Final Report - Baseline Survey

Baseline Survey Results for the Provincial Development Strategy Odtar Meanchey Province

April 2005 CARE/MALTESER/ZOA Written by Matthew Foster ([email protected])

2005 Baseline Survey, Otdar Meanchey Province Table of Contents Page A.

Introduction and methodology

1-5

B.

Results

1 2 3 4 5 6 7 8 9 10

Household data Community development Landmines/uxos Land and natural resources Agriculture and food security Health Education Women Human rights Co-ordination

6-7 8-17 18 19-22 23-32 33-48 49-50 51 52-53 54

C.

Recommendations

55-58

Annexes: A.

ToR

C:\Documents and Settings\abc\My Documents\care

B.

Questionnaire

C:\Documents and Settings\abc\My Documents\care

C.

Focus group discussion guide questions

C:\Documents and Settings\abc\My Documents\care

D.

List of villages sampled in household survey and FGDs

C:\Documents and Settings\abc\My Documents\care

E.

Comparison of logframes

C:\Documents and Settings\abc\My Documents\care

F.

Presentation of results and interpretation

C:\Documents and Settings\abc\My Documents\care

A. Introduction and methodology A.1 Introduction In 2003, Care, Malteser, and ZOA decided to carry out a joint situation analysis and planning process to help define the focus for their programmes and enhance co-ordination of their work in Odtar Meanchey. Key stages have included the Need Assessment (April 2003) a province wide assessment, and in November –December 2003 a multi-stakeholder planning process which helped to define a Provincial Development Strategy (PDS). The PDS is a multi-sectoral strategy document which identifies objectives in 10 components (community development, landmines, land and natural resources, agriculture, support to businesses, health and sanitation, education, women, human rights and governance, and coordination) and specific strategies to achieve these over the next five years. The three organisations will use the PDS to define programme activities and already joint funding proposals have been submitted to raise the resources required to implement the strategies, and it is hoped that additional NGOs working in the province will follow the PDS framework.

A.2 Objectives of the survey The need has been identified to set a baseline within the province for the 10 components to enable achievements towards the objectives in the PDS to be measured. A mid-term evaluation in January 2007 and an end evaluation in January 2009 have been scheduled which will repeat the baseline survey to assess change related to the objectives over time. The objectives of this consultancy as stated in the terms of reference (ToR) were (see annex A for full ToR): • • • • •

to review all relevant project documents including the needs assessment, strategy document, project plan and relevant proposals as well as essential documents of the participating organisations and give them suggestion for the set up of the questionnaires appropriate for the baseline survey to give advice on setting up a database and ensuring accurate data entry to conduct refresher training to project staff in survey and interview techniques which will help prepare them for conducting the survey to develop additional interview questions and conduct additional interviews with relevant stakeholders, including government officials from village, commune, and district as well as provincial level, other NGO staff, village representatives maybe using a RRA method to analyse the data and present the results in a comprehensive report

This resulted in three stages for the baseline survey; the household survey, focus group discussions (FGD), and interviews with provincial government officials. It was determined that the household survey would generate the majority of the information required for the baseline, and consequently the planning, implementation, and analysis of this stage was the major task for the three participating organisations and the consultant. Focus group discussions were used as a way to gather sensitive information, difficult to gather reliably from the household survey, and also to gather data to compliment the household survey results; providing participant’s thoughts on the situation in their villages on issues relevant to the PDS and ideas on explanatory factors. The interviews were added to set a form of baseline for component 10 of the PDS, co-ordination. The methodology for each of these stages is explained in detail below.

1

A.3 Household survey a) Development of the questionnaire The chosen household survey instrument was a structured questionnaire to gather quantitative data for statistical analysis. The development of the questionnaire was carried out following a review of the objectives and indicators within the PDS and the subsequent funding proposal submitted to EC (‘Community empowerment and strengthening of government capacity for the development of Oddar Meanchey province 2005-2007, March 2004). There was some uncertainly surrounding the compatibility of the objectives and indicators in the two documents and attempts were made to refine these (including removal of output indicators). The table in Annex E attempts to harmonise the objectives and indicators detailed in the logframes for both documents, and provides a useful way to cross reference between the logframes and the results section in this report. The table also clarifies that the household survey could not generate all the information required for the Monitoring and Evaluation (MandE) purposes of the PDS and the results of the survey should be seen within the context of a broader MandE framework (where additional information will need to be gathered this is indicated in the third column with ‘MandE framework’). It is considered that the baseline survey is best suited to generating information at the impact level i.e. changes in people life’s/socio-economic situations, the ultimate intention of the PDS, as well as for some of the anticipated outcomes i.e. changes in the behaviour and practise of people and/or service providers. Additional information will need to be gathered by staff or from existing sources to compliment the information generated by the baseline survey and to form a complete Monitoring and Evaluation framework (MandE framework) for the PDS. The process for developing the questionnaire involved the ZOA advisor and the consultant working together to draft questions to meet the expressed needs of the three organisations and the requirement of the logframes. An existing questionnaire used for a previous ZOA community development project was used as a starting point as it contained many relevant questions used previously to good effect, and an initial draft was distributed to the participation organisations for comment. Revision were made and a first draft of the questionnaire was translated into Khmer and back translated into English to check the interpretation of the questions, although this did not eliminate all mistakes (see lesson’s learnt section). The final draft of the questionnaire was field tested twice as part of the training (see below) and revised following suggestions from the interviewers related to interviewees understanding of the questions and appropriateness of terms. In addition, during the survey alterations were made to questions were it was found that mistakes persisted in the questionnaire. A final version of the questionnaire is included in annex B. The table in annex E helps to demonstrate that the process of drafting workable questions had the result of adapting the indicators from those originally contained in the logframe (the actual questions used in the questionnaire should be considered as the indicators). The consultant consider this to be the correct approach; allowing the MandE framework to be driven by the information that can realistically be gathered as opposed to sticking rigidly to the indicators in the original logframes. The consultant recommends that a review of the PDS takes place following the receipt of this report (see also section C). This will allow programme design to be informed by the results of the baseline survey and also provide the opportunity to formally integrate the two logframes and develop a MandE framework for the strategy. b) Sampling frame and strata The measurement unit for the survey was households, and the clusters were villages. The sample size was determined to be 94 clusters (villages) x 10 units (household interviews), resulting in a data set of 940 household questionnaires. The sample size was largely determined by availability of resources i.e. the number of interviewers available over the two weeks timeframe of the survey. The difficulty determining sample sizes for multi-sectoral 2

surveys is well documented (due to the diverse range of indicators) but 1000 units is considered to be a good rule of thumb. ‘Representativeness’ is a crucial characteristic for effective sampling, and consultations were held with the staff from the three participating organisations to consider how to make the sample size representative of the situation throughout the province. It was decided to divide the province into two strata, a strata of eastern districts (Anlong Veng and Trapeang Prasat) and a strata of western districts (Bantey Ampil, Chong Kal, and Samrong), to create strata intended to minimised the variation in socio-economic indicators within and maximised variation between the two groups of districts. The sampling effort was then divided proportionally by the relative populations (from the Commune Database 2004) of each strata. The Commune Database 2004 was used as the basis for the sample frame although it is important to note that a) villages where the three organisations did not intend to work in the future due to their urban location/relative absence of poverty were removed from the sample frame and b) due to the high level of migration to and within the province a number (approx. 20) new villages are not recognised in the database. These villages do not form part of the sample frame for the baseline survey (the villages included in the sampling frame are listed in annex D). The number of households per cluster (i.e. 10 interviews per villages) was determined on the basis of logistics (allowing a team of two interviewers completing 5 interviewers each per day to complete one village per day) but also made good sense from a representative point of view with, typically, greater homogeneity within villages than between them, and therefore ‘from a sampling precision point of view, smaller clusters are to be preferred over larger one….a design with 30 clusters of 20 households each would be preferred to one with 20 clusters and 30 households’ (Magnani, FANTA Sampling guide 1997). The percentage of villages reached by the household survey was approx. 40% of those within the sample frame. c) Village and household selection Villages within the sample frame were randomly selected using probability-proportional to size (PPS); giving each household within the sample frame the same overall probability of selection (the selected villages are listed in annex D). Household within each village were also selected randomly using the following technique: • • • •

estimate the number of households in the village; the sampling frame listed families whereas the baseline survey was assessing household indicators, and this required interviewers to discover the number of households in the village (usually be asking the village chief) prior to household selection divide the number of households by number of sample units i.e. 10 = xth household generate a random number between 1 and the nth number = yth household and count to the yth household from the first house along the main road locate each subsequent xth household working along the main road and then each smaller road/track in turn until all households in the village have been counted

