into three general fields. In Temotu, the groups reported the SICs should add committee members, that is, either adding 50 new members such as women or additional committee members from the community. Overall, 16% of leader groups reported that there should be no change and that the SIC is ‘the best option’. The last category is a “New SIC”. This indicates 20% of all leader groups believe the SIC would be better run under an alternate committee. The majority of suggestions are to use the church committee or any other existing committee instead of the SIC. Table 25: Community Leader suggestions on better options than a SIC Province Add Committee Use Existing New SIC Choiseul 0 5% 40% Malaita 0 55.0% 30% Temotu 35% 5% 10% Western13 0 0 0 All 8.7% 16.2% 20% Source: Community-E7 In summary, the villagers were relatively satisfied with the SIC committee and believed it was an effective method of coordinating the sub-project, however there appears to be a greater need for female and youth involvement. This will lead to greater consensus within the community, ensuring community ownership, participation and sustainability. Households, also believe, that financial and project management and community coordination needs improvement. 5.3.3) Implementation challenges The community leader discussion groups in each village were asked to name the two biggest challenges to sub-project implementation. Figure 22 presents the results. Overall, the biggest challenge appeared to be community participation (55%). This was closely followed by purchasing raw materials (40%) and RDP procedures (31%). At the province level, the results are similar with the biggest impediment to implementation in Choiseul, Malaita and Western being community participation while in Temotu, RDP procedures were reported to provide the biggest challenge. 13 9/20 community groups responded to this questions. However, the answers to this question were not able to be categorised as the answers varied and did not relate to the actual question. 51 Figure 22: Sub-project implementation challenges in incomplete sub-projects 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Community. Contractor Purchasing Manage Fin. Raw SIC not RDP Community Part. raw mat. materials working procedures politics Contribution Choiseul Malaita Temotu Western All Source: Community-E27 As we show later in Figure 28, a large proportion of households reported RDP projects were either incomplete or not functioning. It is therefore important to directly examine the biggest challenges faced by incomplete projects. Results in Figure 22, show that villages with incomplete projects suffered from similar challenges faced by the average villages. Community participation and purchasing raw materials was the biggest challenge implementing a sub-project in both these types of villages. Figure 23: Households’ suggested changes to RDP processes and procedures to improve sub-project implementation. 80% 70% 60% 50% 40% 30% 20% 10% 0% Timing Funding limits Information flow Accountability of Procurement No Changes to and from the committees system community Choiseul Malaita Temotu Western All Source: Household- G15 Households were also asked about RDP processes and what they would change. Subsequently, they were asked to outline procedures to improve sub-project construction and implementation. 52 Results differed significantly across provinces (Figure 23). The surveyed households in Western province indicated timing14, information flow to and from the community and funding limits should improve. Temotu households overwhelmingly reported the procurement system needed the most improvement. On the other hand, the surveyed households in Malaita indicated the funding limits, timing and accountability of the committees needed improvement. Lastly, Choiseul, similar to Temotu, reported procurement and timing were in need of change to improve sub-project construction and implementation. In summary, the biggest challenge to RDP project implementation appears to be community participation. Over 80% of community leader groups reported either poor community participation or the contribution of raw materials by communities were significant impediments to project implementation. These results are true across completed project villages and incomplete project villages. Another important impediment in RDP procedures is timing and funding limits. Improvement in these areas will have significant impacts on the implementation of projects by improving the timeliness of project completion and thus would improve the proportion of sub-projects which have been completed and are functioning. 5.3.4) SIC Bank Account Figure 24: The proportion of households that reported a benefit in the SIC having a sub- project bank account. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All SIC bank benefit Source: Community- E58 Issuing individual bank accounts to communities was not only an important method of distributing RDP funding but also a method enabling villagers to learn financial skills. Bank accounts were opened by CHs in most communities to provide SIC access to implementation funds. 14 Timing is not broken down into specifics in the household survey. 53 To ascertain the usefulness of bank accounts, all community leader groups were asked if there was a benefit for the SIC in having a sub-project bank account. The results are reported in Error! Reference source not found.. Overall, 78% of surveyed village leader groups reported a strong benefit in having a sub-project bank accounts. Figure 25: Benefits of the SIC bank account. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Learnt some accounting Learnt to use cheques Easier to use money Source: Community- E59 Village leader groups were further asked to identify the benefits of having a SIC bank account (Figure 25). Over half of the groups identified that bank accounts improved the community members’ book keeping skills. Furthermore they indicated that as a result of these bank accounts members got the chance to use cheques. However, these results varied significantly between provinces. The biggest impact of having an SIC bank account appeared to be the benefits of learning accounting in both Choiseul, Malaita and Temotu, while in Western province as well as Malaita, bank accounts made it easier for the SIC to access money. Table 26: The proportion of community leader groups who reported the community plan to keep a bank account after the RDP was completed Province Yes Maybe No Obs Choiseul 41% 11% 48% 17 Malaita 60% 33% 7% 15 Temotu 88% 5% 7% 18 Western 42% 0 58% 14 All 59% 12.5% 28.5% 64 Source: Community-E61 Table 26 reports the proportion of community groups that plan to keep a bank account after RDP. Overall, more than half of the villages reported they were likely to keep a bank account after the RDP while 28.5% will not. These results, along with the other results reported in this 54 section, suggest SIC bank accounts were directly correlated with the transfer of skills to the community and were often seen as beneficial to the community. 5.3.5) Ward Development Committee A Ward Development Committee (WDC) was established in each ward, by the decision of an open community meeting in which different clans, associations, women and minority/vulnerable groups were present. WDC was composed of: the Member of Provincial Assembly (MPA); a representative of a traditional authority; a Church representative; a women’s representative; and male and female youth representatives. Other representatives could have been be nominated by the communities. The WDCs usually met in an “open session” once per year to review the sub- project requests submitted by each village in the ward. In addition, they were responsible for considering cost estimates and preparing ward infrastructure maps. All households were asked if they knew any members of the WDC. Overall, 40% of households reported they knew at least one member of the WDC (Figure 26Error! Reference source not found.). Two follow-up questions were asked to those households responded that knew WDC members. Firstly, they were asked if they ‘believed the RDP Ward Development Committee properly represented groups in the village?’ . Figure 27 reports that 75% of households believed the WDC committee properly represented the community. Results are less than homogenous at the province level. Close to half (52%) of the Temotu households believe that the WDC does not properly represent their village. It is possible that these results were driven by the sub-section of households that responded that they knew WDC members. Households, for example, may know WDC members because they had disagreements or similarly live in the village, leading to a bias in the result. Therefore, these results should be interpreted with caution. Figure 26: The proportion of households that they knew members of their RDP WDC. 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All Source: Household-G2 55 Figure 27: The proportion of households who believed the RDP WDC properly represents groups in their village. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Source: Household-G3 A quarter of households indicated that more women are needed in the WDC when asked if the WDC membership should change (Table 27), while 33% believe the composition of the WDC should not change. At the province level, 15% of households in Choiseul indicated no change in composition should occur compared to 63% in Western province. Further, on average, 19% of households believed the WDC should comprise more youths. In summary communities appear to be satisfied with the WDCs. However, satisfaction was lower in both Temotu and Choiseul provinces. Results also show that WDCs should include more women and youth especially in Temotu and Malaita provinces. Table 27: Ward Development Committee membership composition Province No change More Women More Men More Youths Obs Choiseul 14.21% 7.8% 0.5% 4.2% 190 Malaita 44.5% 43.6% 40% 38.1% 110 Temotu 33.3% 50.9% 9.8% 50.9% 51 Western 63.9% 16.2% 9.3% 6.9% 86 All 33.8% 23.5% 13.2% 18.7% 437 Source: Household-G4 5.3.6) Sub-Project Quality Figure 28 examines the proportion of sub-projects which have been completed and are functioning in a village. On average, 52% of households reported projects were completed and functioning in their village. However, there was heterogeneity across provinces. In Western province, 30% of projects were completed and functioning while 45% were completed in Temotu. This contrasts with 69% in Choiseul and 62% in Malaita. In 35% of villages, 56 households reported contradictory results: for these villages some households reported projects were complete and functioning while some reported otherwise. This may be due to households’ varying definitions of a complete and functioning project. It also may be due to a lack of communication within villages. Figure 28: The proportion of sub-projects that were completed and functioning. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Source: Household-E1 Households that reported a project was completed and functioning were subsequently asked if they were satisfied with the quality or standard of the sub-project construction (Figure 29). More than half of all the households reported that they were very satisfied with the construction quality while 20% were partially satisfied. In contrast, 8% of households were not at all satisfied with the quality. Figure 29: How satisfied are households with the quality or standard of the sub-project construction. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Very Satisfied Somewhat satisfied Not at all satisfied Source: Household-E2 When asked what factors improved project quality, nearly half of the respondents indicated community management, while 30% reported direct community purchasing led to the improved 57 quality of the sub-project (Table 28). Training and design also played a role, according to households. These results are fairly consistent throughout the provinces, except in Malaita where less than 1% of households indicated direct community purchasing helped improve the quality. These results are interesting as the majority of factors that led to the improvement in quality can be generalised into community participation, such as training and direct community purchasing and management. This could be a reflection of the effectiveness of community participation and ownership. Table 28: Factors that improved the quality of the sub-project. Direct Community community Province Design Training management purchasing Construction Choiseul 18.4% 14.7% 48.4% 33.6% 1.1.% Malaita 15.8% 25.9% 44.1% 7.9% 7.4% Temotu 23.0% 18.9% 45.1% 43.6% 3.6% Western 14.6% 21.8% 59.1% 35.4% 0 All 18.1% 20.2% 48.8% 30.4% 3.2% Source: Household-G13 A caveat to these results is that households also discussed other factors that could have improved the quality of the sub-project. This question appears to have confused some participants. There were numerous inappropriate responses because some respondents believed they were being asked what methods could be used to improve the sub-project. Thus, the interpretation of the analysis based on this question should be taken with caution. Figure 30: Quality of construction in comparison to other similar infrastructure built in the community or nearby 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Better Same Worse Source: Community- E42 58 Community leader discussion groups were asked to comment on the technical quality of construction compared to other similar infrastructure built in the community or nearby. The results are reported in Figure 30. The majority of communities (73%) reported that the quality of an RDP sub-project was better than similar infrastructure, while 20% of leader groups indicate an RDP project is of similar quality to other infrastructure projects. Figure 31: Factors that negatively impacted the quality of the sub project. 100% 80% 60% 40% 20% 0% Inadequate Lack of Poor SIC not Poor quality Distance from Poor funding technical skills contractors functioning materials suppliers Community cooperation Choiseul Malaita Temotu Western All Source: Household-G14 Every household was asked what factors negatively impacted the quality of the sub-projects (Figure 31). More than 50% of all households reported distance from supplier being a big impediment to sub-project quality while 45% of households indicated inadequate funding negatively impacted project quality. Other factors, such as the SIC not functioning (16.5%) and poor community cooperation (17.2%) were also important. In summary, these results suggest that the majority of households and community leaders were satisfied with the quality of the RDP sub-project. An important reason for this high satisfaction is in part high levels of community engagement. However, inadequate funding and distance from suppliers is still an impediment to project quality especially for those provinces further away from Honiara. 