Non-response error (i.e. no respondents available within a selected household) was dealt with my moving the next house until a present and willing respondent was identified. Respondents within the household were selected by asking the member of the household to self-select from available household members (children less than 15 years old were not considered suitable) to answer questions about the household. The respondents were predominately women (75% of the total with similar percentages within each district). d) Training 18 interviewers and three team leaders were trained in household selection, interview techniques, use of the questionnaire, and checking questionnaires for error, over a three day period. Two aspects of the training proved especially useful; practical use of the questionnaire in the form of field testing; and use of the questionnaire in a 3

controlled setting using a case study. The later in particular allowed the managers of the survey to get a good understanding of the abilities of each member of the survey team (The group completed questionnaires in response to the answers provided by one respondent enabling each questionnaire to be checked for accuracy). Guidance notes were also produced providing explanations for ‘difficult’ questions and a reminder of the process of quality control. Due to a lack of time, the guidelines were not translated into Khmer. e) Supervision and quality control The interviewers were divided into three teams of six to cover the following three areas based on the required sample effort (1. Samroang and Chong Kal; 2. Bantey Ampil; 3. Anlong Veng and Trapeang Prasat). Interviewers worked in sub-teams of two, allowing the coverage of one village per day by each sub-team (based on a target of 5 interviews each per day) from 31 January to 12 February 2005 (although in the first two days of the survey subteams interviewed as a pair). Supervision of interviewers was carried out by three assigned team leaders with responsibilities for logistics, supporting teams of interviewers, and quality control of the interview process. The later involved checking all completed questionnaires for error, including missing or unclear responses, in which case interviewers were to be instructed to return to the household to find the required information. Interviewers were instructed to check their sub-team mates completed questionnaires before submitting to their team leader. In addition, the ZOA advisor supervising the survey was required to carry out extensive checks on completed questionnaires and quality control is discussed further in a lesson’s learnt paper produced by the ZOA advisor. f) Data entry and analysis Data entry was carried out by three staff from the participating organisations. Each data entry staff was responsible for inputting the questionnaires results from one team. A separate PC and spreadsheet was used for each team, and data was input into a template excel spreadsheet. A key problems encountered was the lack of columns available in Excel which made input of multiple responses clumsy i.e. several responses in one column, preventing ease of data analysis. A further problem was caused by the time taken to translate uncoded question responses. Analyse was carried out by the consultant using both Excel and SPSS software. Further data cleaning (in addition to that carried out by the ZOA Advisor) and the time taken to import the data suggest that data entry directly into SPSS would be a more time efficient strategy for future surveys. The results of the analysis are presented in the next section and the results for household data, community development, agriculture and food security, land and natural resources, and health components were also presented back to field interviews and team leaders for interpretation on 22 March (comments have been documented in annex F).

A.4 Focus Group Discussions Eight villages, four from each strata, were selected non-randomly (from those selected for the household survey) for FGDs (see annex D for villages). Discussions with three focus groups; women, men, and community leaders took place in each village, resulting in a total of 24 FGDs. The information required from the FGDs and a suggested questioning route was developed by the consultant and the ZOA community development advisor (included in annex C). A team of 5 facilitators participated in one day on refresher training on FGD techniques. In addition, the team was made familiar with the information required from the discussions and the suggested questioning route (although they were actively encouraged to modify the prepared list of questions), and developed a number of participatory methods to use within the FGDs. A second day was spent carrying out practise FGDs using the techniques developed the previous day. The FGD team worked in groups of 2-3, with one facilitator and one recorder between 31 January and 10 February with supervision provided by the ZOA community development advisor. The women’s discussion groups were run 4

concurrently with one of the other two groups to limit the number of unwanted observers potentially influencing the group especially when sensitive questions were due to be asked. The teams were asked to type up their transcripts from the FGDs and send to the translator the day following each discussion. Limited time was available for training in recording and on reflection this would have been very useful. The transcripts did not do justice to the quality of the FGDs, requiring the consultant to work through each transcript with each facilitator or recorder to fill in gaps either in the recording or lost in the translation from Khmer to English. The transcripts were analysed using standard qualitative techniques, grouping responses into common themes. It is important to re-iterate that FGDs were not intended to provide data for statistical analyse but to provide insights from a small number of case studies to compliment and possible provide explanation to the quantitative data derived from the household survey. The results from the FGDs are therefore presented separately from the household survey results within each section of the results chapter.

A.5 Key informant interviews To set a baseline for the co-ordination components of the PDS is was considered essential to interview provincial office directors and this was considered to be a more appropriate task for an ’independent’ consultant rather than a staff member from one of the participation organisations. A questioning route, detailed below, was developed by the consultant specifically focused on the co-ordination objectives and indicators and was conducted by the consultant with 4 provincial officials (the Provincial Governor and the Director of the Health, Rural Development, and Agriculture Departments) and the results are summarised in section C.10; Co-ordination. Guide questions for informant interviews: • • • • • • • • •

Which organisations do you currently work with? What is their role? What needs to happen for your work with NGOs to be successful? How do you co-ordinate NGO activities? What data do you collect to help you measure progress? How could I get a list of NGO activities in the province related to your sector? How have you been involved in the development of the PDS? Can you provide any feedback on a) the process of developing the PDS and b) the content of the final plan What do you think the role of NGOs in the province should be?

A.6 Lessons learnt During the baseline survey a series of reviews took place and this in addition to the reflections of individual staff involved in the survey has enabled a paper on lessons learnt to be complied by the ZOA Advisor. See also section C for recommendations on methodology.

5

B. Results 1. Household data 1.1 The sample population A population of 5330 people in 940 households was covered by the household survey, as described by gender, age group, and district in table 1 below. Table 1: Population statistics for households surveys (n = 940) Age group/gender Anlong Veng Bantey Ampil Chong Kal Samroang Trapeang Prasat Total

M<1

F<1

M 1-6

F 1-6

M 6-12

F 6-12

M 13+

F 13+

Total M

Total F

Total

18 17 17 31

13 23 13 15

51 86 47 83

46 96 63 98

51 151 66 123

58 130 76 106

232 521 242 410

253 574 278 438

352 775 372 647

370 823 430 657

722 1598 802 1304

14

8

81

51

85

89

297

279

477

427

904

97

72

348

354

476

459

1702

1822

2623

2707

5330

The population covered by the household survey accounts for approx. 4.95% of the total population within the sample frame (based on MoP/SEILA 2004 Commune Database; using a mean of 5 people per household), see also section A methodology. The sample population data is highly consistent with the data for the whole province described within the 2004 database. There is a very close male: female ratio in the sample total and within each district (approx 0.95 men: women) with the exception of Trapeang Prasat district which is the only district with a greater ratio of men to women (the commune database also describes Anlong Veng as having marginally more men than women). Household size varies between 1 and 14 people, with a mean size of 5.7 within the sampled households. The mean value from the commune database is 5 people per family. This difference may be reflected in the fact that the database is measuring families whereas the 2005 household survey is using households as the ultimate sampling unit (approx. 2% difference per village). There are slightly larger household sizes in the east strata (mean value of 6 people per household in both Anlong Veng and Trapeang Prasat) than in the west, with the lowest household sizes in Bantey Ampil (mean value of 5.3 people). Approx. two-thirds of the population within the sample are older than 13 years old (a ratio of 1:2 <12 years old : >13 years old), with the youngest population in Trapeang Prasat (a ratio of 1:1.8 <12 years old : >13 years old). Figure 1: The population appears quite stable with only 30 households having moved to their village in the last year (3% of households in the sample). Samroang has the highest rate of new settlers, with 6% (14 households) of its households arriving to live in their village in the last year. The principal motive for those who have moved to live in their village in the last year has been to seek land (59% - 18 households). 83% of household are headed by males (780 households), and figure 1 shows that there is slight variation by district with the highest percentage of male headed households in Chong Kal (89%) and the lowest in Anlong Veng (77%). 6

1.2 Household roof type Figure 2: Interviewers were asked to make observations of roof type as this is widely regarded to be a proxy indicator of the relative wealth of a household (see for example Participatory Poverty Assessment: Cambodia, ADB 2001). The results (summarised in figure 2) show that 60% of all households sampled have grass roofs (the 2004 Commune database records this figure as 63.5%). 38% of households in the survey have metal roofs. There are significant difference at district level, with the highest percentage of grass roofs being in Trapeang Prasat (83%) and the lowest in Chong Kal (41%) which is also the district with the highest percentage of metal (53%) and tile roofs (5%).

1.3 Household assets Table 2 below summarised the responses to a series of questions about the ownership of household assets. Table 2: percentage of households owning households assets by district (n = 940) Asset

Pans

Bicycle

Moto

Oxcart

Kuyorn

car/truck

Water pump

Radio

TV

Anlong Veng Bantey Ampil Chong Kal Samroang Trapeang Prasat Total

49% 50% 51% 53% 58% 52%

68% 60% 53% 70% 66% 63%

8% 11% 12% 20% 15% 14%

14% 46% 57% 37% 16% 37%

10% 9% 8% 5% 5% 7%

0 0 0 1% 0 1%

4% 1% 0 3% 2% 2%

27% 28% 27% 27% 35% 29%

32% 14% 26% 15% 21% 20%

The above data can be summarised as follows: • • • • • • •

Less than 1% of households own a car/truck (count = 1) Less than 10% of households own koyorns, and water pumps Less than 15% of households own a moto, ranging by district between 8% in Anlong Veng to 20% in Samroang 20% of households own a TV, ranging by district between 14% in Bantey Ampil to 32% in Anlong Veng 29% of households own a radio, with 27-8% of households in all districts owning a radio with the exception of Trapeang Prasat where 35% own one Over 50% of households own pans with very little variation between district (the question asked about frying pans due to a misinterpretation of the questionnaire) Over 60% of households own a bicycle, ranging from 53% ownership in Chong Kal to 70% in Samroang