5.3.7) Sub-Project Maintenance Community leader groups were asked a series of questions about sub-project maintenance, beginning with ‘Has any maintenance already been carried out on this subproject?’ Figure 32 shows relatively low levels of maintenance were carried out. Overall, 20% of respondents reported maintenance had been carried out on sub-projects. These results are not unexpected as 59 a significant proportion of projects are incomplete. Looking only at those villages with complete projects close to 26% of community leader groups reported maintenance was carried out. Figure 32: The proportion of community leader groups that reported maintenance was carried out on the sub-project. 30% 25% 20% 15% 10% 5% 0% Choiseul Malaita Temotu Western All Source: Community-E50 The second question in this set asked community leader groups if maintenance was planned for the sub-project (Figure 33). Close to 70% of community leader groups plan to implement maintenance. The main difference between provinces was that less than half of the surveyed community groups in Malaita plan future maintenance compared to 90% in Western province. In comparison to other similar projects in the community (see Figure 34Error! Reference source not found.), maintenance plans are more common for RDP programs (67% vs 50%). Figure 33:The proportion of community leader groups that reported a plan for future maintenance of this sub-project. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Source: Community-E52 60 Figure 34: The proportion of community leader groups that reported there is an operations and maintenance plan for other similar infrastructure in the village. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Source: Community-E55 Under the auspice that maintenance will occur in the future, village community leader groups were subsequently asked how this maintenance will be funded (Figure 35). Methods of funding differ widely at the province level. Close to half of all community leader groups in Western province reported that maintenance will be funded through monthly subscription fees. In Temotu, funding is expected to come through fundraising and user fees, while in Malaita, maintenance will be funded through monthly fees and fundraising. Figure 35: Funds for maintenance 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All Monthly Fees User Fees Fundraising Source: Community-E53 The last question in this subsection asked community groups about who would be responsible for maintenance. The results are shown in Figure 36Error! Reference source not found.. Overall, the SIC was held responsible for this duty in 43% of communities, according to community 61 leader groups, while other committees are predominately responsible for this duty in Western province. Figure 36: Who will be responsible for carrying out maintenance? 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All SIC Other committee Chief/community leader Source: Community- E54. Note: 60% of community leader groups in Choiseul did not answer this question. The results from this section are interesting. Only a quarter of communities undertook maintenance for their sub-projects while more than half reported maintenance will be carried out in the future. It is also worrying that a large number of communities are either not planning to carrying out maintenance in the future or are unsure from where the funding will come. Further, they often don’t have experience implementing maintenance in other similar non-RDP projects. Poor implementation of maintenance may be, in part, due to a lack of skills and experience, especially by the SIC committee who are held responsible for implementing the maintenance in most villages. These results seems to be consistent with the baseline in which households were asked what needs to be done to keep buildings and facilities in this village in good condition and working properly?’. In most cases, households were aware maintenance was important yet lacked the skills and organisation to implement this maintenance. 62 5.4) Sub-project Participation 5.4.1) Participation Table 29: Intermediate Result Indicator Three PDO Level Results D=Dropped Unit of Cumulative Target Values Indicators C=Continue Measure Evaluation N=Revised Baseline R=Revised Yr 1 Yr 3 Yr 5 Yr 7 Core Intermediate Result R Percent Zero 47.9% 80 indicator Three: Percentage of households in target planning units participating in RDP Community Development Sub- project decision making In this section we analyse the level of participation in RDP projects across study communities. In particular we look at processes that work and what some of the avenues for future improvements. However, the starting point of our analysis is the medium term indicator that RDP has set for itself in the results framework: percentage of households in target planning units participating in RDP Community Development Sub-project decision making. Figure 37 represents the level of participation across all four study provinces. From the table we can see that overall participation rate was well below RDP targets. Only 47.9 households reported that they participated in the selection of the RDP sub-project. The highest participation rate was in Temotu (58.8 %) and lowest in Malaita (39.0 %). The main reason why people did not participate was because they were away (24.0 %), being busy (14.0 %) or lack of awareness (9.0 %). If we restrict our sample to just have people who were present during the sub-project selection process then participation rate increases to 71.9 % which is 8.1 percentage points lower than RDP targets. 63 Figure 37: Percentage of Households participating in RDP community development Sub-project decision making 70% 60% 50% 40% 30% 20% 10% 0% -10% Participated Did not Know Away Busy Ill Other Choiseul Malaita Temotu Western Total Source: Household survey F1 and F3 As a part of the survey, households were also asked to identify ways in which participation could have been improved. The three most important avenues that households identified were for the SIC/RDP/CHs to provide 2 weeks’ notice (37.5 %), liaise with the church to make announcements (42.7 %) and hold meetings in the evenings (36.1 %) (Figure 38). The suggestions made by the households are consistent with the main reasons for not participating in meetings. The best possible strategy would be to provide two weeks’ notice through the churches for meetings to be held in evenings. Figure 38: Strategies to encourage people to participate in selection of projects 80% 70% 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western Total 2 Weeks’ Notice Church Announcements Meeting in Evenings Source: Household survey F1 and F6 Majority of the people who participated in meetings actively took part in them. 54.6 % of the households reported that they both spoke and voted during project selection meetings. 21 % reported they voted while 12.4 % reported they spoke. 12.1 % of the participants did not actively 64 contribute in the meetings. Participants in Choiseul province were most active and in Western province were the least active (Figure 39). Figure 39: Type of participation in sub-project selection meetings 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western Total Participant Voted Participant Spoke Participant Did Both Participant Did neither Source: Household survey F1 and F2 Households were also asked to identify who had the most influence over selection of RDP sub- projects. Most households indicated that the entire community had the most influence over the choice of sub-projects (72.4 %). The second and the third most important influence came from village leadership (43.4 %) and religious leadership (22.0 %). Their influence was particularly high in Temotu province. The WDC and the MPA had very little influence in the selection of projects at the village level (Table 30). Based on these responses it is clear that indeed the selection of sub-project is based on community’s needs and desires. However, our results would have been stronger if we had over 50 % of the community participating in the selection process. Table 30: Important influences on the selection of sub-projects Entire Religious Province Community Chief leader WDC MPA Choiseul 84.2% 21.2% 4.3% 0.0% 1.6% Malaita 53.2% 33.7% 13.7% 16.1% 0.5% Temotu 72.4% 74.4% 45.7% 1.0% 2.5% Western 82.9% 42.9% 23.5% 1.2% 2.4% Total 72.4% 43.4% 22.0% 4.9% 1.7% Source: Household survey F4 65 In terms of participation in implementation of sub-projects most households helped out either by providing labour (83.8 %) and/or food (57.9 %). Another 10 % of the households reported that they provided money or material towards the sub-project implementation. Finally, 92.1 % of the households believed that community participation was essential in implementing the project and 94.6 % of them believed that community contributions should be mandatory by RDP as it used to be. In summary, the level of participation in the sub-project selection was well below RDP targets. However, it seems that the non-participation was not due to lack of interest amongst the members of the community. If meetings were better advertised and were held at better times (evenings) participation rates would have been substantially higher. The fact that majority of the community understand the need for community participation in the implementation part reinforces this observation. 5.5) Gender It has long been understood 15 that to ensure effective development outcomes, development projects must encourage the voice and participation of women. Women’s integration into development projects establishes an equal gender power relation and is an equality-based approach. The RDP program has set guidelines that encourage women to participate and contribute to the development projects that directly involve them. 5.5.1) Engagement of Women in RDP The community leader survey asked participants if they thought RDP processes enable women to influence decision-making more than other community projects. In nearly all cases, the community leader groups reported RDP processes enabled women’s involvement (Figure 40). This is consistent throughout all of the surveyed provinces. When asked how RDP processes enabled female engagement, close to 45% of all community groups listed women in the SIC as being a driving factor (Figure 41) while 30% of community groups reported that the RDP processes provided an opportunity for women to present their views and 51% reported RDP processes encouraged women to be more active. 15 Sen AK (1992). Missing Women. British Medical Journal;304: 586-7 66 Figure 40: The proportion of community leader groups that think RDP processes enable women to influence decision-making more than other community projects 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Source: Community-E14 Figure 41: RDP processes, women and their influence on decision making in comparison to other community projects. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All Women in SIC Opportunity to present voice Women more active Source: Community-E15. Note: This table specifically examines the question ‘ Why do RDP processes enable women to influence decision-making more than other community projects?’ An important RDP recommendation is that the SIC should be comprised of different interest groups within the community, including women. Figure 42 reports the answers to E16 from the community leader survey which asked participants if the SIC includes women. Overall, 90% of the surveyed SIC committees include women. At the province level, Choiseul had the lowest proportion of SICs with female members (80%) while Temotu had the highest (95%). 67 Figure 42: The proportion of Sub-project Implementation Committees which included women. 100% 95% 90% 85% 80% 75% 70% Choiseul Malaita Temotu Western All Source: Community-E16 The results from Figure 43 show that RDP SIC membership is often the first major community responsibility of the women involved. In Temotu, the province with the highest female SIC membership, in 85% of cases, it is these women’s first major community responsibility while in Choiseul, the province with the lowest female SIC membership, it is the first major community responsibility for 61% of female SIC members. Figure 43: The proportion of community leader groups which reported SIC’s include women and it was their first major community responsibility. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Source: Community-E17 These results show that the RDP processes have increased female community engagement through SIC committee membership. However, these findings demonstrate that women are active in the RDP process but not in other areas of the community. To examine if engagement increased female participation in other village activities, community leader groups were asked ‘If there was/is a women on the SIC, has her/their activity in the village changed since joining the SIC?’ (Figure 44). In 81% of villages, the community leader groups reported women on the SIC 68 participate more in village activities than before joining the SIC. At the province level, 44% of community leader groups reported the female SIC members, on average, participated to the same degree in community activities. Figure 44: Women in the SIC and their activity in the village since joining the SIC. 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All More Same Less Source: Community-E18. Note: This table examines the question- ‘If there was/is a women on the SIC, has her/their activity in the village changed since joining the SIC?’ A follow-up question ‘If there was/is a women on the SIC, has her/their activity outside of the village changed since joining the SIC?’ shows that on average, 80% of female SIC members participate more in activities outside their community (see Figure 45). However, the results for this question may have confused participants, and they may have been unable to differentiate this question from the previous one. Figure 45: Women in the SIC and their activity outside the village since joining the SIC. 80% 70% 60% 50% 40% 30% 20% 10% 0% Often Sometimes Rarely SIC women more active outside the village SIC women are not more active outside the village Source: Community-E19. Note: This table examines the question- ‘ If there was/is a women on the SIC, has her/their activity outside of the village changed since joining the SIC? ’ 69 Figure 46: The proportion of community leader groups that reported that women who participated in the SIC increased their status in the community. 100% 95% 90% 85% 80% 75% Choiseul Malaita Temotu Western All Source: Community-E20 The last question in this set examines how female SIC members’ status had changed in the community since participating in the SIC. The results are shown in 70 Figure 46. Close to all community leader groups reported being in the SIC committee increased the status of the female members. 