7

2. Community development 2.1 Awareness of VDC

Values

Village Development Committees (VDCs) have been established with support from SEILA throughout the province and provide a potential mechanism to organise and support community development activities. The household survey asked a number of questions to gage the level of awareness of VDCs and to find out what actual activities they carry out. Figure 3 below shows the 43% of the sample answered ‘yes’ to the question ‘do you have a VDC?’. It also shows clear district variations, with over half the respondents in Samroang (51.7%) answering yes, whereas less than one third of respondents Figure 3: answered yes in Chong Kal (31%) and just over onethird in Trapeang Prasat (38%). To test the level of awareness of VDCs a following question was asked 'Do you have a VDC?'; ‘can you name a member of the VDC?’. Table 3 percentage by district (n = 940) demonstrates that only a small percentage in each 100% district, and 5% in total (20 respondents) who 47% 42% 31% 52% 38% 43% positively identified the presence of a VDC in their 75% village could not name at least one of the members. This suggests that in a small number of cases VDC’s 14% 14% have been established but have been inactive and/or 50% 12% 13% 55% 13% have little presence throughout their villages (discussed 13% 48% 46% 44% 41% further in section 2.8). It is anticipated that village 25% 36% level factors (especially commitment of individual yes no VDC members, attitude of VC, and local level don't know 0% community development support provided by NGOs) Anlong Ve ng Chong Kal Trape ang Prasat Bante y Ampil Samraong Total are more likely to affect the presence of VDCs than District district level factors. It is therefore also interesting to look at the variation in response to this question within each village. Table 3: Comparison of 'do you have a VDC? and 'can you name a member of the VDC?'; percentage of households per district

District Anlong Veng Bantey Ampil Chong Kal Samraong Trapeang Prasat Total

know name of at least one VDC member

Household responses can not name a VDC member

don't have or don't know if have VDC

43% 38% 28% 44% 31% 38%

4% 5% 3% 7% 7% 5%

53% 58% 69% 48% 63% 57%

Figure 4 below provides a description of the presence of VDCs within each village (i.e. the percentage of respondents able to name at least one VDC in their village). The data suggests that VDCs are not widely known to all households within each village. The median value (i.e. the middle value) is 30% of all respondents, which means that in half of the villages sampled (46 villages) a maximum of 30% of respondents are able to name at least one VDC member. In 4 villages (Am. Thmei, Chhou Stap, Kok Mon, and Peam Knong) 100% of the respondents 8

were not able to identify the presence of the VDC and/or name a VDC member, whereas in 1 village all respondents could name at least one member (Sambour Meas). The variation within villages is markedly different within Trapeang Prasat and Chong Kal which both have a median value of 20%, whereas Anlong Veng, Bantey Ampil, and Samroang have a median value of 40%.

Figure 4: Percentage of respondants per village able to name at least one member of VDC (n = 345)

25

No. of villages

20

21

15 13

10

12

13 10

8

5 4

4

4

70

80

1

1

90

100

0 0

10

20

30

40

50

60

Percentage of respondants per village

2.2 The role/activities of VDCs In addition to presence of VDCs it is also important to understand the type of activities that they carry out within their villages. Table 4 below summarises the number of different roles provided by households in response to the question ‘what does the VDC do in your village?’ Over 70% of respondents (approx 278 households) were able to name at least one role for the VDC (30.3% of the total household sample). There are slight variations between districts, ranging between a maximum of 78.8% in Chong Kal and minimum of 66.1% in Trapeang Prasat of respondents able to identify at least one role for their VDC. It is also interesting to note that although Chong Kal has the lowest VDC presence indicators, the data suggests that where people are aware of their VDC, they see their VDC as being active (as indicated by the number of active roles identified by respondents). Table 4: Number of active roles of VDC identified by household respondent by district (n = 398)

District Anlong Veng Bantey Ampil Chong Kal Samraong Trapeang Prasat Total

Number of active roles of VDC identified by household respondent 4 roles don't know None 1 role 2 roles 3 roles 10.7% 4.1% 4.3% 11.0% 17.9% 9.0%

17.9% 24.0% 17.0% 17.8% 16.1% 19.3%

46.4% 38.0% 66.0% 33.9% 42.9% 42.0%

23.2% 33.1% 12.8% 31.4% 23.2% 27.4%

1.8% 0.8% 0% 4.2% 0% 1.8%

0% 0% 0% 1.7% 0% 0.5%

Total can name role 71.4% 71.9% 78.8% 71.2% 66.1% 71.7%

It is anticipated that the actual variation between number of roles will be highly dependant on the probing skills of individual interviewers, however, table 5 below suggests that a that high degree of generalisation can be made about the type of roles for VDCs within the sample. 9

Table 5: Frequency of roles for VDC by district (n = 403): Role of VDC Makes the VDP Representing people Nothing Managing community resources Don’t know Resolve conflicts within households Other Resolve conflicts with authorities

Percentage by district Bantey Anlong Ampil Veng

Trepeang Prasat

Total

31% 21% 14%

39% 17% 13%

36% 19% 15%

24%

8%

4%

13%

4%

3%

8%

16%

7%

9%

4%

3%

4%

9%

6%

3%

0%

1%

10%

3%

3%

2%

2%

1%

3%

0%

1%

Samraong

Chong Kal

36% 19% 13%

40% 30% 17%

34% 16% 18%

11%

4%

8%

Table 5 above demonstrates that the most frequently provided role for the VDC is to make the Village Development Plan (VDP), accounting for 36% of all responses (145 households). Representing people was the second most frequently cited role, 19% of all responses, with the exception of Bantey Ampil where managing community resources was the second most frequent accounting for 24% of all responses. The data demonstrates that VDCs are generally not taking a role in conflict resolution (both within households and with authorities). A key indicator of community development within the PDS is the level of consultation of people within the VDP process (leading to the development of Commune Plans). 15% of the total sample (141 households) are aware of the VDCs role in making the VDP and this is explored in more detail below.

2.3 The Village Development Plan Those who identified the presence of a VDC in their village were later asked ‘do you know what is written in the VDP?’ and the results are provided in figure 5 below. From this data we can not conclude if VDPs are being produced, but it does show that only 35% of the sample (141 households) are able to identify at least one thing that is included in the plan (15% of the total sample). Figure 5: There is a small amount of district variation with the highest level of awareness in Trapeang Prasat (41%) and lowest in Bantey Ampil (28.5%). The 141 respondents provided the following details of what they saw as the main benefits in the plans for their household: • 49 water related – 35% of responses • 32 road/path related – 23% of responses • 28 wells – 20% of responses • 16 school construction – 11% of responses • 10 improve life in family – 7% of responses • 7 no use – 5% of responses The majority of responses appear to be construction related which suggests, as expected, that the VDP is being used as a mechanism to request infrastructure type projects via the commune development fund. 10

2.4 Village meetings Figure 6:

Interviewees were asked ‘do you attend village meetings?’ 71% of the sample go to village meetings and the results by district and gender are displayed in figure 6. This provides an indicator of the level of participation in decision-making and also within the village development plan (although people may not be aware of the VDP itself their opinions may still be heard at the village planning meetings and used in the formation of the plans by the VDC). It is important to note that although gender disparities appear at district level, the overall sample showed only a slight difference between men and women (with 75% and 74% attending village planning meetings respectively). The main gender disparities are in Chong Kal, with significantly more women than men attending the meetings (71% to 55% respectively). Whereas in Anlong Veng and Bantey Ampil there are very high attendance rates for men (87 and 88% respectively) compared to those for women (73 and 71% respectively). Further data on attendance by gender is provided by the FGDs and is considered in section 2.8.3. It is a big step to assume that attending meetings means that people actively participate within them. To explore this assumption in more detail, a following question was asked ‘do you talk in the meetings?’. The results show that 54% of respondents that go to the meeting also talk in them (39% of the total sample) and the results by district and sex are provided in figure 7. Again although there are gender differences at Figure 7: district level, it is important to note that within overall sample there was very similar results for the percentage of men and women talking in the meetings (54% and 53% respectively). The results vary greatly at district level with only 34% of respondents who participate in the meetings also talking in them in Bantey Ampil, whereas in Trapeang Prasat 73% of respondents talk during the meetings. The main gender difference at district level appears in Samroang where only 48% of women who attend village meetings also talk in them (compared to 74% of the men). Figures 8 and 9 below show the reasons provided by respondents why they do not attend the meetings and reasons why they do not talk in the meetings respectively. The results measure the number of responses provided separated by men and women.

11

Although the actual counts differ dramatically, the overall percentage for men and women are very similar (due to the high percentage of women interviewed within the survey). It is clear that the most frequent reason given for not attending the meetings is ‘not knowing about them’ (45% of respondents provided this reason). This supports the data in section 1.1. showing the results for the levels of awareness of VDCs within each village.

Number of responses

Figure 8: Reasons for not attending village m eetings; num ber of responses by gender (n = 269)

120 100 80 60 40 20 0

97

24

men

20

39

28

31 8

3 don't know about them

not invited

women

not interested

2

no time

2

too shy/not able to speak in group

15

1

don't know

3

other

3

2

1

3

new to area relative goes instead

The main reasons given for not talking in the meetings are that people are too shy to speak with 115 responses (45% of respondents gave this reason), they do not feel it’s their role with 51 responses (20% of respondents gave this reason) and they don’t understand the meetings with 42 responses (17% of respondents gave this reason). The results shown in figure 7 above do not appear to be influenced strongly by literacy (when cross-tabulated with the literacy indicators in section 7.3).

Figure 9: Reasons for not talking in village m eetings; num ber of responses by gender (n = 253)

120

No. of responses

men

96

100

women

80 60

37

40 20

30

14 3

11

6

8

12

19

17 1

9

0

9

14

0 it's not my role

no one will listen to me

I'm not interested

I don't understand the Too shy / can not speak

don't know

other

nothing to say/ no ideas

meetings

2.5 Participation in community projects The PDS considers that the number of community initiatives to be an indicator of self-help capacity. The results of the household survey (figure 10) show that 65% of households included a member who participated in a community project in the last year. The highest percentage occurred in Bantey Ampil (75%) and the lowest in Samroang (55%). The remaining three districts have very similar results (61-65%).