5.5.2) Effects of Female Participation in the RDP Process. Section 5.5.1 established that women play an important role in SIC committees. It is often the first form of community engagement for women which leads to spill-over effects, increasing female community participation. This section will examine the effects of increased female participation in the RDP process16. Figure 47: Women in the SIC and reported SIC effectiveness 100% 80% 60% 40% 20% 0% Women SIC No Women in SIC Source: Community-E16 and E6 Figure 47 examines SIC composition and the proportion of community groups which consider SIC to be effective. When SICs included women, in 94% of cases, community groups reported the SIC to be effective. On the other hand, when the SICs excluded women, only 62% of community groups reported the SIC to be effective. Table 31: Women in the SIC and frequency of meetings Often Sometimes Rarely Women in SIC 38.3% 41.2% 20.6% No Women in SIC 25.0% 0.0% 75.0% Source:Community-E16 and E1 The SIC is often responsible for organising meetings and informing community groups of important activities and changes. When women were a part of SICs, close to 40% of community leader groups identified community meetings occurred often, while 41% report community 16 We report the results at the country level rather than province level due to the small number of observations. 71 meetings occurred sometimes. In contrast, when SICs did not include women, meetings rarely occurred 75% of the time (see Table 31). These results however, may be driven by a small subset of villages which could have performed poorly, irrespective of SIC composition. It is therefore important to examine this issue further. Table 32 shows that when female SIC members are reported to be more active outside the village community, meetings are held often in 44% of villages and sometimes in 75% of villages while in communities where SIC membership does not increase female engagement inside the community, meetings are held often in these villages 17% of the time and sometimes 75% of the time. Table 32: Active women in SIC and frequency of meetings Often Sometimes Rarely SIC women more active inside the village 44.0% 35.0% 21.0% SIC women are active inside the village to the same degree as before. 17.0% 75.0% 8.0% Source: Community-E19 and E1 Conflict within communities, such as disagreements or disputes before or during the construction, is a large impediment to RDP project completion. Interestingly, when female SIC members are more active outside the community, conflict occurs in 30% of villages compared to 41% in communities where female SIC members are less active (Table 33). The hypothesis behind these results is that SIC female members being more active inside the community leads to improved village participation and greater ownership. This results in less frequent community conflict, as members feel included in the process and design. However, due to the small sample and possibility that villages in which women are more active are inherintantly different to villages in which women are less active. Table 33: Active women in SIC and conflict in the implementation process Conflict SIC women more active inside the village 29% SIC women are active inside the village to the same degree as before. 41% Source: Community- E19 and E24 Another important issue is the involvement of women in the selection process. A key design of the RDP program is community-driven development. Community driven means that the communities will be responsible for the identification and the implementation of their own sub- projects, thereby encouraging community ownership. Table 34 uses the baseline question WA 5 from the female discussion group to examine the proportion of communities that reported 72 women were involved in the selection process 17 and if the project is now complete. Overall, when women were involved in the selection process, 68.5% of villages reported the project was complete while when women were not involved in the selection process, only 57% of the villages reported the project was complete. Table 34: Participation of women in the RDP selection process and completion of sub- projects Women involved in selection Project Complete Women involved in selection 68.42% Not involved 57.14% Source: Baseline: Women discussion group- WA 5. Wave 1: Household-E1 Table 35: Women involvement in the selection process and conflict in implementation RDP Women involved in Selection Process Conflict Women involved in selection 36.8% Not involved 57.1% Source: Baseline: Women discussion group- WA 5. Wave 1: Community- E24. Note: This table examines the question ‘If women are involved in the RDP selection process, is there less conflict in the implementation process? ’ Similar to the results shown in Table 35, when women are involved in the selection process, 37% of village leader groups reported there were disagreements or disputes before or during the construction. In contrast, when women were not involved in the selection process, 57% of community leader groups reported some conflict before or during the construction. Figure 48: Participation in sub-project selection meetings by gender 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% Men Women It is important to note that this question is different to that looked at above. It is directly taken from the baseline 17 women’s discussion group. 73 Source: Household Questionnaire F1 In terms of participation we can see that women are less likely to participate. On average 53.2 % of men participated in sub-project selection meetings compared to 39.7% for women (Figure 48). Even in terms of intensity of participation, women seemed to be more passive then men. On average 24 % of women reported to have neither spoken nor voted in selection meetings. The corresponding number for men was only 6 %. Men also generally spoke more than women in meetings. 78 % of men reported to have spoken in meetings compared to 47 % for women (Figure 49). Figure 49: Type of participation in sub-project meetings by gender 70% 60% 50% 40% 30% 20% 10% 0% Spoke and Voted Spoke Voted Neither Males Females Source: Household-F1 and F2 and A2 In summary, the RDP process encourages greater female community involvement, which occurs predominately through the SIC committee. Female members that join the SIC increasingly participate within and outside the community. This increased female participation directly correlates with greater RDP project outcomes. Specifically, when SIC committees comprise female representatives, there is likely to be less disagreement and disputes and more community meetings. However, it may also be the case that communities that have greater female participation have other characteristics that lead to greater project outcomes. This could be an area of further study to examine these villages and identify such characteristics. Nevertheless, these results showing female participation in the selection process and membership of the SIC are a strong signal that the RDP process creates wider community participation. This directly increases the probability that the RDP project will be successful. 74 5.6) Economic Activity and Services: 5.6.1) Primary School Attendance It has been widely recognised that education is a powerful instrument for reducing poverty. However, unfortunately access to education is often limited in the Solomon Islands. Not all the surveyed villages had primary schools and only a few had secondary schools. For those without access to schools, it can be difficult and expensive to ensure the education of children. A number of RDP projects focus on improving access to primary education through building libraries, staff houses and other improvements in primary school facilities. Table 36 and Table 37 examine if these projects had impacts on the proportion of eligible children attending primary school18. In general, 67% of eligible females were attending primary school while 62% of eligible males were attending. The proportion of females attending primary school is similar to the baseline while the proportion of males attending declined slightly. These percentages are similar compared with national estimates. Table 36: The proportion of eligible children per household attending primary school. Baseline Baseline Wave 1 Wave 1 Female Male Female Male All 65.2% 72.3% 67.1% 62.0% Household- C4 and A6. Table 37 specifically looks at primary school sub-project villages. In the baseline, these villages had much lower female participation compared to other villages. By the time of the follow-up survey, female participation had increased by 15% however, participation is still lower compared to non-sub project villages 19 . Male participation, on the other hand, decreased between the baseline and the follow up. However, the decline was smaller in RDP sub-project villages. Table 37: The proportion of eligible children per household attending primary school in RDP primary school sub-project villages. Wave Wave 1 Baseline Baseline Province Female Male Male Female All 63.33% 64.58% 77.7% 55.5% Source: Household- C4, A6 and E1. 18 Eligible children are defined as children of primary school age and or those that are attending primary school 19 This difference is not statistically significant at the 5 % level and should be interpreted with caution. 75 Sub-projects that provide sanitation, health services and other improved services may also have an impact on primary school attendance. For example, improved water supply and health services may reduce how often a child is sick, making it less likely they will drop out of school 20. We utilise a tobit21 regression to analyse the factors that affect the proportion of eligible children in each household who attend primary school. The results are reported in Table 38 for females and Table 39 for males. For females a reduction in travel time to primary school is statistically associated with improved primary school attendance. Furthermore, Malaita and Western provinces are more likely to have eligible females attending primary school. Other than this, improved access to services such as health clinics and women’s centres are not directly associated with female attendance at primary school. Table 38: Tobit regression- Dependent variable: The proportion of eligible female children per household who attend primary school. Variables Coefficients P-value Malaita 0.77 0.01 Temotu 0.31 0.24 Western 0.77 0.02 Agricultural Advice -0.01 0.95 School Access Improved -0.11 0.62 RDP Water Tank 0.09 0.76 RDP Health Clinic 0.44 0.16 RDP Community Hall 0.15 0.65 RDP Solar Station 0.19 0.59 RDP Primary School -0.10 0.84 RDP Radio 0.62 0.40 RDP Women Centre -0.49 0.32 Travel time to primary -0.00 0.10 Electricity improved 0.16 0.41 Sanitation Improved -0.16 0.61 No. Children -0.11 0.26 Roads Improved -0.50 0.03 Obs 315 Table 39: Tobit regression- Dependent variable: The proportion of eligible male children per household who attend primary school. 20 We test this only on the follow-up survey which appears to be a more reliable indicator of primary school attendance. This is because in the baseline, households were not asked how many children live in each household, thus it is difficult to gauge the number of children in each household who do not attend primary school. 21 Tobit models are statistical estimation models used when the dependent variable contains many censored observations (zero’s). Interpretation is identical to ordinary least squares 76 Variables Coefficients P-value Malaita 0.40 0.30 Temotu 0.20 0.59 Western -0.21 0.61 Agricultural Advice -0.02 0.94 School Access Improved 0.01 0.98 RDP Water Tank 0.37 0.34 RDP Health Clinic 0.87 0.07 RDP Community Hall 0.97 0.06 RDP Solar Station -0.14 0.77 RDP Primary School 0.35 0.59 RDP Radio 0.27 0.80 RDP Women Centre -0.34 0.71 Travel time to primary 0.00 0.84 Electricity improved -0.30 0.28 Sanitation Improved -0.11 0.78 No. Children -0.11 0.38 Roads Improved -0.06 0.84 Obs 334 RDP health clinic and community centre are two factors that appear statistically correlated with improved attendance at primary school for males. This is likely because communities with this infrastructure are likely to have other positive characteristics that encourage males to attend primary school. In summary, these results provide evidence that improved access to some infrastructure is not a leading determinant of female primary school participation. It is likely that a number of other factors that play a role in determining outcomes are not measured here. However, communities that improve health clinics and community halls are likely to have a greater proportion of males attending primary school, a signal that improved infrastructure is associated with primary school attendance. 77 6.1) Component 2 6.1.1) Agricultural Advice and Farming Practices Table 40: Results Indicator Three PDO Level Results D=Dropped Unit of Cumulative Target Values Indicators C=Continue Measure Evaluation N=Revised Baseline R=Revised Yr 1 Yr 3 Yr 5 Yr 7 Core Indicator Three: N Percent 2 11.2% 35 Percentage of Households who have changed agricultural practices including varieties, as a result of agricultural advice. Table 41: Proportion of Households that received advice and changed farming practices. Baseline Wave 1 Difference Province Choiseul 2.19% 5.26% 140.1% Malaita 1.6% 19.19% 916.6% Temotu 2.06% 12.63% 513.1% Western 2.8% 6.94% 147.8% All 2.18% 11.20% 413.7% Source: Household-H8 and Household-H5. To be considered receiving advice and changing practices H5==1 & H8==1 Agricultural innovation is an important tool to increase production and incomes at the village level and makes up a key component of the RDP project. The main driver of agricultural innovation under an RDP comes from agricultural officers who visit villages to disseminate training and advice. Indicator three examines this issue asking “What are the Percentage of Households who changed agricultural practices including varieties, as a result of agricultural advice”. Those households which change agricultural practices and receive d advice are considered to have changed practices as results of advice. While all other households which do not change practices and receive advice are considered to not have changed agricultural practices as a result of advice. In other words, households which change practices and do not receive advice or households which receive advice and do not change practices are not considered to have changed practices as a result of advice. 78 The baseline shows that, on average, 2% of households changed agricultural practices as a result of receiving advice. At the province level, 2% of households changed practices after receiving advice in Choiseul and Temotu while 2.8% changed practices in Western province. By the time of the follow-up survey, the proportion of households that received advice and changed practices improved markedly to 11.20%, an increase of 413% from the baseline. Large advances were seen in Choiseul (5.26%), Malaita (19.19%) and Temotu (12.63), increasing the proportion of households that received advice. A much smaller increase was seen in Western province where the number of households changing practices after receiving advice increased by 147% (Table 41) Figure 50: Households whose farming methods have changed over the last 4 years. 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All Baseline Wave 1 Source: Household-H6 To explain the above changes, it is important to examine the proportion of households that changed practices and then identify the factors driving this change. Figure 50 shows the proportion of households that have changed agricultural practices over the last 4 years. Farming methods have changed on average by 192% from the baseline. The largest changes were seen in both Choiseul (561%) and Malaita (180%). Strikingly, these results are much higher than those shown in Table 41. For example, in wave 1, 40% of households changed farming methods, yet only 11.20% changed farming methods due to the advice received. This, in part, is due to the proportion of households that actually received advice. Figure 51 shows that, on average, Malaita, Temotu and Western all saw increases in the proportion of households that received advice, however Choiseul saw a decrease in the proportion of households that received advice. Strikingly, the proportion of households that received advice is still much lower than those that 79 changed farming methods inferring many households change practices due to organisation other than RDP. Figure 51: The proportion of households that received or asked for advice from an agricultural officer or advisor in the past 12 months. 40% 35% 30% 25% 20% 15% 10% 5% 0% Choiseul Malaita Temotu Western All Baseline Wave 1 Source: Household-H8 Figure 50: Households whose farming methods have changed over the last 4 years. shows that households are changing farming practices but, this is only partly due to agricultural advice. However, when households actually received advice (see Table 43) are, on average, likely to change their practices 50% of the time. There are considerable differences at the province level, with Choiseul and Malaita increasing the proportion of households that changed practices when they received advice by 310% and 110% respectively, while much smaller differences are seen in Temotu and Western. Of those that received advice and changed practices, satisfaction in the advice has grown since the baseline survey. On average, 61% of households are always satisfied with agricultural advice, up from 47% in the baseline. Western province is the most satisfied with agricultural advice, with 76% of households saying they are satisfied in contrast to 45% of Temotu households which are always satisfied with farming advice. In comparison, households are less likely to be sometimes satisfied or never satisfied with agricultural advice, decreasing by 16% and 50% respectively from the baseline survey. 80 Table 42: The proportion of households that reported receiving satisfactory advice. Baseline Wave 1 Difference Province Always Choiseul 46.34% 64.00% 38.1% Malaita 53.30% 69.13% 29.7% Temotu 41.67% 45.47% 9.1% Western 48.20% 75.76% 57.2% All 47.58% 61.02% 28.2% Sometimes Choiseul 29.27% 16.00% -45.3% Malaita 10.00% 19.11% 91.1% Temotu 37.50% 40.90% 9.1% Western 37.90% 21.21% -44.0% All 28.23% 23.73% -15.9% Never Choiseul 21.95% 20.00% -8.9% Malaita 36.67% 11.76% -67.9% Temotu 20.83% 13.63% -34.6% Western 10.34% 3.03% -70.7% All 22.58% 11.30% -50.0% Source: Household-H10 Table 43: The proportion of households that reported receiving advice and changing their farming practices. (For those that received advice only) Baseline Wave 1 Difference Province Choiseul 9.76% 40.00% 309.8% Malaita 25.81% 54.29% 110.3% Temotu 47.83% 52.08% 8.9% Western 35.71% 36.36% 1.8% All 26.83% 48.30% 80.0% Source: Household- H8 and H5. This sample is restricted to only those that received advice. To be considered receiving advice and changing practices H5==1 & H8==1 Interestingly, this increase in satisfaction in Western and Temotu is correlated with an increase in farm output due to agricultural advice. The results are reported in Figure 52. Overall, 73% of households reported an increase in crop or farm output due to agricultural advice, this result has not changed since the baseline, even though a larger proportion of households are receiving agricultural advice. 81 Figure 52: The proportion of households who reported increased crop or livestock production. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Choiseul Malaita Temotu Western All Baseline Wave 1 Source: Household-H11. This sample is restricted to only those households that answered H8==1. In summary, the results for indicator 3 are modest. The proportion of households that received agricultural advice and changed their farming practices has increased from the baseline. However, there are two important reasons for the slow pick-up in agricultural methods: i) only 25% of households have received or asked for advice from agricultural officers. If households don’t receive advice, their opportunities to improve farming practices are limited; ii) on average, 50% of households take up the advice and change practices. This means that a large proportion of households never change their practices even after receiving advice. This directly contrasts with the high level of satisfaction and improvement in output due to the advice received. This alludes to the important need to improve take up and agricultural advice to a larger number of households. Evidence from past take up may help convince households of the advantages of changing their farming methods; iii) when households are asked if MAL officers provided consultation in a village, household answers varied in nearly all villages, with some households reporting ‘yes’ to this question while others reported ‘no’. This provides evidence that MAL officers are holding consultation sessions in villages, however, participation in these consultation sessions is low. This could be a reason why the take up of MAL advice is seen in only a quarter of households. 82 7.1) Intermediate Indicator 7.1.1) Agricultural Services Table 44: Intermediate Result Indicator One. PDO Level Results D=Dropped Unit of Cumulative Target Indicators C=Continue Measure Values Evaluation N=Revised Baseline R=Revised Yr Yr Yr 5 Yr Core 1 3 7 Intermediate Result indicator R Percent Zero 49% 20 50 One: Percentage of villages in participating provinces to whom agricultural services have been provided. Indicator one examines the percentage of villages in participating provinces to whom agricultural services have been provided. Province level results for intermediate indicator one are shown in Figure 53. Agricultural services are defined as all the services provided by MAL and other service providers including receiving agricultural advice from MAL officers and any training, demonstrations, workshops, trials or other new activities or developments in agriculture in this village within the past four years. Overall, 49% of villages received agricultural services from MAL. Agricultural service provision is lowest in Western province (38%) and highest in Malaita (59%). As mentioned previously, an issue with this measure is the percentage of villages receiving a service is defined from the household survey. In most cases some villagers reported receiving a service while others did not. To overcome this, for each village, we tallied the total number of responses which indicated that they received a MAL service. If this tallied number was more than 50 % of the total number of respondents in that village, then the village was classified as having received a MAL service. However, it could be argued that if at least one household receives an agricultural service in a village then that village has received a service. Using this metric, only four villages (2 in Malaita, 1 in Choiseul and 1 in Western) do not receive any MAL services. This represents over a 95% MAL service coverage, significantly larger than the 49% rate reported above. These results infer, that MAL service coverage is wide, nearly all surveyed villages receive at least some MAL service, however within each village only a small proportion are involved in agricultural services. 83 Figure 53: The proportion of households that received a MAL service. 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% Choiseul Malaita Temotu Western All Baseline Wave 1 Source: Wave 1: Household- H8 ,H14, H15. Baseline-C7, C14, C15. Note: Villages where more than half of all households received a service are classified as receiving a MAL service. It is also important to note that these results are almost identical to the results at the household level. All householders were asked if there had been any new developments such as training, demonstrations, workshops and trials in agriculture in their village in the past 12 months. 22 Similar to the baseline, there was considerable inconsistency in responses at the village level. In 85% of villages, results were contradictory with some households saying ‘yes’ to the survey, while others in the same village saying ‘no’. This suggests that many villagers may not be aware of the methods used by their neighbours and further, participation in consultations was limited. Table 45 shows that the proportion of households that reported new developments and activities in their village increased by 146%, on average, throughout all provinces. The results were largest in Choiseul and Malaita where new developments in agricultural increased by 150% and 350%, respectively. These results may be, in part, driven by an increasing spill-over of knowledge of agricultural activities in villages, however changes are also likely due to an increasing number of activities and developments. 22 Note, this is a shorter reference period than for household-level innovation, which referred to the past five years. 84 Table 45: The proportion of households that reported training, demonstrations, workshops, trials or other new activities or developments in agriculture in their village within the past four years Province Baseline Wave 1 Difference Choiseul 20.7% 51.7% 150.1% Malaita 12.5% 56.1% 349.0% Temotu 23.6% 38.4% 62.7% Western 24.8% 36.7% 48.0% All 18.7% 46.1% 146.3% Source: Household-H15 Those who said there had been new developments in agriculture were asked to indicate who had funded the developments (see Table 52). In over 85% of cases, new developments, such as training and workshops, were run by the Ministry of Agriculture and Livestock (MAL). This is a significant improvement from the baseline where, on average, 53% of households indicated new activities were carried out by the Ministry. Breaking down changes at the province level, Choiseul and Western had the largest increase in new activities implemented by MAL. The large increase in MAL activities coincides with a decrease in new developments implemented by NGO’s and other private donors. The proportion of households receiving training and other developments by NGOs declined by 65% while other private donors decreased agricultural activities by, on average, 30%. 7.2.2) Consultation All household respondents were asked if anyone from their household had participated in consultations with MAL. Figure 54 shows that overall 17.6% of respondents said ‘yes’ in the follow-up survey compared to 10.6% in the baseline. Households in Choiseul were most likely to have joined a consultation (19.5%) and those in Temotu least likely (11.85%). This pattern is interesting, especially in Malaita where MAL increased consultations by, on average, 462.67%, a significant increase and now it has the highest consultation rate of all the provinces. Such a large increase in a relatively short period should be looked at as a lesson for other provinces. 85 Figure 54: The proportion of households that reported taking part in community consultation with MAL. 30% 25% 20% 15% 10% 5% 0% Choiseul Malaita Temotu Western All Baseline Wave 1 Source: Household- H12 Of those that took part in community consultation, on average, 78% of baseline households described their needs and priorities to MAL, compared to 64% in the follow-up survey. This is a decrease of 18.5% and shows that communities are less likely to make suggestions to the Ministry. This decrease can be explained, at least partially, by the large increase in the total number of households that took part in consultation with MAL in the follow-up survey compared to the baseline. Figure 55: The proportion of households that received community consultation and described their needs and priorities to the MAL in the last four years. 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Proportion of HH that received community Proportion of HH that described their needs and consultation and described their needs and priorities priorities to MAL and reported that MAL responded to the MAL in the last 4 years. to these suggestions. Baseline Wave 1 Source: Household-H13 & H14 86 Error! Reference source not found. shows that when communities describe their needs and priorities to MAL, these suggestions are likely to be acted upon 57% of the time. This is an increase of 58% from 36% in the baseline. These results should be taken with caution due to the low response rates, as the results may be biased. Also household responses differ within villages, suggesting responses were not always disseminated throughout a village and thus participation in consultations was somewhat limited. 7.1.2) Changes in farming practices As seen in Figure 50, on average, farming methods have changed in 40% of households, and these changes are, in part, due to the RDP in addition to other NGOs and governments. Changes in farming methods are wide-ranging, as shown in Table 46. Fertilizer and herbicide usage decreased, on average, by 32% and 100%, respectively in the surveyed households while pesticides saw a small increase in take up. Table 46: How farming methods have changed over the last four years. Baseline Wave 1 Difference Fertilizer Choiseul 0.00% 1.58% 0% Malaita 2.03% 3.50% 72.7% Temotu 6.17% 0.51% -91.8% Western 2.65% 1.10% -58.5% All 2.50% 1.69% -32.4% Pesticide Choiseul 0.00% 2.63% 0% Malaita 2.03% 3.50% 72.7% Temotu 2.47% 1.01% -59.1% Western 1.99% 4.40% 121.2% All 1.63% 2.86% 75.3% Herbicide Choiseul 0.00% 0.00% 0% Malaita 0.34% 0.00% -100.0% Temotu 0.62% 0.00% -100.0% Western 1.32% 0.00% -100.0% All 87 Drainage Choiseul 0.53% 0.53% -0.5% Malaita 1.01% 0.00% -100.0% Temotu 1.23% 0.00% -100.0% Western 0.66% 2.20% 231.9% All 0.88% 0.65% -26.3% Irrigation Choiseul 0.00% 0.00% 0% Malaita 1.01% 0.00% -100.0% Temotu 1.23% 0.00% -100.0% Western 0.00% 1.10% 0% All 0.63% 0.26% -58.8% Cultivation Choiseul 2.65% 33.68% 1173.3% Malaita 9.12% 2.00% -78.1% Temotu 2.45% 38.38% 1464.1% Western 3.97% 9.34% 135.1% All 5.25% 20.91% 298.3% Improved Seed Choiseul 3.70% 3.7% -0.5% Malaita 1.01% 8.0% 689.3% Temotu 5.52% 18.7% 238.4% Western 6.62% 3.8% -41.9% All 4.75% 8.70% 83.2% Other (including- Soil improvement, Crop Varieties and Live Stock Changes) Choiseul 3.70% 21.6% 482.6% Malaita 3.72% 35.0% 841.8% Temotu 10.43% 46.0% 340.7% Western 9.93% 19.2% 93.6% All 7.38% 30.8% 317.2% Source: Household-H6 Changes in irrigation and drainage declined by 26% and 58%, respectively to zero percentage for all surveyed households expect those in Western province. This decline in take up, albeit from a low base, may in part be due to the large initial costs of the change, or because irrigation is less important in wet environments. The largest changes in agricultural methods were seen in cultivation in which close to 21% of households changed methods in this area, an increase of 300%. Similar changes in soil improvements, crop varieties and livestock were seen in all provinces (on average, 30% of households changed their practices in these areas). 