12

Figure 10:

It is useful to break this total figure down to village level to allow increases in participation of community projects to be measured on a per village basis in the future. Figure 11 below shows the frequency distribution (number of villages) by percentage of households participating in a community project in the last year per village. The median value is 70% (in over half the villages at least 70% of household participate in a community project). The types of community projects that respondents participate in are mainly construction projects (roads, fences, schools, etc) which account for 92% of household members’ participation in community projects. This suggests that community participation takes place to fulfil the co-funding agreement for village development funding provided by PDRD/SEILA. Figure 11: Percentage of respondants taking part in community projects per village (n = 588)

No. of villages

20 18

15

15 13

10 9 5 0

1

1

4

3

0

10

20

30

13

8 5

40

50

60

70

80

90

100

Percentage of respondants per village

2.6 Participation in self-help projects A further indicator of self-help capacity is provided by the question ‘have you carried out any new projects in the last year to improve life in your household?’ The overall results are provided in figure 12, which show that 38% of the households sample carried out new projects in the last year to improve life in their households. The results by district range from a low of 28% in Bantey Ampil to a high of 52% in Anlong Veng. It is again useful to break this total figure down to village level to allow increases in self-help household projects to be measured on a per village basis as this may be stimulated by village level factors (such as the presence of an NGO). Figure 13 below shows the frequency distribution (number of villages) by percentage of households carrying out self-help projects in the last year per village. The median value is 40% (in over half the villages at least 40% of households carry out selfhelp projects). In 4 villages none of the households sampled carried out self-help projects in the last year.

13

Figure 12:

The projects have been split into type in figure 14 below. It is important to note the broad interpretation of self-help projects which is dominated by livelihood activities not restricted to in/around the household itself (the original intension of the question) including growing new crops, setting up a small business, improved existing agriculture, and use of forest/forest products (within this including chamkar rice cultivation and wood cutting).

Figure 13: Percentage of respondants per village carrying out self-help projects (n = 940) 16

No. of villages

14

15

14

12

12

10

15

12 10

8 6 4 2

5

4

4

1

0

90

100

0 0

10

20

30

40

50

60

70

80

Percentage of respondants per village

Figure 14: Number of new projects initiated for household self improvement in the last year (n = 350)

No. of responses

160 140

136

120 100 80

64

65 47

60

35

40

19

11

7

20

3

3

1

0 Sta rte d

g ro

Sta rte d win g

ne w

Im p

rov ed t

as ma ll b cro p

Im p

rov ed h

he ho us us e ine ss

Ot he r yg ien e

Ma ki n g

Liv Im Us Sil Fis pro kw es eo hin t oc v f o g e for rm ha de kp e nd f xis st rod a icr r mi / fo tin uc aft ng ga res tio s gri n tp cu ro d ltu uc re ts

2.7 Summary of community development indicators 14

The situation in each village related to the functioning of community development structure such as VDC is likely to be dependant on village level factors and so care must be taken with summary statistics for the sample or on a district level. Table 6: summary of community development indicators

% can name VDC member % of above can name at least 1 role of VDC % aware of VDP % go to village meetings % of above talk in meetings

Anlong Veng

Bantey Ampil

Chong Kal

Samroang

Trapeang Prasat

43% 71%

38% 72%

28% 79%

44% 71%

31% 66%

35% 76% 59%

29% 76% 34%

30% 67% 68%

39% 66% 54%

42% 72% 73%

Table 6 summarises by district from the sample a number of the community development indicators, which shows that: • • • •

approx. one third of respondents are aware of at least one VDC member by name approx. two thirds of those aware of the VDC members by name are also aware of at least 1 activity they carry out approx. one third of respondents are aware of the VDP more than two thirds off households in all districts go to village meetings at which approx. half of these people talk.

The most notable district variations are: • • •

in Bantey Ampil where a low percentage of the households sampled who attend meetings talk in them (34%) in Chong Kal where a low percentage of men interviewed attend village meetings (55%) in Samroang where a low percentage of women interviewed who attend village meetings talk in them (48%)

2.8 Data from the FGDs and the Needs Assessment The FGDs specifically focused on the topic of community development within both the men’s and women’s group meetings (a total of 16 FGDs), asking questions related to the establishment of the VDCs, their roles, participation within decision-making, and any specific role towards supporting the poor. This data compliments and enriches much of the data derived from the household survey. Consideration is also made of the data within the needs assessment. 2.8.1 Establishment of VDC All FGD groups explained that they had a VDCs which was established with support from SEILA, including the election of members (between 1999 – 2001) by village members. It is clear from the FGDs that no subsequent elections have taken place in the 8 villages (although some members have resigned due to lack of available time) and in most cases have not been replaced by new members (or as in one case by unelected members). Members of VDC also expressed the concern of a lack of support since their establishment and appeared to be lacking a sense of direction, capacity and purpose. The length of time since establishment without subsequent elections may 15

provide the reason for only 38% of the household survey being able to name at least one member of VDC in their village. Further potential reasons are identified by the Needs Assessment; many villagers may have moved into their area since the establishment of the VDC and are therefore unaware of it’s existence. The later explanation would also suggest that VDCs are not active or are not serving the entirety of the people in their villages. The within village variation of awareness of VDC (see section 2.1) suggests that the people serving on the committees and their work is not known by the majority of people within their villages. This can be further explored by looking at the roles that the VDC take according to FGD participants. 2.8.2 The role of VDCs The FGD responses overwhelming identified the role of the VDCs as communicating with NGOs. This is in contrast to the household survey (see section 2.2, >10 responses identified this as a role contained within the other category in table 5) which clearly identified the VDP as the main role for the VDC. This may be in part be down to terminology with FGD responses including ‘making activity plans’ and describing planning for specific activities, these type of responses will have been coded as VDP in the household survey, and suggests that the FGD also identified VDP as a key activity for the VDC. 3 out of 8 villages in which FGDs took place (37%) talked specifically about VDC led meetings with villagers, whereas in 6 villages the participants talked about the high degree of consultation and joint decision making that takes place with the VDC and VC within their village which indicates (as does the household survey) that in the majority of villages village meetings do take place and certainly involve people beyond VDC members alone. It is important to note that much of the responses group the VC with the VDC and (as identified in the needs assessment) the two often seem synonymous or even one and the same. There were few examples provided by the FGDs of VDCs providing specific support for the poor in their villages. In Prasat Ibuek it was explained that ‘when villagers suffer starvation only ask neighbours, but also village has united rice store bank to support suffering families’ and in Tamann it was explained that ‘if starvation or serious illness occurs the VC and VDC write down the family name and sent to inform commune chief or red cross’. The HC equity funds were recalled by 3 villages as a mechanism to provide free health care for the poorest people but the important people identified to facilitate this were the VC and/or VHSG. The FGD data suggests that neighbours are the main source of support for those suffering lack of food, whereas the household survey results identify the main strategy to be to relatives (borrowing food and/or money from relatives) as described in section 5.1. It is important to state that in 15 FGDs (out of 16 focused on VDCs) participants saw the role of the VDC as important or very important, and only one group out of the 16 expressed the view that ‘they do nothing here’. However, the FGDs clearly identified that they saw the most important role of the VDC as communicating with NGOs, and a potential concern emanating from the FGD transcripts is villagers perception of NGOs, as one women explained; ‘VDC activity is very important because they are the parents of the villagers in promoting and urging to make relationships with other NGOs as a result of writing requests for NGOs to disseminate programs while having meetings with NGOs that come for developing the village as promoters in the community and propose to continue their obligations for ever’. The above statements and others like them (together with the results from the household survey which shows that one-third of respondents carried out a self-help project in the last year) resonate with the concerns identified in the Needs Assessment that many people do not believe that they have the ability to improve their own lives which may have been in part caused by the focus on relief approaches in the province and a lack of emphasis to date on empowering communities to support themselves. 16

2.8.3 Participation in decision-making The FGD data strongly suggest that VDCs are highly consultative with people in their villages, in 6 villages the FGD respondents specifically mention their involvement in discussing actions and decision making with the VDC at village meetings. This is consistent with the household survey which identified that over 70% of all respondents attend village meetings. In most cases women appear more active than men in village planning and decisionmaking (6 villages, including all in the east strata). It was explained in the FGDs that this was due to the lack of time men had to participate in meetings due to their role in income generating activities such as leaving the village for migrant work or being busy in the rice field, forest or as a policeman or soldier. This is consistent with the household survey data which identified higher percentages of women’s participation than men’s in the east strata. In all the 8 villages the women’s FGD also explained that they had the opportunity to take part in decision making and share their ideas. The responses given seem to indicate that this situation has changed within recent memory, many responses positively describing women’s participation and role in their village/society end with words such as ‘now’ or proceeded by ‘nowadays’. In terms of representation on the committees by gender, all VDCs in the FGD villages have both men and women serving on the committee, and in most cases there is a good gender balance (certainly more balanced that the data from PRDC, included in the needs assessment, suggests). There is one example of a male dominated committee (Phnom Kor in Chong Kal) and two examples of female dominated committees, both in the east strata (Toul Tasek, Thnal Tteng), the latter possibly for the same reason identified in the paragraph above (i.e. high female participation in community life due to men being busy with income generation commitments). Regardless of gender balance, the lack of subsequent elections (in some cases since 1999) questions the representativeness and accountability of VDCs. This is extremely important when considering if people from different social strata are represented on the committees (or if it is possible for them to be elected due to illiteracy, time, or social status), and because there was little evidence from either household survey or FGD data that the role of VDCs includes helping the poorest people in their villages (although there is no link between the group within the ‘poorest’ group identified by the survey and attendance in village meetings, see section 5.8). VDCs are often established as a way to empower communities, and to provide a mechanism to reach and support the poorest rural people, yet there is the risk that their formation concentrates power and opportunity into the hands of a few. The needs assessment summarised this predicament well; the poorest people often feel neglected by NGOs whilst NGO workers often find it difficult to work with or target the poor. The PDS has a clear objective to decrease the marginalization of vulnerable groups within its community development objective and its needs to plan carefully how the VDCs it supports can be encouraged to fulfil the objective, if at all (see recommendations in section C).