88 7.2) MAL services 7.2.1) Type of Advice Sought All householders were asked if they had received advice from MAL in the past 12 months. Those that responded ‘yes’ were then asked to indicate the type of advice they received. Figure 56 : Type of advice received by MAL 60% 50% 40% 30% 20% 10% 0% Cocoa Crop Training Planting Soil Livestock Baseline Wave 1 Source:Household-H9 Figure 56 reports these results. Responses were diverse, with categories often not being mutually exclusive. The crop category consists of households who received advice specifically about crop varieties or crop management. Training refers to any unspecified training undertaken by MAL. Responses in this category include “how to use the right tool for farming” to ‘workshop on how to grow cabbages’. The largest category, planting, refers to how to plant agricultural crops including pruning and planting seeds. The second last category, soil, refers directly to advice about soil type including managing and improving soil fertility while livestock includes advice on livestock, such as pigs and poultry. The results show that advice on cocoa is relatively stable and makes up around 20% of advice disseminated. Advice on training, crops and soil all increased markedly, ranging from 200% (crops) to 94% (soil). As mentioned, the largest group is planting which makes up over 50% of advice received, with the majority of advice being ‘how to plant ‘ and ‘how to prune’. Livestock advice declined from 17% to 10% of the advice received. This suggests that villagers are mostly concerned with agricultural activities rather than livestock or perhaps MAL programs are more likely to focus on agricultural activities. Finally, as shown and mentioned in Error! Reference 89 source not found., in those households that received advice, farming output increased on average by 73%. 7.3.1) Food Security An important issue directly related to food production is food security. All households were asked whether their household food supply was adequate. This variable was converted into a ‘yes’/’no’ variable, with ‘sometimes’ counted as ‘no’. On average, in the baseline, 40% of households reported they are either sometimes unable to grow, catch, and/or buy enough food for everybody. This is in contrast to 15% of households who reported the same in the follow-up survey (Figure 57). Figure 57 : The proportion of households that are able to grow, catch and/or buy enough food for everyone at baseline 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Yes Sometimes No Figure 58 : The proportion of households that are able to grow, catch and/or buy enough food for everyone at wave1 100% 80% 60% 40% 20% 0% Choiseul Malaita Temotu Western All Yes Sometimes No 90 Figure 58 examines the relationship between households’ adequate food supply and advice received from MAL. Column 2 of Table 47 reports the proportion of households that received advice from MAL and have an adequate food supply, and the third column reports the proportion of households that did not receive advice and have an adequate food supply. The p- value reports the probability that the means are statistically different. The results show that in the follow up, on average, 90% of households who received advice are able to grow, catch, and/or buy enough food for everybody. In comparison, 85% of households that did not receive advice have an adequate food supply. The means are statistically different, inferring households that received advice are more likely to provide an adequate food supply for everybody in the household. Table 47: The proportion of households (follow-up survey) that are able to grow, catch and/or buy enough food for everyone and either received or didn’t receive MAL advice. Province Receive Advice No Advice P-value Obs- Receive Obs-No Advice Choiseul 100.0% 97.5% 25 164 Malaita 90.0% 84.3% 70 127 Temotu 85.1% 67.8% 47 146 Western 91.4% 89.6% 35 125 All 90.3% 85.5% 0.04 177 562 Source: Household- H8 and H1. These results are in direct contrast to Table 48 which examines the same questions for the baseline sample. In the follow-up survey, on average, 60% of households that received advice are able to grow, catch or buy enough food for everyone, while those that did not receive advice have an adequate food supply 58% of the time. These means are not statistically different to each other, implying that at the baseline, households who received MAL advice are not more likely to have an adequate food supply. Table 48: The proportion of households (baseline survey) that are able to grow, catch and/or buy enough food for everyone and either received or didn’t receive MAL advice. Province Receive Advice No Advice P-value Obs- Receive Obs-No Advice Choiseul 46.3% 58.3% 41 144 Malaita 61.2% 54.9% 31 264 Temotu 75% 73.3% 24 124 Western 66.6% 50.8% 27 120 All 60.1% 58.4% 0.36 123 652 Source: Baseline Household- C1 and C7 The difference between the baseline and follow-up survey are important and need to be examined further to identify the mechanisms driving an adequate food supply. To analyse this 91 further, we run a standard Probit model which allows us to control for households and other factors that also determine food supply. The results are reported in Table 49. The first column in the table reports the variables used in this model. The second provides the marginal effects, in other words, the impact of a one-unit change in a variable on the probability of having an adequate food supply. The last column provides the p-values. A p-value less than 0.05 indicates the variable has a statistically significant impact on having an adequate food supply. The Probit model shows that a 1-unit increase in the probability of receiving advice leads to a 6% increase in the probability of having an adequate food supply. This is statistically significant at the 5% level. Further, households that sold food crops are more likely to have an adequate food supply. Specifically, an increase in the probability of selling food crops leads to a 10% greater probability of having an adequate food supply. Differences also exist between provinces. Malaita and Temotu are statistically more likely to be food insecure compared to Choiseul 23. Other variables, such as MAL consultation and household factors such as the number of people in the household, are statistically insignificant determinates of having enough food for everyone in the household. Table 49: Probit model: dependent variable - households are able to grow, catch and/or buy enough food for everyone. Variables Marginal Effects P-value Household received advice 0.06 0.04 Gender of respondent 0.04 0.25 No. of males in Household -0.00 0.85 No. of females in Household 0.01 0.62 Sell Livestock 0.01 0.80 Sell Crops 0.11 0.00 Males in Primary 0.01 0.40 Females in Primary -0.00 0.96 Head of House Female -0.01 0.90 MAL consultation -0.05 0.14 Malaita -0.11 0.01 Temotu -0.24 0.00 Western -0.06 0.18 Source: Household- H1 (dependent variable). 23 Choiseul is the benchmark case that all other provinces are compared against. For example a negative marginal effect for Temotu indicates that Temotu is less food secure compared to Choiseul. 92 In summary, these results are similar to those reported in the baseline report (see page 25). The main difference is the impact of MAL advice. In the baseline, MAL advice was not correlated with food security, contradictory to the results above. A number of factors could be driving this: i) households that seek advice may be inherently different to those that don’t, for example, they may be more motivated to improve farming practices, thus leading to more food for the household; ii) increasing food security could also lead to an increase in the take up of advice by MAL. The rationale underlying this is that as households have greater food security, they are more likely to take risks and change their practices leading to a greater increase in the advice taken up. Evidence to prove both these hypothesis is difficult as motivations and risk preferences are often unobserved; iii) MAL advice may be leading to increased farming output, improving food supply for the family. To examine if this is the case, we ran a separate Probit (results not reported) including the variable “Did the advice help your household produce more or better crops and/or livestock?” This variable is statistically significant at the 1% level, indicating households that receive advice and report an increase in crop output are 15% less likely to be food secure. This provides evidence for argument iii). 7.3.2) Food Security and Markets. Households in the Solomon Islands are predominately subsistence farmers. However, over the last decade, this has begun to change. Of those households surveyed in the baseline, 57% reported selling food compared to 71% of households during the follow-up survey. This is a large increase that could have an important impact on the livelihoods of rural households. To examine this relationship, we compare those households that received advice and sell crops. The results are reported below. Table 50: The proportion of households that sell crops and either received advice or didn’t receive advice. All Receive Advice No advice p-value Obs- receive Obs-not receive Baseline 0.65 0.57 0.06 124 654 Wave 1 0.77 0.71 0.05 176 558 Table 50 shows that households that received advice are more likely to sell crops. In the baseline, 64% of those that received advice sell crops while only 56% of those that didn’t receive advice 93 sell crops. A similar difference is seen in the follow up. In both cases, the difference in the mean between those that received advice is statistically significant at the 10% level. Table 51: Probit: Dependent variable - households that reported selling crops. Variables Marginal Effects P-value Household received advice 0.18 0.73 Gender of respondent -0.08 0.04 No. of males in Household -0.01 0.55 No. of females in Household 0.02 0.17 Sell Livestock 0.17 0.00 Households has adequate food 0.19 0.00 Males in Primary 0.02 0.48 Females in Primary -0.00 0.10 Head of House Female -0.04 0.71 MAL consultation 0.13 0.03 Malaita -0.17 0.02 Temotu -0.17 0.02 Western -0.06 0.34 Source: Household-H3 (dependent variable) Table 51 shows a Probit Regression model of the correlates of households selling crops. A household’s ability to sell crops is derived from household question H3 ‘Does this household grow crops? (including food and cash crops such as cocoa and copra). The results show that the gender of the survey respondent is negatively correlated with the probability of a household selling crops. If a household sells livestock or if they have an adequate food supply, they are, on average, more likely to sell crops. In addition, if a household attends a MAL consultation, they are, on average, 12% more likely to sell crops. At the province level, Malaita and Temotu are statistically less likely to sell crops compared to Choiseul. In summary, these results are very similar to the food security analysis above and thus, the conclusions are similar. Households that receive some MAL consultation are more likely to sell crops. However, the results should be interpreted with caution as we are unable to ascertain a direct causal relationship. 94 Table 52: Organisations which introduced the reported training, demonstrations, workshops, trials or other new activities or developments in their village Province Ministry of Agriculture Private NGO's Instrument Baseline Wave 1 Difference Baseline Wave 1 Difference Baseline Wave 1 Difference Obs Obs Baseline Choiseul 54.3% 93.4% 72.1% 5.7% 8.8% 53.9% 22.9% 2.2% -90.4% 91 35 Malaita 58.3% 73.2% 25.5% 22.2% 20.6% -7.3% 30.6% 12.4% -59.5% 97 36 Temotu 83.9% 91.9% 9.6% 3.2% 0.0% -100.0% 12.9% 10.8% -16.2% 74 31 Western 18.2% 86.2% 374.2% 21.2% 5.2% -75.6% 45.5% 15.8% -65.3% 58 33 All 53.3% 86.7% 62.6% 13.3% 9.4% -29.5% 28.1% 9.8% -65.0% 316 135 Source: Household-H17.Note: This sample is restricted to only those household that report they received training, demonstrations, workshops, trials or other new activities or developments in the village (H15==1). 95 Section 8: Recommendations 8.1.1) Component 1: Infrastructure Services Recommandations. 1) A clear definition of what constitutes clean water needs to be identified. This is important to understand whether the RDP is funding projects that provide only marginal improvements in water quality. Additionally, if multiple options do exist in terms of project choice, communities will be able to choose projects that provide the highest benefits to costs ratio. 2) One of the shortcomings of the RDPs results framework is that it identifies percentage increases/decreases between baseline and follow up surveys only at the national level. We propose two additions: (i) We recommend that the RDP introduce targets based on overall coverage rate in addition to percentage improvements between baseline and follow up surveys. ii) We recommend the RDP creates a results framework at the province level. This will ensure that the improvements in basic infrastructure are equitable across all the provinces. The framework should be modelled on the current framework used at the country level. 3) RDP projects are currently classified as community projects. However, in terms of usage these projects vary greatly. Some such as electricity are often user pay while others such as water tanks are free to use. This distinction between private and public usage projects needs to be made and based on these distinctions best practice strategies for improving outcomes in both private projects and public projects will have to be formulated. Experiences from similar developing countries can be consulted in developing such strategies. 4) Allocation of water, especially in the dry season, is vital in ensuring the health benefits from most water projects. This can be done in at least two ways: (i) Training: It important that RDP train communities on best practices in management and allocation of water resources. Given the potential of conflict it would be best to involve village chiefs in this process. 96 (ii) User fee: Community members could be charged a small user fee for every unit of water they consume. This will reduce wastage and ensure water is used only for purposes that generate benefits. At the same time it will also generate funds for maintaining quality of the water sources. 8.2) Component 1 RDP Processes- Recommendations: 1) We find that improved participation from members of the community leads to greater project completion rates as well as higher quality construction. The RDP should devise strategies to improve participation. This is especially pertinent going forward, when most projects will be moving into a maintenance phase. 2) Communities lack skills, experience and knowledge in project maintenance and management. It is important that those who are given responsibility (O&M committees) are trained and understand the importance of timely maintenance to ensure project sustainability. 3) Communication and dissemination of information is identified as one of the major sources of community dissatisfaction with the RDP processes. For example, communities often do not understand the reasons behind RDPs procurement process or how it is implemented leading to reports of dissatisfaction. To overcome this issue CH’s and TL’s must ensure information is dispersed throughout the community. 4) Connected to the previous point, villages that are further away (in terms of distance) from PSUs often have more issues with CHs’ and consequently lack information. RDP must devise a strategy to improve monitoring of these CHs’. A possible suggestion would be for communities to maintain log books to record CH attendance in meetings. This will allow the community and RDP to monitor CHs’ effort levels. In addition, random visits by TL’s or other monitors without prior notice will improve monitoring. 5) We find that when women are involved in RDP processes including the SIC, projects implementation is more likely to be successful. In addition women’s status in the community improves due to participation in these processes. As a result, we would recommend RDP to make it mandatory for women to be included in future project implementation and maintenance committees. 97 8.3) Component 2 Recommendations: 1. Close to 25% of households received or asked for advice from agricultural officers. If households don’t receive advice their opportunities to improve farming practices are limited. MAL officers must therefore increase participation at MAL events including training and consultation. 2. Half of all households that receive agricultural advice do not use the information to change agricultural practices. However, those that do change practices are more than satisfied with the results. The low take up of advice may in part be because farmers are risk adverse and changing practices is costly. We recommend undertaking research into costs involved in changing farming practices to identify why take up of advice is limited. 3. Following point ii) to encourage take up of practices MAL officers should provide information to households on the benefit of changing practices. This could be in the form of posters. Alternatively, MAL should encourage households that have changed practices to explain the method of change in addition to the benefits. This could take place in meetings and consultation organised by MAL 4. To improve outreach MAL should provide passive ways through which farmers can reach out to them. For example, a help exchange could be set up where farmers from any village are able to telephone MAL officers seeking advice. Alternatively particular farmers from villages could be appointed by MAL to liaise with MAL officers. The important issue here is to provide a quick and easy mechanism through which advice can be sought. 5. Finally, as a general recommendation it is important to stress the importance of improving agricultural productivity in improving livelihoods. At this point in time most villagers do not see the link between higher yield and better livelihood outcomes. This is because most villages do not have access to markets nor do they have access to proper storage facilities. Unless an appropriate supply chain network is developed, whereby farmers can sell their products in big markets the true effect of providing agricultural advice will not be realised. 98 8.4) Future Methodology: This section will provide an outline on what survey questions and methodology were successful and subsequent recommendations. Response rates in the community survey were relatively high, on average 85% of community groups responded to questions. Response rates were either low or inconsistent in a number of cases: 1) When community groups were asked to provide percentages or estimate numbers (such as A11) 2) Open ended questions about the RDP process. For example, Community groups were confused when asked questions about what could have been done better, as they often used this to discuss negative or positive aspects of the program rather than provide suggestions. 3) Response rates often declined significantly towards the end of the survey, implying the survey may have been too long. Response rates in the household survey were slightly lower than that of the community. On average between 70-80% of survey questions were answered. Response rates were either low or inconsistent in a number of similar questions: 1) Questions based on village level impacts. These question types are difficult to interpret in household surveys due to inconsistencies, as some households believe a service was provided to a community while others did not. 2) Questions where households reported time to access infrastructure such as question D2 varied significantly, in many cases reported time ranged from10 minutes to over 60 minutes to access the same resource from the same village. 3) Questions about RDP process and implementation such as G8-G15 were often broad categories. This lowered the power of these questions, as specific responses were not able to be measured. Recommendations: 1) One of the main issues we faced with regards to assessing improvements in satisfactory infrastructure is that there was no set definition of satisfactory access. These terms need to be defined for the analysis to be objective. For example to ascertain access to electricity households can be asked whether they have enough electricity to power 2 light bulbs during the night. If RDP is indeed hesitant of using a top down approach in defining these terms, it can always conduct research within the community and get a sense of what the communities think satisfactory access should be. 2) Another significant issue with the household survey is the lack of questions to ascertain household's wealth and income. Questions should be asked such as the number of mobile phones, type of roofing and method of cooking (i.e. gas or wood) to ascertain 99 measures of wealth. This is important as it will allow for future researchers to see the impact of RDP services on wealth which is a significant factor that determines poverty. 3) One of the problems with households surveys is that respondents have tendencies to misreport. Self-image issues, strategic deception and even plain mistakes do occur when answering questions. To ensure that these kinds of errors are minimised provisions for spot checking should be incorporated within the survey framework. For example if a household reports having access to electricity it would be worthwhile to actually check whether they do. 4) Open ended questions often led to confusion and thus lead to discussions that are often off topic. These questions need to be more focused. For example individual questions about SIC procurement process should be asked rather than asking respondents to describe the benefits and drawbacks of SIC procurement process. 5) It is general practice in health questionnaires to have multiple shorter recall periods. It is also customary to ask further questions to check for robustness. For example, to ascertain whether any children health, it is customary to ask the primary caregiver whether the child (to be called out by his/her name) was sick (by mentioning symptom) in the last two days? Followed up by robustness questions on prevalence of the sickness yesterday and today? 6) The health section in this survey was very poorly designed. There was very few questions on hygiene practices which is a important factor that affects overall health of households. In future evaluations which most likely will try to quantify health improvements, having a well thought out section on getting information hygiene behaviour would be vital. 7) Quantify improvements in access to infrastructure. Provide participants with the information that very satisfied with water access is the same as rating access between 8- 10 out of 10. While satisfied is equivalent to 5-7 out of 10, and not satisfied is equivalent to rating access less than 5/10. 8) Questions regarding water sources should identify the quality of these water sources. In other words, respondents should also be asked if water sources are ever polluted, or if the water source is covered (such as a bore hole). This will provide better identification of quality. 9) School attendance questions must ask household the number of children who do not attend Kindy but who are eligible to attend. Without this question, it is difficult to infer the proportion of students that attend Kindy. 100 10) Community leader surveys provide more consistent and higher order responses to those questions about community infrastructure. Community leader surveys should be used when community access and infrastructure is of interest rather than only household surveys. 101 Appendix 1: Session Logs All evaluation teams took notes on key RDP observations and discussions with community members. Summaries of key issues are provided below24. Choiseul In a number of communities, respondents were unsure how projects are prioritized by the communities and then how they are ranked and chosen by the WDC. Enumerators found that communities that were unhappy with the SIC were unwilling to publicly criticize the committee. Therefore, questions based on the SIC may be underestimates of how satisfied community members are with the committee. According to enumerators, the distribution of information is a key issue. In some villages, community members were unsure of RDP’s role. Malaita In general, RDP issues often arose because of poor communication between the CH and the community. For example, Sulugwalu and Foufu communities did not generally understand the RDP implementation process. Sulugwalu, in particular, complained that the selection of SIC was done by the CH without the knowledge of many villagers. -In Kikiri, a rain water tank was built. Community members were generally happy with the project but were unable to manage the water supply and complain about the people who waste the limited amount of water in the water tank. Temotu: In general, the procurement process was communities main issue. Materials often took months to arrive in Lata and villagers were often not updated on the current status. For example, the Malapu community were unhappy with the procurement process, believing they should be able to collect their own materials. Another important observation was the differences between the Reef Islands and Santa Cruz. Distance and basic infrastructure on the Reef significantly increased the time it took to complete 24 Western session logs are inaccessible. 102 a project. This is likely one of the main reasons why close to 70% of projects on the Reef are reported as incomplete. In general, community members in the Reef Islands were more unhappy with CHs compared to those communities in Santa Cruz. This, in part, may be due to monitoring. CHs in the Reef Islands were monitored less often compared to Santa Cruz, increasing free riding by the CHs . 103 104 Appendix 2 Annex 1. Distribution of Village Size Village Size Percent < 25 24.4% 26-50 21.8% 51-75 23.1% 76-100 12.8% >100 17.9% Annex 2: Percentage of villages with satisfactory access to primary education (using 70% cut-off) Percentage of Villages with Satisfactory Access Province Baseline Wave 1 Difference Choiseul 71.0% 75.0% 5.6% Malaita 67.0% 70.0% 4.5% Temotu 83.0% 60.0% -27.7% Western 88.0% 79.0% -10.2% Full 78.0% 71.0% -9.0% Full w/o Temotu 68.0% 75.0% 10.3% Annex 3: Percentage of villages with satisfactory access to health services (using 70% cut-off) Percentage of Villages with Satisfactory Access Province Baseline Wave 1 Difference Choiseul 43.0% 50.0% 16.3% Malaita 40.0% 45.0% 12.5% Temotu 48.0% 25.0% -47.9% Western 47.0% 58.0% 23.4% Full 45.0% 44.0% -2.2% Full w/o Temotu 39.0% 51.0% 30.8% 105 Annex 4 Percentage of villages with satisfactory access to roads (using 70% cut-off) Percentage of Villages with Satisfactory Access Province Baseline Wave 1 Difference Choiseul 0.0% 90.0% Malaita 67.0% 70.0% 4.5% Temotu 30.0% 45.0% 50.0% Western 24.0% 58.0% 141.7% Full 28.0% 65.0% 132.1% Full w/o Temotu 24.0% 72.0% 200.0% Annex 5 Percentage of villages with satisfactory access to markets (using 70% cut-off) Percentage of Villages with Satisfactory Access Province Baseline Wave 1 Difference Choiseul 10% 70% 600% Malaita 53% 55% 4% Temotu 35% 0% -100% Western 29% 42% 45% Full 30% 41% 37% Full w/o Temotu 25% 55% 120% Annex 6: Percentage of villages with satisfactory access to water (using 70% cut-off) Percentage of Villages with Satisfactory Access Province Baseline Wave 1 Difference Choiseul 29.6% 90.0% 204% Malaita 26.7% 50.0% 87% Temotu 13.4% 5.0% -63% Western 23.5% 68.4% 192% Full 22.4% 53.16% 137% Full w/o Temotu 26.4% 69.49% 163% Annex 7: Percentage of villages with satisfactory access to sanitation facilities (using 70% cut-off) Percentage of Villages with Satisfactory Access Province Baseline Wave 1 Difference Choiseul 0.0% 10.0% Malaita 0.0% 0.0% Temotu 0.0% 0.0% Western 6.0% 16.0% 166.7% Full 1.3% 6.0% 361.5% Full w/o Temotu 1.7% 8.0% 370.6% 106 Annex 8: Percentage of villages with satisfactory access to electricity (using 70% cut-off) Percentage of Villages with Satisfactory Access Province Baseline Wave 1 Difference Choiseul 0.0% 90.0% Malaita 7.0% 5.0% -28.6% Temotu 0.0% 5.0% Western 12.0% 84.0% 600.0% Full 3.9% 45.0% 1053.8% Full w/o Temotu 5.1% 59.0% 1056.9% Annex 9:Proportion of Household that Report the CH is Important in the RDP Process. Province Yes Obs Choiseul 85% 20 Malaita 95% 20 Temotu 85% 20 Western 93.7% 16 All 89.4% 76 Source: Community-E21 107 Appendix 3 Section A: General Province Question Response Mean All A10 village governed by elected leader 44% All A10 village governed by traditional