17

3. Landmines and UXOs According the household survey data, in the last year: • •

1 child was injured by a landmine/uxo, 1 had an amputation, and there were no reported deaths to landmines/uxos amongst children (>15 years old). 10 adults were injured, the majority in the east strata (4 in Anlong Veng and 5 in Trapeang Prasat), 22 adults had amputations due to landmines/uxos (more than half of these in Bantey Ampil) and 3 were killed

The household survey also attempted to measure the number of people within the sample population with a landmine/uxo related disability. 99 people (53% of all people with disabilities) were identified by respondents as having landmine/uxo related disabilities. This represents 1.9% of people in the sample (population of 5330) and 4% of households having a member who is a casualty of landmines/uxos (not including death). The greatest number by population occurred in Trapeang Prasat district, where 3% of people in the sample have a disability caused by landmines/uxos.

18

4. Land and natural resources 4.1 Land titles 17% of the sampled household responded ‘yes’ to the question ‘do you have a land title?. There are marked variations between districts as shown in figure 15 below. Figure 15:

Figure 16:

The highest percentage of land titles exists in Samroang where 45% of respondents claim to have a land title (although the data set is smaller due to an error in the wording of the questionnaire which discount the first 100 responses, therefore the number is 59 households), and low responses were recorded in Anlong Veng (7%), Trapeang Prasat (7%) and also Banteay Ampil (9%). The following question, ‘is the title temporary or permanent?’ shows that 24% of those with a land title have a temporary title (142 households) and 76% have a permanent title (638 households), as displayed in figure 16. There is a marked district variation in Trapeang Prasat district with 90% (9 households) of respondents having a temporary title and 10% (1 household) permanent, however it should be noted that this result is based on a very small data set due to the low number of positive responses to the previous question in this district.

4.2 Households with no land Only 2% of households (21) do not appear to have access to land to grow rice, fruit or vegetables. The highest number of these households are in Samroang (14), followed by Bantey Ampil (5), and 1 each in Anlong Veng and Trapeang Prasat. 98% of households in the survey therefore have access to land for agricultural production, quite how stable this situation remains unclear due to the uncertainly surrounding the legal status of land titles as described in annex F. It is also unclear if not having access to land is am accurate poverty indicator. In many of the cases within Samroang it appears that households are not accessing land to grow rice, fruit or vegetables because they are running a small business or employed as a professional worker or civil servant.

4.3 Deforestation An explicit objective in the PDS is to protect the environment and reduce deforestation. The principle causes of deforestation are commercial logging and clearance for chamkar rice production. The household survey asked 19

questions to allow us understand more about the important of chamkar rice production to households living in this area and to estimate the extent of forest area being cleared. The results show that a total number of 270 households are engaged in chamkar rice production. 27% (251 households) of the households sampled grew upland wet season rice (chamkar) in the last year on their own land with an additional 2% (20 households) growing on borrowed or rented land. These figures include 1 household growing chamkar on both their own and rented/borrowed land. The summary statistics presented in table 7 below show that there is very small difference in the size of chamkar rice production according to ownership of land. Table 7: Summary statistics for chamkar rice production (n = 270) Own land 0.88 hectares 0.96 hectares 1 hectare 222 hectares

Mean Median Mode Sum (total area of land)

Borrowed/rented land 0.86 hectares 0.72 hectares 1 hectares 17 hectares

The data provided by the household survey is also useful to locate geographically the areas where deforestation caused by chamkar rice production is the greatest. Table 8 shows the area of land used for chamkar rice production in each district Table 8: Area of land used for chamkar rice production (n = 270) Number of households Anlong Veng Bantey Ampil Chong Kal Samroang Trapeang Prasat Total

78 48 12 43 119 270

% of household’s sampled within district 65% 17% 8% 19% 79% 29%

Total area of land (hectares)

% of total area

63 24 5 37 109 239

26% 10% 2% 15% 46% 100%

Chamkar/househ olds in sample (hectares) 0.5 0.08 0.03 0.16 0.72 0.25

The data shows that clearance of forest for chamkar rice production is most affecting deforestation in Trapeang Prasat district (109 hectares of land cultivated for chamkar) and in Anlong Veng (63 hectares), a clear division by strata (the eastern strata accounting for 72% of the total area of land used for chamkar cultivation within the sample). The data also allows us to pinpoint the locations of high chamkar rice production. The following 8 communes (of which 7 are in the east strata) account for 175 hectares of chamkar rice production (over 73% of the total area of land used for chamkar cultivation within the sample): • • •

In Anlong Veng; the communes of Anlong Veng and Trapeang Prey In Samroang; Kon Kriel commune (and specifically in Anlong Veng village) Trapeang Prasat; in 5 out of the 6 communes (all communes except Thlat)

4.4 Income from sustainable use of natural resources 20

The analysis of income sources for sampled households is provided in section 5.3. An indicator in the PDS for assessing income from sustainable use of natural resource management is total and average per family income from resin tapping increases in 15 villages. The household survey could not generate the level of detail required to set a baseline for this indicator. However, the results of the income sources do show that resin tapping is a significant income source for 14 households in the survey (1.5%), with most of these households being located in the east strata (8 in Anlong Veng and 4 in Trapeang Prasat). In addition, 3 households responded with resin tapping as their first significant income source.

4.5 Data from the FGDs and the Needs Assessment The men and community leader’s groups (a total of 16 FGDs) were asked specific questions about use of natural resources, and in particular about the importance of forests and fishing to their livelihoods. This information compliments the data provided on deforestation in section 4.3 above as well as more generally the significant income sources data described in section 5.3. Consideration is also made of the data within the needs assessment. 4.5.1 Forest resources The participants in all groups in the west strata described the destruction of the forest in the period 1993-4, whereas it appeared to be generally later in the east strata 1999-2000 (with the exception of Phdeak Chru group which described the peak period of forest destruction in the early 1990’s). All groups explained that the destruction was still continuing in the proximity to all villages by large scale logging but also by clearance to make chamkar. Participants in one village also described a few cases of people in the village owning chainsaws for use in small scale logging. It is clear from the FGDs (as also concluded in the needs assessment) that the forests continue to play an important role in people’s livelihoods strategies. Participants in all villages were able to name a high number of roles that forest resources play in their lives including provision of rainwater, materials for building houses, vegetables for eating, provision of quality farming land, charcoal, fertilizer (‘leaves especially make the best quality’), can keep animals there, prevent natural disasters such as floods etc. In addition, all groups also demonstrated a high level of understanding of the need and desire to sustainably manage forest resources. All the groups expressed their concerns over illegal logging and the difficulty to control it due to the involvement of ‘power people’ and the army. The FGD confirmed again the information in the needs assessment that large scale illegal logging continues to happen with strong protection by perpetrators. However, 6 of the groups specifically talked about their desire to establish community forests near their village (mostly with support from the government or NGOs). Two groups in the east strata explained that the poor families rely of chamkar, with the view given in one of the groups that they ‘should continue do it this way because they are so poor’. In the other group they thought that enabling poor households to access draught animals would be the only way to reduce deforestation, allowing these households to farm the same plot of land over a longer period of time. Despite the clear importance of forest resources, it is interesting that the household survey did not show forest related activities as a significant income source for many households (see figure 28). This may be partly down to coding and forest related activities are concealed within the ‘labourer’ and ‘other’ categories. Resin tapping certainly only provides a very small number of responses, whereas the needs assessment identified this as ‘an important source of income….with at least some families in almost all villages in he eastern districts (up to 30 per village are involved’. In addition, charcoal production provided few responses but again is identified as another significant income source in the needs assessment. These contradictions will need to be discussed with field staff to better understand if the difference is due to survey error, a case of highly localised livelihoods strategies, or demonstrating already the effects of deforestation (i.e. the cutting of resin trees). 21

4.5.2 Fishing Fishing appears to be a generally less important livelihood strategy for participants in the groups. Only three of the groups (Thnal Tteng, Tamann, Phdeak Chrum) stated the importance of fish but it is unclear from the FGD transcripts if this refers to fishing as a livelihood strategy or the importance of eating fish (which may have been bought from elsewhere as seems to be the case in the province). In addition, only one group identified availability of natural resources such as fish in the strengths analysis (albeit with a rank of 4 and score of 3 out of 41). The group in the east strata, and in Tamann and Romchek in west strata, explained that illegal fishing and an associated decline in fish stocks started in 1999-2000, mainly being carried out by the army, soldiers, and ‘power people’, although the group also added that most villages, having seen these techniques, now also used them but not to the same extent due to the lack of weapons/equipment. The two remaining groups from the west strata (Phoum Kor and Prast Ibeuk) recalled that illegal fishing and decrease in fish stocks started occurring much earlier, since 1980-1. All groups identified that illegal fishing still continues and their descriptions of the situation confer with the needs assessment which identified that ‘throughout the province fisheries resources are under heavy pressure from people (usually soldiers or police) fishing with electronic gear and explosives over which there is virtually no control’. As stated above, fishing now appears to be less important to most of the groups, and this agrees with the household survey which identified that 5% of responses for significant income sources were fishing. It is difficult to make generalisation across the province because the importance of fishing is expected to be highly localised (as described in the needs assessment), with 60% of households in the survey also reported to have eaten fish in the day before the survey. However, this may also indicated that people are now buying fish from elsewhere such as Seam Reap province or Thailand. All the groups stated their desire to reduce illegal fishing, although again they identified their own, and the fisheries authorities, lack of power to enforce the law or report offenders. Four groups suggested the establishment of community fisheries (3 in the east strata) and two groups proposed making a pond in their village for fish farming.