chief 95% All A10 village governed by church leader 14% All A10 village governed by village committee 10% Choiseul A10 village governed by elected leader 5% Choiseul A10 village governed by traditional chief 100% Choiseul A10 village governed by church leader 0% Choiseul A10 village governed by village committee 0% Malaita A10 village governed by elected leader 5% Malaita A10 village governed by traditional chief 90% Malaita A10 village governed by church leader 10% Malaita A10 village governed by village committee 10% Temotu A10 village governed by elected leader 80% Temotu A10 village governed by traditional chief 85% Temotu A10 village governed by church leader 20% Temotu A10 village governed by village committee 15% Western A10 village governed by elected leader 17% Western A10 village governed by traditional chief 67% Western A10 village governed by church leader 33% Western A10 village governed by village committee 11% All A11 main denominations in village: anglican church 58% All A11 main denominations in village: catholic church 35% All A11 main denominations in village: charismatic church 27% All A11 main denominations in village: methodist church 3% All A11 main denominations in village: seventh day adventist 38% All A11 main denominations in village: ssec 25% All A11 main denominations in village: united church 33% Choiseul A11 main denominations in village: anglican church 0% Choiseul A11 main denominations in village: catholic church 30% Choiseul A11 main denominations in village: charismatic church 0% Choiseul A11 main denominations in village: methodist church 5% Choiseul A11 main denominations in village: seventh day adventist 25% Choiseul A11 main denominations in village: ssec 5% Choiseul A11 main denominations in village: united church 65% Malaita A11 main denominations in village: anglican church 40% Malaita A11 main denominations in village: catholic church 45% Malaita A11 main denominations in village: charismatic church 5% Malaita A11 main denominations in village: methodist church 0% Malaita A11 main denominations in village: seventh day adventist 40% 108 Malaita A11 main denominations in village: ssec 0% Malaita A11 main denominations in village: united church 0% Temotu A11 main denominations in village: anglican church 100% Temotu A11 main denominations in village: catholic church 10% Temotu A11 main denominations in village: charismatic church 50% Temotu A11 main denominations in village: methodist church 0% Temotu A11 main denominations in village: seventh day adventist 25% Temotu A11 main denominations in village: ssec 0% Temotu A11 main denominations in village: united church 0% Western A11 main denominations in village: anglican church NA Western A11 main denominations in village: catholic church NA Western A11 main denominations in village: charismatic church NA Western A11 main denominations in village: methodist church NA Western A11 main denominations in village: seventh day adventist NA Western A11 main denominations in village: ssec NA Western A11 main denominations in village: united church NA All A12 no. people living in Honiara 58.57 Choiseul A12 no. people living in Honiara 11.20 Malaita A12 no. people living in Honiara 49.85 Temotu A12 no. people living in Honiara 114.65 Western A12 no. people living in Honiara 176.71 All A13 shock over past year: drought 27% All A13 shock over past year: earthquake 22% All A13 shock over past year: flood 27% All A13 shock over past year: typhoon 0% All A13 shock over past year: landslide 14% All A13 shock over past year: tsunami 21% All A13 shock over past year: highrain 54% All A13 shock over past year: volcano 0% Choiseul A13 shock over past year: drought 40% Choiseul A13 shock over past year: earthquake 30% Choiseul A13 shock over past year: flood 40% Choiseul A13 shock over past year: typhoon 0% Choiseul A13 shock over past year: landslide 10% Choiseul A13 shock over past year: tsunami 20% Choiseul A13 shock over past year: highrain 15% Choiseul A13 shock over past year: volcano 0% Malaita A13 shock over past year: drought 100% Malaita A13 shock over past year: earthquake 0% Malaita A13 shock over past year: flood 100% Malaita A13 shock over past year: typhoon 0% Malaita A13 shock over past year: landslide 100% Malaita A13 shock over past year: tsunami 100% Malaita A13 shock over past year: highrain 100% Malaita A13 shock over past year: volcano 0% Temotu A13 shock over past year: drought 0% 109 Temotu A13 shock over past year: earthquake 8% Temotu A13 shock over past year: flood 0% Temotu A13 shock over past year: typhoon 0% Temotu A13 shock over past year: landslide 0% Temotu A13 shock over past year: tsunami 8% Temotu A13 shock over past year: highrain 92% Temotu A13 shock over past year: volcano 0% Western A13 shock over past year: drought 16% Western A13 shock over past year: earthquake 16% Western A13 shock over past year: flood 0% Western A13 shock over past year: typhoon 0% Western A13 shock over past year: landslide 0% Western A13 shock over past year: tsunami 0% Western A13 shock over past year: highrain 88% Western A13 shock over past year: volcano 0% All A2 no. HHS in village 87.92 Choiseul A2 no. HHS in village 96.25 Malaita A2 no. HHS in village 103.85 Temotu A2 no. HHS in village 63.65 Western A2 no. HHS in village 79.27 All A3 no. people in village 490.87 Choiseul A3 no. people in village 590.90 Malaita A3 no. people in village 485.85 Temotu A3 no. people in village 395.85 Western A3 no. people in village 622.22 All A4 Number of tribal groups living in village 4.82 Choiseul A4 Number of tribal groups living in village 4.05 Malaita A4 Number of tribal groups living in village 4.05 Temotu A4 Number of tribal groups living in village 6.35 Western A4 Number of tribal groups living in village 4.58 All A8 % of land that is customary 75-100% Choiseul A8 % of land that is customary 75-100% Malaita A8 % of land that is customary 75-100% Temotu A8 % of land that is customary 76-100% Western A8 % of land that is customary 75-100% Section B: Economic Activities Province Question Response Mean All B2 people depending on subsistence economy 0.75 Choiseul B2 people depending on subsistence economy 75-100% Malaita B2 people depending on subsistence economy 1.19 Temotu B2 people depending on subsistence economy 75.00 Western B2 people depending on subsistence economy 0.75 All B4 no. female-owned businesses 0.88 110 Choiseul B4 no. female-owned businesses 0.35 Malaita B4 no. female-owned businesses 1.80 Temotu B4 no. female-owned businesses 0.41 Western B4 no. female-owned businesses 0.50 All B5 no. jointly-owned businesses 5.19 Choiseul B5 no. jointly-owned businesses 7.85 Malaita B5 no. jointly-owned businesses 4.05 Temotu B5 no. jointly-owned businesses 3.41 Western B5 no. jointly-owned businesses 6.80 Section C: Participation Ques Province Response Mean tion All C1 village benefitted from project: rwss 27% All C1 village benefitted from project: provincial govt 13% All C1 village benefitted from project: ramp, mpp1, mpp2 8% All C1 village benefitted from project: constituency fund 55% All C1 village benefitted from project: ngo 40% All C1 village benefitted from project: other donor 59% All C1 village benefitted from project: national govt 30% Choiseul C1 village benefitted from project: rwss 25% Choiseul C1 village benefitted from project: provincial govt 30% Choiseul C1 village benefitted from project: ramp, mpp1, mpp2 0% Choiseul C1 village benefitted from project: constituency fund 60% Choiseul C1 village benefitted from project: ngo 45% Choiseul C1 village benefitted from project: other donor 35% Choiseul C1 village benefitted from project: national govt 30% Malaita C1 village benefitted from project: rwss 30% Malaita C1 village benefitted from project: provincial govt 5% Malaita C1 village benefitted from project: ramp, mpp1, mpp2 25% Malaita C1 village benefitted from project: constituency fund 30% Malaita C1 village benefitted from project: ngo 35% Malaita C1 village benefitted from project: other donor 63% Malaita C1 village benefitted from project: national govt 15% Temotu C1 village benefitted from project: rwss 26% Temotu C1 village benefitted from project: provincial govt 5% Temotu C1 village benefitted from project: ramp, mpp1, mpp2 0% Temotu C1 village benefitted from project: constituency fund 75% Temotu C1 village benefitted from project: ngo 40% Temotu C1 village benefitted from project: other donor 80% Temotu C1 village benefitted from project: national govt 45% Western C1 village benefitted from project: rwss 22% Western C1 village benefitted from project: provincial govt 22% Western C1 village benefitted from project: ramp, mpp1, mpp2 33% 111 Western C1 village benefitted from project: constituency fund 38% Western C1 village benefitted from project: ngo 38% Western C1 village benefitted from project: other donor 38% Western C1 village benefitted from project: national govt 67% All C10 mp distributed foor to village 66% Choiseul C10 mp distributed foor to village 75% Malaita C10 mp distributed foor to village 68% Temotu C10 mp distributed foor to village 55% Western C10 mp distributed foor to village 1% no. mpa visits over previous year (365=lives in All C11 community) 35.07 Choiseul C11 no. mpa visits over previous year (365=lives 104.15 Malaita C11 no. mpa visits over previous year (365=lives 0.05 Temotu C11 no. mpa visits over previous year (365=lives 1.00 Western C11 no. mpa visits over previous year (365=lives 4.50 All C12 mpa led development project 23% Choiseul C12 mpa led development project 40% Malaita C12 mpa led development project 5% Temotu C12 mpa led development project 25% Western C12 mpa led development project 22% All C13 mpa distributed foor to village 23% Choiseul C13 mpa distributed foor to village 45% Malaita C13 mpa distributed foor to village 5% Temotu C13 mpa distributed foor to village 20% Western C13 mpa distributed foor to village 22% All C15 no. times church distributed food past yr 2.05 More than once a Choiseul C15 no. times church distributed food past yr year More than once a Malaita C15 no. times church distributed food past yr year Temotu C15 no. times church distributed food past yr Less than once a year More than once a Western C15 no. times church distributed food past yr year All C2 someone in village involved w/ logging 23% Choiseul C2 someone in village involved w/ logging 55% Malaita C2 someone in village involved w/ logging 15% Temotu C2 someone in village involved w/ logging 0% Western C2 someone in village involved w/ logging 72% All C4 someone in village involved w/ mining 13% Choiseul C4 someone in village involved w/ mining 30% Malaita C4 someone in village involved w/ mining 0% Temotu C4 someone in village involved w/ mining 10% Western C4 someone in village involved w/ mining 28% All C7 no. mp visits over previous year 0.62 Choiseul C7 no. mp visits over previous year 0.65 Malaita C7 no. mp visits over previous year 0.85 Temotu C7 no. mp visits over previous year 0.35 112 Western C7 no. mp visits over previous year 0.10 All C8 mp has family in village 48% Choiseul C8 mp has family in village 40% Malaita C8 mp has family in village 55% Temotu C8 mp has family in village 50% Western C8 mp has family in village 67% All C9 majority of people voted for mp 56% Choiseul C9 majority of people voted for mp 61% Malaita C9 majority of people voted for mp 53% Temotu C9 majority of people voted for mp 55% Western C9 majority of people voted for mp 67% Section D: Access to infrastructure Province Question Response Mean All D1 access to school improved 72% Choiseul D1 access to school improved 80% Malaita D1 access to school improved 80% Temotu D1 access to school improved 55% Western D1 access to school improved 44% All D10 access to water improved 49% Choiseul D10 access to water improved 80% Malaita D10 access to water improved 53% Temotu D10 access to water improved 15% Western D10 access to water improved 56% All D13 access to sanitation improved 3% Choiseul D13 access to sanitation improved 5% Malaita D13 access to sanitation improved 5% Temotu D13 access to sanitation improved 0% Western D13 access to sanitation improved 6% All D16 access to electricity improved 68% Choiseul D16 access to electricity improved 95% Malaita D16 access to electricity improved 85% Temotu D16 access to electricity improved 25% Western D16 access to electricity improved 78% All D19 community buildings improved 28% Choiseul D19 community buildings improved 26% Malaita D19 community buildings improved 32% Temotu D19 community buildings improved 25% Western D19 community buildings improved 22% All D4 access to health care improved 2.08 Choiseul D4 access to health care improved 5.35 Malaita D4 access to health care improved 0.55 Temotu D4 access to health care improved 0.35 Western D4 access to health care improved 43.75 113 All D7 roads have improved 29% Choiseul D7 roads have improved 15% Malaita D7 roads have improved 40% Temotu D7 roads have improved 32% Western D7 roads have improved 20% Section E: Organisation of RDP Sub-project Province Question Response Mean All E1 frequency of rdp meetings Sometimes Choiseul E1 frequency of rdp meetings Sometimes to Rarely Malaita E1 frequency of rdp meetings Sometimes Temotu E1 frequency of rdp meetings Sometimes Western E1 frequency of rdp meetings Sometimes All E12 community supported SIC 95% Choiseul E12 community supported SIC 100% Malaita E12 community supported SIC 100% Temotu E12 community supported SIC 85% Western E12 community supported SIC 94% All E21 CH important in subproject implementation 88% Choiseul E21 CH important in subproject implementation 85% Malaita E21 CH important in subproject implementation 95% Temotu E21 CH important in subproject implementation 85% Western E21 CH important in subproject implementation 94% All E24 there were disagreements before or during construction 36% Choiseul E24 there were disagreements before or during construction 30% Malaita E24 there were disagreements before or during construction 35% Temotu E24 there were disagreements before or during construction 44% Western E24 there were disagreements before or during construction 19% All E3 attended rdp meeting: leaders 44% All E3 attended rdp meeting: SIC 61% All E3 attended rdp meeting: men 15% All E3 attended rdp meeting: women 12% All E3 attended rdp meeting: youths 7% All E3 attended rdp meeting: everybody 56% Choiseul E3 attended rdp meeting: leaders 40% Choiseul E3 attended rdp meeting: SIC 65% Choiseul E3 attended rdp meeting: men 0% Choiseul E3 attended rdp meeting: women 0% Choiseul E3 attended rdp meeting: youths 0% Choiseul E3 attended rdp meeting: everybody 15% Malaita E3 attended rdp meeting: leaders 100% Malaita E3 attended rdp meeting: SIC 100% Malaita E3 attended rdp meeting: men 100% Malaita E3 attended rdp meeting: women 100% 114 Malaita E3 attended rdp meeting: youths 100% Malaita E3 attended rdp meeting: everybody 100% Temotu E3 attended rdp