22

5. Agriculture and food security 5.1 Food security 74% of households in the sample have experienced food shortages in the last year (696 households) as shown in figure 17. The data suggests that food shortage is widespread throughout the province, being slightly higher in the east strata (both Anlong Veng and Trapeang Prasat the incidence is ≥80%), whereas in the west strata the incidents ranges between 76% (Chong Kal and Samroang) and 65% in Bantey Ampil. The mean and median values for the number of months of food shortage in the last year are both 4 months; with the maximum number of 6 months also the most frequent accounting for 220 households (32% of households experiencing food shortages and 23% of the entire sample). Figure 17:

Figure 18:

Table 9: Total number of months of food shortage per district District Total months of food shortage Mean number

Anlong Veng

Bantey Ampil

Chong Kal

Samroang

Trapeang Prasat

Total

456 5

716 4

407 4

634 4

535 4

2748 4

Table 9 above shows that the total number of months of food shortage within the sample is 2748 months, with each district having a mean value of 4 months, with the exception of Anlong Veng which has a mean value of 5 months. The results in figure 17 show that Anlong Veng is clearly the district most severely affected by food shortages, although food shortage also appears to be a highly significant issue throughout all districts. The main reasons for food shortage perceived by respondents were low crop yields (511 responses), with lack of land (224 responses), lack of access to credit (178 responses), illness (159 responses), and lack of draught animals (91 responses). Together they account for 89% of all responses. Figure 18 shows that the distribution of rice bank committee’s varies greatly throughout the districts, with the greatest number located in the two districts experiencing the greatest problems with food shortage (Anlong Veng and Trapeang Prasat). The lowest number, according to the data, are located in Chong Kal where only 3% of 23

respondents answered positively to having a rice bank committee. Only 24 household respondents could not name could not name a member of the rice bank committee which indicates that where the committee’s exist they have a strong presence throughout their village. The presence of a rice bank committee however does not appear to effect coping strategies to food shortages. Figure 19 below shows responses to food shortages, with borrowing rice from relatives being the strategy used by the majority of households (420 households; accounting for 34% of all responses). Whereas borrowing from the rice bank accounts for only 6% of responses (used by 81 households).

Frequency of responses

Figure 19: 'When your family experienced food shortage last year, how did you try to solve the problem?'; frequency of responses (n = 696) 500 400 300 200 100 0

420

68 Se ll a

70

102

81

67

83

5

82

Bo Bo Ea Ot Bo Fin No he rro Borro rr o t le r thi da r wm wf ss ss ng w m row r n ets r f e i o o c on w o m o e( ne d ey f o r ic rel y( od (le eb ativ re l c ro nd an ati es er) p k ves ) )

110 18 Fis

hin g

La bo ur

81

Co l lec

39 Ma ke /s e t/u ll h se an na di c tur raf al r ts es ou rce s

5.2 Household nutrition 5.2.1 Vegetable production 52% of households have land available for growing vegetables, however, only 64% of these households actually grew vegetables in the last year (312 households), with slight variation by district ranging from 74% in Bantey Ampil to 54% in Chong Kal. The majority of this land is owned by the households (98%) with only 12 respondents (2%) growing vegetables on borrowed or rented land. Figure 20:

Figure 21:

24

The size of land is typical of small scale production, predominantly <0.5 hectares in size (89% of those growing vegetables in the last year). This date is important to Figure 22: understand the role of vegetables as both an income source (see also section 5.3) and for household nutrition. The motivation to grow vegetables appears to be mainly to sell rather than for consumption (as also identified in the needs assessment); there is a clear pattern within all districts of households growing vegetables only to sell and to both sell and for their own consumption (54% and 45% respectively as shown in figure 22). The majority of households growing vegetables are growing between 2-4 types (40% grow 3 types of vegetables, 27% grow 2 types and 16% grow 4 types). The number of types of vegetables is higher in the east strata (58% grow 3 types in Anlong Veng and 50% grow 3 types in Trapeang Prasat). It is also interesting to see that there is a more year round vegetable production in the east strata, as shown in figure 23. 75% and 64% of Figure 23: households growing vegetables do so in the dry season in Anlong Veng and Trapeang Prasat respectively. Whereas in the west strata dry season vegetable production is much lower (the east districts accounts for 63% of households growing in the dry season, and the three western districts account for 36%). The type of vegetable being grown is shown in figure 24 below. 311 households are growing a total of over 14 types of vegetables and generated 886 responses to the question ‘what type of vegetables do you grow?’. 6 types account for over 75% of the responses; mung beans (161 households are growing), cabbages (125 households), cucumber (117 households), pumpkin (99 households), waxbottle gourd (85 households) and eggplant (84 households). Figure 24: Types of vegetables being grown; frequency of responses (n = 311)

55 5

0 ni on

28

7

O

gu W or ax d /b ot tle gu or d M or ni ng G lo ry C ab ba ge s Sw ee tp ot at o

Sp on ge

pk in

3

Pu m

Eg gp la nt C uc um be r

M un g

be an

3

M us hr oo m

56

125

85

M ai ze

99

C as sa va

58

be an lo ng Ya rd

84

117

C hi lli

161

200 150 100 50 0

25

5.2.2 Fruit production 76% of households in the sample (714) have an area of land where they grow fruit trees as shown in figure 25. The results are very similar in all districts, with the exception of Trapeang Prasat where 91% of the households have an area of land for growing fruit. However, 12% of the households (117) don’t actually have any productive fruit trees on their land, with a big difference between strata, in the east strata 47% and 28% don’t currently have any productive fruit trees on their land for growing fruit in Figure 25: Trapeang Prasat and Anlong Veng respectively. The total number of fruit trees gown in the sample is 7939 trees; with a maximum number of 158 trees, a mean of 11 trees, and a median value of 7 trees. In addition, the 75th percentile is 15 trees which indicates that a small number of households have a large number of trees (possibly orchards) distorting the mean above the median (making the later a more reliable measure of the average value). There is marked variation in the number of trees by district with a mean of 14 trees in Anlong Veng and Chong Kal, 13 trees in Samroang, 9 trees in Bantey Ampil, and 8 trees in Trapeang Prasat. The number of types of trees grown has both a mean and median of 2 types of tree, and a maximum value of 9 trees. Figure 26:

Figure 26 shows that 76% of households growing fruit within the sample do so for their own household consumption. The only variation at district level in Trapeang Prasat where 69% grow for household consumption. In total 12% of all households in the sample are growing vegetables and fruit for family consumption. 14% of households in Anlong Veng and Samroang are growing both for family consumption, 13% in Trapeang Prasat, 12% in Chong Kal, and 8% in Bantey Ampil.

5.2.3 Diet 99% of the sample (933 households) ate 2-3 times in the day before the survey (59% are twice and 40% ate three times). There is only slight variation at strata or district level, with Chong Kal having the highest percentage of households eating twice in the day before the survey (74% - 111 households) and Bantey Ampil having the highest percentage eating three times (49% - 141 households). In terms of food eaten the day before the survey, 99% ate rice, and <10% ate bor-bor (8.4%), noodles (2%), eggs (9.6%). Table 10 below summarises the results for fish, prahok, meat, vegetables, fruit and oil. 26

Table 10: Summary results for food ate the day before the survey (n = 940) Food type

Percentage ate in day before survey

No. of households

Fish

60%

567

Prahok

65%

610

Meat Vegetables Fruit Oil

18% 65% 15% 29%

165 613 157 271

Highest percentage by district(s)

Lowest percentage by district(s)

Chong Kal – 73% Samroang – 69% Chong Kal – 81% Samroang – 72% Anlong Veng – 23% Chong Kal – 77% Bantey Ampil – 19% Samroang – 34%

Trapeang Prasat – 51% Anlong Veng – 52% Chong Kal – 10% Bantey Ampil – 53% Trapeang Prasat 7% Anlong Veng – 19%

5.3 Financial status Figure 27 shows the percentages of the number of different significant income source per household. The majority of households have 2 significant income sources (55%, 517 households), followed by one income source (26%). Chong Kal district shows the greatest variation from the Figure 27: overall situation, with higher percentage of households having more significant income sources (61% have 2 sources, and 27% have three). Households in Bantey Ampil appear the most vulnerable to failures in rice harvests, having the highest percentage of households (38%) with only one significant income source (predominately paddy rice production). The types of household’s significant income sources are summarised in figure 28 below. Paddy rice production dominates the responses, with over 40% of the total responses and a significant income source for 777 households (83% of the sample). In addition, it is the first significant income source given by 700 households (74% of the sample). Other important income sources are selling livestock (231 responses), labour (156 responses) and selling fruit and/or vegetables (134 responses).

Frequency of response

Figure 28: 'In the last year, w hich activities generated significant incom e for your household?'; frequency of all responses (n = 940) 1000 800

777

600 400 200

134 156 3

231 77 34 23

2

7

49

3

52

9

14

6

7

75

6

18 10

2

15

0

2

2

116

0 Ric Ric Ho La Pro Sh Sm Ta Mo Fis Pro Fo Ch Re Se Bu Buy Tha Sew Ric Ric Ba Ma G e Buy Re O th tt t to p he c e res arc s in l li n yin in tc i n e e e p e p me bo u f./G o p all xi e tt / st od rma ss in t /w o al tap g l i g/s g/se h m g win mi ll ery c we a mone s ell r ./Bar r rod rod g a rer ov (b u bus e m ing h a vi n y i ce /C g f oo ma pin ve s el li n l li n ak i n uc uc rd e ( da ern ys / i ne as rg i g fro is h d c k in g toc g tion tion n c y - . w sel s s g n ak ino m ng utt ing k ve g ric e g g rel - ri - c rops to -d ork e ls go i n / a at i c e ha fru g r o y ( d ve it pa mk veg wo s) s d d ar /fr rk) ui t y )

27

Figure 29:

Figure 30:

Figure 29 shows that 23% of households (216 households) have experienced increased income over the last year, whilst 26% feel that their income has decreased (244 households). Income appears to have increased for the most households in Samroang (25%), and the least in Chong Kal (19%). Income has decreased the most in Trapeang Prasat (36%) and the least in Chong Kal (16%) of households. Improved rice yields (52%) and increased livestock production (17%) are the two main reasons given by respondents for income increasing. Conversely the main reasons for those experiencing decreasing income are decreased rice yields (39% of responses), lack of land (19% of responses), lack of access to credit (14%) and illness (12% of responses). In 13% of households, members contributed to household income through migrant work, with fewer in the east strata (8% in Anlong Veng and 5% in Trapeang Prasat), as shown in figure 30 above. Table 11 demonstrates that the main types of migrant work are non-farm work in Thailand (42%), farm work in Thailand (23%) and non-farm work in Otdar Meanchey (19%). Table 11: Count/percentage of type of migrant work by district (n = 124) Farm work in Otdar Meanchey Anlong Veng Bantey Ampil Chong Kal Samraong Trapeang Prasat Total

Count % in District Count % in District Count % in District Count % in District Count % in District Count % in District

0 0% 3 7% 2 9% 3 8% 0 0% 8 6%

Type/location of migrant work? Farm work Other work Other work in another Farm work in Otdar in another province in Thailand Meanchey province 1 11% 2 4% 0 0% 1 3% 0 0% 4 3%

2 22% 17 37% 1 4% 7 18% 2 29% 29 23%

4 44% 3 7% 1 4% 13 33% 2 29% 23 19%

1 11% 1 2% 1 4% 4 10% 1 14% 8 6%

Other work in Thailand 1 11% 20 43% 18 78% 11 28% 2 29% 52 42%

5.5 Improved rice yields 28

90% of households grew rice last year (see figure 31). The situation in each district is very similar with the exception of Samroang where 74% of households grew rice last year. There were only two households in the sample growing rice during the dry season. The production of wet season paddy rice varies Figure 31: significantly between strata as demonstrated by the % of district row in table 12 below. Wet season paddy rice production is significantly more widespread in the west strata, whereas chamkar production is significantly more widespread in the eastern strata. For district level variation in chamkar rice production also see section 4.3. In term of total yields by rice type: • •

Of the 640 households (68% of the sample) growing wet season paddy rice there is a mean of 1580 kg Of the 304 households (32% of the sample) growing chamkar wet season rice production there is a mean of 1066 kg.

Table 12: summary statistics of rice yields by district District No. households % in district Mean Standard deviation Maximum Minimum Median Sum Area of production (hectares) Yield kg/per hectare

Paddy wet season yields (n = 640) Chong Bantey Anlong Trapeang Samroang Kal Ampil Veng Prasat

Chamkar wet season yields (n = 304) Chong Bantey Anlong Trapeang Samroang Kal Ampil Veng Prasat

152

142

248

51

47

45

12

48

78

121

66%

95%

86%

43%

31%

20%

8%

17%

65%

81%

1588

1264

1914

1375

976

882

458

877

1128

1238

1337

1003

1781

1276

571

712

290

658

1028

1002

7500 113 1230 241403

8000 160 955 179558

17850 80 1467 474620

5000 100 960 70135

2800 180 1000 45870

2400 150 700 39675

1200 160 355 5500

3600 40 850 41220

6400 100 855 87990

5400 200 1000 149852

304

298

936

59

37

37

5

24

63

109

794

603

507

1189

1240

1072

1100

1718

1397

1375

Total yields for paddy wet season rice production are the lowest in Trapeang Prasat (with a mean value of 976kg), which conversely is the district with the highest total yields for chamkar rice production (mean value of 1238kg). There is a clear divide by strata for chamkar rice total yields with Anlong Veng also having a mean of 1128 kg, whilst the mean west districts total yields range between 458kg in Chong Kal to 882 in Samroang. The highest total yields for paddy wet season rice production are in Bantey Ampil with a mean value of 1914 kg and the lowest 29

in Trapeang Prasat with a mean value of 976kg. The statistics also suggest a high degree of variability within the rice yield distributions, the standard deviation values are very low for the low yielding districts (Trapeang Prasat for paddy rice, Chong Kal and Bantey Ampil for chamkar) and high for the high yielding districts (Bantey Ampil for paddy rice) indicating that a small number of high yielding households are distorting the mean value (this is further indicated by low median values compared to the mean). In terms of rice yields per hectare, table 12 also demonstrates that the highest yields for wet season paddy rice production are in the east strata (which is interesting as Trapeang Prasat also has the lowest total yield), whereas the highest yield per hectare for wet season chamkar rice production is in Bantey Ampil. The household survey data is in agreement with the needs assessment which identified that for people with land rice farming is central part of their livelihoods strategy. Although the ownership of land can not be reliably established, the household survey can conclude that for the vast majority of people growing rice, it is a central part of their livelihood strategy (83% of households growing rice in the last year gave rice production as their first response to significant income sources as described in section 5.3 above).

5.6 Soil improvement 992 responses were received from 604 households for the question ‘in rice, vegetable, or fruit production what do you do to improve the soil?’ 64% of households are therefore using a form of soil improvement in crop production, and the types by district are presented in table 13 below: Table 13: number of households using soil improvement techniques

Anlong Veng Bantey Ampil Chong Kal Samraong Trapeang Prasat Total % of households in sample using

Produce Compost

Use manure

Use green manure

Use organic pesticides

Use chemical pesticides

Use chemical fertilizer (rice)

Use chemical fertilizer (veg.)

Total

7 1 0 2 4

31 202 103 118 52

0 1 0 2 3

0 1 0 3 4

0 86 3 9 4

5 113 110 59 11

13 16 5 11 13

56 420 221 204 91

1%

54%

0.6%

0.8%

11%

32%

6%

-

There is very little use of compost, green manure organic pesticides and use of chemical fertilizers on vegetables crops throughout the districts. The main soil improvement technique is the use of manure which is used by 506 households (84% of those using soil improvement techniques). As a percentage of all households, the use of manure is most common in Bantey Ampil (70% of households) and Chong Kal (69% of households). The use of chemical fertilizer on rice fields also appears widespread in these two districts, with 39% of households using this technique in Bantey Ampil and 73% in Chong Kal. The use of chemical pesticides is only used widely in Bantey Ampil, where it is used by 39% of households. There is no direct link between rice yields and use of soil improvement techniques; for example, within the three western districts which are the major paddy rice producing districts (total yield >100,000kg), Samroang has the highest yield per hectare but the lowest use of chemical fertilizer (25%) and manure (50%).

5.7 Ownership of livestock 30

89% of households own livestock (as shown in figure 32). There is a small degree of variation at district level with Bantey Ampil having the highest percentage of households owning livestock (92%). Household ownership of livestock can be broken down into the following types: • 48% own pigs Figure 32: • 75% own chickens • 17% own ducks • 50% own cows • 6% own buffalos • 32% own draught cows • 5% own draught buffalos Table 14 shows that Bantey Ampil’s high percentage of households owning livestock is accounting for by over 70% of households owning pigs and chickens. It also has the highest percentage of duck ownership (21%), and buffalo ownership (10%). The two eastern districts have lower percentages of ownerships by livestock type, especially for pigs and draught cows, with the exception of chickens. Table 14: Summary statistics for livestock by district (n = 940) Pigs Dist. Freq. % Mean S,d. Max Med. Sum

Chickens

Cows

CK

BA

AV

TP

SR

CK

BA

AV

TP

SR

CK

BA

AV

TP

SR

CK

BA

AV

TP

103 45 2 1 8 1 192

82 55 2 2 15 1 145

174 76 2 2 25 1 326

39 33 1 1 6 1 54

55 37 2 2 13 1 104

171 74 8 8 40 5 1338

104 69 6 8 70 1 681

213 73 6 7 35 3 1281

94 78 4 4 20 3 383

120 80 5 5 50 4 591

37 16 5 3 15 5 179

28 19 5 4 20 2 139

62 21 5 3 15 5 332

12 10 4 2 8 3 45

23 15 4 4 22 3 91

125 54 3 2 10 2 355

92 61 3 1 10 1 241

148 51 3 2 13 2 395

47 39 3 2 7 2 126

61 40 3 2 12 2 183

Buffalos Dist. Freq. % Mean S.d. Max Med. Sum

Ducks

SR

Other

Draught Cows

Draught Buffalos

SR

CK

BA

AV

TP

SR

CK

BA

AV

TP

SR

CK

BA

AV

TP

SR

CK

BA

AV

TP

11 5 3 1 5 2 32

14 9 2 1 3 1 26

28 10 3 1 6 3 82

1 1 7 7 7 7

2 1 1 1 1 1

5 2 3 2 7 2 15

0 0 0 0

1 0.3 1 1 1 1

0 0 0

0

85 37 2 1 4 2 172

77 51 2 0 4 1 153

104 36 2 1 5 2 211

12 10 2 0 2 2 24

22 15 2 0 2 2 41

9 4 2 0 2 2 20

11 7 2 0 2 1 20

24 8 2 1 4 2 51

0 0 0

2 1 1 1 1 1

0 0

Chickens are the most frequently owned livestock type (over 73% of households in all districts) as well as the highest number of livestock with a mean value of 6 birds per household. Buffalos are the least frequently owned livestock type, owned by 6% of all households, with a mean of value of 3 animals per household owning buffalo. Households owning livestock in the two eastern districts also generally own a lower number of livestock, having lower mean values for chicken and duck ownership. In addition, 30 households own a fish pond (3.2% of the sample), with very slight variation by district ranging from a low of 2.7% of households in Trapeang Prasat to 6.7% of households in Anlong Veng. The figures for ownership of pigs and chickens above are lower than those produced by UNICEF is a baseline survey of the eastern districts in 2000 (56% owned pigs, 93% and 81% raised chickens in Anlong Veng and 31

Trapeang Prasat respectively). The figures for chicken and duck ownership are also lower than ZOA’s previous baseline survey in Samroang and Bantey Ampil (81% of families have chickens and 24% have ducks).