meeting: leaders 40% Temotu E3 attended rdp meeting: SIC 40% Temotu E3 attended rdp meeting: men 25% Temotu E3 attended rdp meeting: women 20% Temotu E3 attended rdp meeting: youths 10% Temotu E3 attended rdp meeting: everybody 65% Western E3 attended rdp meeting: leaders 31% Western E3 attended rdp meeting: SIC 41% Western E3 attended rdp meeting: men 35% Western E3 attended rdp meeting: women 41% Western E3 attended rdp meeting: youths 24% Western E3 attended rdp meeting: everybody 59% All E30 travel time to provincial center 7.15 Choiseul E30 travel time to provincial center 14.90 Malaita E30 travel time to provincial center 4.20 Temotu E30 travel time to provincial center 2.34 Western E30 travel time to provincial center 12.63 hrs All E31 travel time to honiara 69.43 Choiseul E31 travel time to honiara 123.00 Malaita E31 travel time to honiara 8.50 Temotu E31 travel time to honiara 76.80 Western E31 travel time to honiara 28.27 hrs All E34 project material available locally 38% Choiseul E34 project material available locally 85% Malaita E34 project material available locally 10% Temotu E34 project material available locally 20% Western E34 project material available locally 81% All E38 used contractor for subproject 88% Choiseul E38 used contractor for subproject 75% Malaita E38 used contractor for subproject 90% Temotu E38 used contractor for subproject 100% Western E38 used contractor for subproject 25% All E39 satisfied with contractor for subproject Satisfied Choiseul E39 satisfied with contractor for subproject Satisfied Malaita E39 satisfied with contractor for subproject Satisfied Temotu E39 satisfied with contractor for subproject Satisfied Western E39 satisfied with contractor for subproject Satisfied All E40 land for subproject available 97% Choiseul E40 land for subproject available 100% Malaita E40 land for subproject available 100% Temotu E40 land for subproject available 90% Western E40 land for subproject available 100% All E44 rdp project used by: men 5% All E44 rdp project used by: women 3% 115 All E44 rdp project used by: children 7% All E44 rdp project used by: all 96% Choiseul E44 rdp project used by: men 5% Choiseul E44 rdp project used by: women 0% Choiseul E44 rdp project used by: children 0% Choiseul E44 rdp project used by: all 95% Malaita E44 rdp project used by: men 100% Malaita E44 rdp project used by: women 0% Malaita E44 rdp project used by: children 100% Malaita E44 rdp project used by: all 100% Temotu E44 rdp project used by: men 0% Temotu E44 rdp project used by: women 5% Temotu E44 rdp project used by: children 5% Temotu E44 rdp project used by: all 95% Western E44 rdp project used by: men 7% Western E44 rdp project used by: women 7% Western E44 rdp project used by: children 13% Western E44 rdp project used by: all 93% All E45 village members have to pay to use 46% Choiseul E45 village members have to pay to use 30% Malaita E45 village members have to pay to use 37% Temotu E45 village members have to pay to use 70% Western E45 village members have to pay to use 20% All E46 people from outside village have to pay 57% Choiseul E46 people from outside village have to pay 60% Malaita E46 people from outside village have to pay 40% Temotu E46 people from outside village have to pay 70% Western E46 people from outside village have to pay 53% All E5 rdp info made public 47% Choiseul E5 rdp info made public 20% Malaita E5 rdp info made public 35% Temotu E5 rdp info made public 85% Western E5 rdp info made public 41% All E6 SIC effective way to coordinate rdp 92% Choiseul E6 SIC effective way to coordinate rdp 90% Malaita E6 SIC effective way to coordinate rdp 90% Temotu E6 SIC effective way to coordinate rdp 95% Western E6 SIC effective way to coordinate rdp 88% Annex 10: Household Survey Results. Section A: General Province Question Response Mean 116 Total A2 respondent's gender 1.42 Choiseul A2 respondent's gender 1.34 Malaita A2 respondent's gender 1.45 Temotu A2 respondent's gender 1.42 Western A2 respondent's gender 1.46 Total A4 respondent's age in years 43.70 Choiseul A4 respondent's age in years 45.43 Malaita A4 respondent's age in years 39.40 Temotu A4 respondent's age in years 51.20 Western A4 respondent's age in years 45.94 Total A5 no. of male hh members 2.77 Total A5 no. of female hh members 2.76 Choiseul A5 no. of male hh members 2.85 Choiseul A5 no. of female hh members 2.80 Malaita A5 no. of male hh members 2.76 Malaita A5 no. of female hh members 2.83 Temotu A5 no. of male hh members 2.73 Temotu A5 no. of female hh members 2.74 Western A5 no. of male hh members 2.77 Western A5 no. of female hh members 2.68 Total A6 no. of boys aged 0-4 in hh 0.32 Total A6 no. of girls aged 0-4 in hh 0.25 Total A6 no. of boys aged 4-6 in hh 0.25 Total A6 no. of girls aged 4-6 in hh 0.20 Total A6 no. of boys aged 7-12 in hh 0.51 Total A6 no. of girls aged 7-12 in hh 0.48 Total A6 no. of boys aged 13-20 in hh 0.29 Total A6 no. of girls aged 13-20 in hh 0.26 Choiseul A6 no. of boys aged 0-4 in hh 0.42 Choiseul A6 no. of girls aged 0-4 in hh 0.29 Choiseul A6 no. of boys aged 41429 in hh 0.27 Choiseul A6 no. of girls aged 41429 in hh 0.21 Choiseul A6 no. of boys aged 41615 in hh 0.47 Choiseul A6 no. of girls aged 41615 in hh 0.42 Choiseul A6 no. of boys aged 13-20 in hh 0.36 Choiseul A6 no. of girls aged 13-20 in hh 0.38 Malaita A6 no. of boys aged 0-4 in hh 0.31 Malaita A6 no. of girls aged 0-4 in hh 0.22 Malaita A6 no. of boys aged 41429 in hh 0.28 Malaita A6 no. of girls aged 41429 in hh 0.14 Malaita A6 no. of boys aged 41615 in hh 0.55 Malaita A6 no. of girls aged 41615 in hh 0.51 Malaita A6 no. of boys aged 13-20 in hh 0.30 Malaita A6 no. of girls aged 13-20 in hh 0.20 Temotu A6 no. of boys aged 0-4 in hh 0.27 Temotu A6 no. of girls aged 0-4 in hh 0.27 117 Temotu A6 no. of boys aged 41429 in hh 0.25 Temotu A6 no. of girls aged 41429 in hh 0.28 Temotu A6 no. of boys aged 41615 in hh 0.54 Temotu A6 no. of girls aged 41615 in hh 0.54 Temotu A6 no. of boys aged 13-20 in hh 0.30 Temotu A6 no. of girls aged 13-20 in hh 0.22 Western A6 no. of boys aged 0-4 in hh 0.27 Western A6 no. of girls aged 0-4 in hh 0.22 Western A6 no. of boys aged 41429 in hh 0.19 Western A6 no. of girls aged 41429 in hh 0.19 Western A6 no. of boys aged 41615 in hh 0.46 Western A6 no. of girls aged 41615 in hh 0.44 Western A6 no. of boys aged 13-20 in hh 0.20 Western A6 no. of girls aged 13-20 in hh 0.23 Section B: Access to Water, Electricity and Sanitation. Province Question Response Mean Total B10 source of lightning: solar 85% Total B10 source of lightning: generator 5% Total B10 source of lightning: torch/battery 25% Total B10 source of lightning: kerosene lamp 23% Total B10 source of lightning: no electricity 1% Total B10 source of lightning: hydropower 0% Choiseul B10 source of lightning: solar 83% Choiseul B10 source of lightning: generator 3% Choiseul B10 source of lightning: torch/battery 9% Choiseul B10 source of lightning: kerosene lamp 14% Choiseul B10 source of lightning: no electricity 1% Choiseul B10 source of lightning: hydropower 0% Malaita B10 source of lightning: solar 86% Malaita B10 source of lightning: generator 4% Malaita B10 source of lightning: torch/battery 10% Malaita B10 source of lightning: kerosene lamp 8% Malaita B10 source of lightning: no electricity 1% Malaita B10 source of lightning: hydropower 1% Temotu B10 source of lightning: solar 81% Temotu B10 source of lightning: generator 1% Temotu B10 source of lightning: torch/battery 34% Temotu B10 source of lightning: kerosene lamp 13% Temotu B10 source of lightning: no electricity 0% Temotu B10 source of lightning: hydropower 0% Western B10 source of lightning: solar 78% Western B10 source of lightning: generator 9% Western B10 source of lightning: torch/battery 22% 118 Western B10 source of lightning: kerosene lamp 37% Western B10 source of lightning: no electricity 0% Western B10 source of lightning: hydropower 0% Total B11 source of electricity: solar 76% Total B11 source of electricity: generator 13% Total B11 source of electricity: no electricity 28% Total B11 source of electricity: hydropower 0% Choiseul B11 source of electricity: solar 83% Choiseul B11 source of electricity: generator 5% Choiseul B11 source of electricity: no electricity 15% Choiseul B11 source of electricity: hydropower 0% Malaita B11 source of electricity: solar 78% Malaita B11 source of electricity: generator 15% Malaita B11 source of electricity: no electricity 8% Malaita B11 source of electricity: hydropower 0% Temotu B11 source of electricity: solar 52% Temotu B11 source of electricity: generator 2% Temotu B11 source of electricity: no electricity 45% Temotu B11 source of electricity: hydropower 0% Western B11 source of electricity: solar 74% Western B11 source of electricity: generator 18% Western B11 source of electricity: no electricity 15% Western B11 source of electricity: hydropower 0% Total B14 washes hands: after toilet 1.50 Choiseul B14 washes hands: after toilet 1.67 Malaita B14 washes hands: after toilet 1.29 Temotu B14 washes hands: after toilet 1.70 Western B14 washes hands: after toilet 1.34 Total B15 washes hands: before food 1.56 Choiseul B15 washes hands: before food 1.64 Malaita B15 washes hands: before food 1.30 Temotu B15 washes hands: before food 1.73 Western B15 washes hands: before food 1.58 Total B16 washes hands: before looking after kids 1.63 Choiseul B16 washes hands: before looking after kids 1.68 Malaita B16 washes hands: before looking after kids 1.26 Temotu B16 washes hands: before looking after kids 1.89 Western B16 washes hands: before looking after kids 1.73 Total B7 no. times hh fetches water daily 2.35 Choiseul B7 no. times hh fetches water daily 2.59 Malaita B7 no. times hh fetches water daily 2.52 Temotu B7 no. times hh fetches water daily 1.90 Western B7 no. times hh fetches water daily 2.52 119 Section C: Children Province Question Response Mean Total C1 children in the hh had diarrhea over past week 7% Choiseul C1 children in the hh had diarrhea over past week 3% Malaita C1 children in the hh had diarrhea over past week 10% Temotu C1 children in the hh had diarrhea over past week 5% Western C1 children in the hh had diarrhea over past week 9% Total C2 no. of boys attending kindy 0.30 Total C2 no. of girls attending kindy 0.25 Choiseul C2 no. of boys attending kindy 0.31 Choiseul C2 no. of girls attending kindy 0.27 Malaita C2 no. of boys attending kindy 0.28 Malaita C2 no. of girls attending kindy 0.17 Temotu C2 no. of boys attending kindy 0.36 Temotu C2 no. of girls attending kindy 0.35 Western C2 no. of boys attending kindy 0.24 Western C2 no. of girls attending kindy 0.20 Total C3 days kids in the hh attended kindy 2.12 Choiseul C3 days kids in the hh attended kindy 1.76 Malaita C3 days kids in the hh attended kindy 3.00 Temotu C3 days kids in the hh attended kindy 1.74 Western C3 days kids in the hh attended kindy 1.94 Total C4 no. of boys attending primary sch 0.61 Total C4 no. of girls attending primary sch 0.58 Choiseul C4 no. of boys attending primary sch 0.63 Choiseul C4 no. of girls attending primary sch 0.57 Malaita C4 no. of boys attending primary sch 0.62 Malaita C4 no. of girls attending primary sch 0.57 Temotu C4 no. of boys attending primary sch 0.65 Temotu C4 no. of girls attending primary sch 0.59 Western C4 no. of boys attending primary sch 0.53 Western C4 no. of girls attending primary sch 0.61 Total C5 attend days kids in the hh attended primary sch 1.32 Choiseul C5 attend days kids in the hh attended primary sch 1.17 Malaita C5 attend days kids in the hh attended primary sch 1.05 Temotu C5 attend days kids in the hh attended primary sch 1.39 Western C5 attend days kids in the hh attended primary sch 1.56 Total C6 no. of boys attending secondary sch 0.23 Total C6 no. of girls attending secondary sch 0.19 Choiseul C6 no. of boys attending secondary sch 0.13 Choiseul C6 no. of girls attending secondary sch 0.21 Malaita C6 no. of boys attending secondary sch 0.26 Malaita C6 no. of girls attending secondary sch 0.17 Temotu C6 no. of boys attending secondary sch 0.35 Temotu C6 no. of girls attending secondary sch 0.23 120 Western C6 no. of boys attending secondary sch 0.19 Western C6 no. of girls attending secondary sch 0.15 Total C7 attend days kids in the hh attended secondary sch 1.21 Choiseul C7 attend days kids in the hh attended secondary sch 1.10 Malaita C7 attend days kids in the hh attended secondary sch 1.30 Temotu C7 attend days kids in the hh attended secondary sch 1.13 Western C7 attend days kids in the hh attended secondary sch 1.41 Section E: RDP Sub-project Province Question Response Mean Total E5 kept rdp project kept in good conditions 0.96 Choiseul E5 kept rdp project kept in good conditions 0.99 Malaita E5 kept rdp project kept in good conditions 0.96 Temotu E5 kept rdp project kept in good conditions 0.93 Western E5 kept rdp project kept in good conditions 0.95 Total E6 why rdp project not kept well: no money 5% Total E6 why rdp project not kept well: bad weather 0% Total E6 why rdp project not kept well: no coordination 7% Total E6 why rdp project not kept well: no skills 0% Total E6 why rdp project not kept well: no time 0% Total E6 why rdp project not kept well: no resources 2% Total E6 why rdp project not kept well: other 3% Choiseul E6 why rdp project not kept well: no money 2% Choiseul E6 why rdp project not kept well: bad weather 0% Choiseul E6 why rdp project not kept well: no coordination 5% Choiseul E6 why rdp project not kept well: no skills 1% Choiseul E6 why rdp project not kept well: no time 0% Choiseul E6 why rdp project not kept well: no resources 0% Choiseul E6 why rdp project not kept well: other 1% Malaita E6 why rdp project not kept well: no money 3% Malaita E6 why rdp project not kept well: bad weather 0% Malaita E6 why rdp project not kept well: no coordination 1% Malaita E6 why rdp project not kept well: no skills 0% Malaita E6 why rdp project not kept well: no time 0% Malaita E6 why rdp project not kept well: no resources 1% Malaita E6 why rdp project not kept well: other 1% Temotu E6 why rdp project not kept well: no money 13% Temotu E6 why rdp project not kept well: bad weather 0% Temotu E6 why rdp project not kept well: no coordination 50% Temotu E6 why rdp project not kept well: no skills 0% Temotu E6 why rdp project not kept well: no time 0% Temotu E6 why rdp project not kept well: no resources 25% Temotu E6 why rdp project not kept well: other 38% Temotu E6 why rdp project not kept well: no resources 25% 121 Western E6 why rdp project not kept well: no money 25% Western E6 why rdp project not kept well: bad weather 0% Western E6 why rdp project not kept well: no coordination 25% Western E6 why rdp project not kept well: no skills 0% Western E6 why rdp project not kept well: no time 0% Western E6 why rdp project not kept well: no resources 25% Western E6 why rdp project not kept well: other 0% Western E6 why rdp project not kept well: no resources 25% 122