5.8 Crop diversity Figure 33 shows that 27% of households in the survey grew rice, vegetables and fruit in the last year (as an indicator of crop diversity). The results show a marked difference at strata level, with 49% and 39% of households growing rice, vegetables, and fruit in the last year in Trapeang Prasat and Anlong Veng respectively. The lowest percentage of households growing all three crops is in Figure 33: Bantey Ampil (17%). It is interesting to compare the results for crop diversity with the food shortage data. There does appear to be a close association between crop diversity and food shortage in all districts (with the exception of Trapeang Prasat). The greatest difference is in Bantey Ampil where 15% of those who grew rice, vegetables, and fruit in the last year suffered from food shortage whereas 85% of those who did not grow all three did suffer from food shortage. The difference is also >50% in Anlong Veng, Chong Kal and Samroang. There is no clear association between crop diversity and change in income.

5.9 Identifying the poorest of the poor An analysis of the household survey data was made to see if it was possible to identify a group within the survey which could be categorised as the ‘poorest of the poor’. A combination of indicators was used to identify this group; grass roof, ≤0.5 hectares of land for rice cultivation, no draught animals, and more than 3 months food shortage a year. This identified 45 households from the survey (5% of households). The majority of these households are located in the east strata (25 in Anlong Veng and 15 in Trapeang Prasat), with only 5 located in the west strata (4 in Samroang and 1 in Chong Kal). Within Trapeang Prasat these household are widely distributed amongst the villages sampled whereas in Anlong Veng the following 6 villages in 3 communes have at least 3 households (30% of sample within each village) which fit the above definition of the poorest of the poor: • • •

in Trapeang Tao commune; Trapeang Tao and Ou Angre in Anlong Veng commune; Romcheck and Yeang Kang Cheung in Trapeang Prey commune; Cheung Phnom and Toul Sala

When cross-tabulating the households fitting the definition of poorest of the poor there is no clear association with any of the following variables; female headed household, number of years lived in the village, literacy indicators, knowledge of VDC, attendance and talking in village meetings, number of times ate in the day before the survey, or infant mortality. There is however, a close association between this group and reduced ownership of many of the household assets, in comparison to the results shown in section 1.3, with 0% of this group owning a kuyon or water pump, less than 5% owning a motorbike or oxcart, and 8% owning a TV.

6. Health 32

6.1. Impact indicators 6.1.1. Severe illness This section of the report will begin by presenting some of the health impact indicators (i.e. the actual health status of people) before working through a series of outcome indicators included in the PDS. Figure 34:

Figure 34 shows that 52% of households (487) had a member who had a serious illness in the last year. There is distinct district variation, with Chong Kal having the highest percentage of households with a member having a severe illness (64%), followed by Anlong Veng (61%). The lowest percentage severe illness was in Samroang (51%). Table 15 shows more specific data for the number of adults and children (by gender) suffering from severe fever and severe diarrhoea in the last six months. When compared to population data derived from the survey, rates of illness can be calculated. The results in table 16 show the highest rates of sever fever to be in the east strata (15% of total sample population in both Anlong Veng and Trapeang Prasat) and the lowest in Chong Kal (9% of the sample population). There is very little variation by gender.

Table 15: number of people with severe fever/diarrhoea in the last 6 months and 2 weeks respectively District Men Anlong Veng Bantey Ampil Chong Kal Samroang Trapeang Prasat Total

Severe fever Women Boys (<15)

40 50 13 40 42 185

38 71 27 55 38 229

Girls (<15)

16 34 17 30 28 125

17 37 12 32 24 122

Men 4 19 4 13 9 49

Severe diarrhea Women Boys (<15) 13 23 9 11 13 69

Girls (<15)

8 26 12 23 15 84

7 21 14 14 7 63

Table 16: rates of severe fever and severe diarrhoea in the last 6 months and 2 weeks respectively District Anlong Veng Bantey Ampil Chong Kal Samroang Trapeang Prasat Total

% of male pop.

Severe fever % of female pop.

Severe diarrhea % of female % of total pop. pop.

% of total pop.

% of male pop.

16% 11%

15% 13%

15% 12%

3% 6%

5% 5%

4% 6%

8% 11% 15%

9% 13% 15%

9% 12% 15%

4% 6% 5%

5% 4% 5%

5% 5% 5%

12%

13%

13%

5%

5%

5%

5% of the total sample population have suffered from severe diarrhoea in the last 6 months, with very little variation between districts and by gender. The highest rates are in Bantey Ampil (6% of district sample population) and the lowest rates are effecting males in Anlong Veng (3%). 33

6.1.2 Mortality Information on maternal and child mortality was also collected and is presented in section 6.2 below. In addition, information on deaths caused by landmines/uxos in included in section 3. 6.1.3 Disability 4% of people in the sample had a disability, with large gender variations; 157 males had a disabilities (6% of the male population in sample) and 47 women had a disability (1.7% of the female population in the sample). 53% of these disabilities (99) were caused by landmines/uxos, 18% (34) are described as physical disability, and 13% (25) as related to ‘seeing’.

6.2 Maternal and child health The PDS aims to improve maternal and child health and the household survey asked a series of question both to establish a baseline for maternal and child health indicators and also to understand more about pregnant women’s health seeking behaviour. The survey results show that there have been 214 pregnancies in the last year in 212 of the households sampled (23% of households). A higher percentage of households had a pregnancy in the east strata (33% in Anlong Veng and 27% in Trapeang Prasat, 40 households in each district). It is very difficult to assess fertility rate based on the household survey because the population statistics described in section 1.1 grouped all females over the age of 13 together rather than separately counting women of reproductive age. It should therefore be made clear that table 17 below is displaying the rates of pregnancies in females over the age of 13 to allow comparisons to be made between districts. It appears that higher rates occur in the east strata, Anlong Veng has the highest rate with 17% of females over the age of 13 becoming pregnant in the last year and a rate of 14% in Trapeang Prasat, whereas the rates in the west strata range between 9-12%. Table 17: percentage of females (over 13 years old) pregnant in the last year District

Population count (F 13+)

No. pregnancies

% of pregnancies in F 13+

Anlong Veng Bantey Ampil Chong Kal Samroang Trapeang Prasat Total

253 574 278 438 279 1822

42 49 33 50 40 214

17% 9% 12% 11% 14% 12%

The survey recorded that one woman died as a result of child birth in the last year (in Trapeang Prasat) and 20 children died (widely distributed with at least 1 death in each district). The data set is too small for meaningful comparison between groups or districts; however, it is possible to estimate the infant mortality rate for the survey population. There are 169 children under the age of one year living in the household survey area which means that 10% of children have died in the first year of life (calculated from a total of 189 births). The infant mortality rate would therefore be 106 deaths in the first year of life per 1,000 live births. Figure 35 below shows that 62% (126) of pregnant women visited a health facility prior to giving birth. The high number of responses indicates that women recently pregnant have also answered this question; this may have had the effect of lowering the actual rate as they may not yet have needed to visit a health facility. The lowest rates are found in the east strata (44% in Trapeang Prasat and 58% in Anlong Veng), whereas in the west districts the rates are all >66%. 34

Figure 35:

Of the 126 women visiting a health facility, the majority visited health centres (81% - 102 women or 50% of all pregnant women). The number of times visited ranged between 1 – 3, with 52% visiting 3 times, 25% twice and 24% once. The lowest number of visits occurred in Chong Kal and Trapeang Prasat (33% and 35% of women respectively have visited once). 202 respondents answered the question ‘who supported the pregnant women during her last delivery?’ This question assumes that the women pregnant in the last year have already given birth, which may explain the lower number of responses (out of 212 households with a pregnancy) or more likely the responses include the anticipated action for a number of respondent who have yet to give birth. 84% of women are (or expect to be)

supported in delivery by TBAs. This is slightly lower in Samroang (79%), which along with Bantey Ampil has the highest number of deliveries in the HC (both 13%). Chong Kal has the highest number using private clinics for delivery (9%, although the count is only 3 women). Figure 36 shows when women who have given birth within the last year start to breastfeed. 62% of women (123) started breastfeeding within 24 hours of giving birth. Within this number 45% started breastfeeding immediately, which indicates that mothers gave their baby colostrum, although this was not specifically researched. The only significant variation at district level is in Bantey Ampil where 15 out of 48 mothers (31%) either started breastfeeding after 2 days (25%) or not at all (6%). Figure 36:

Figure 37:

Figures 37 shows the length of time that mothers in the sample exclusively breastfeed. 13% of mothers from the total sample breastfeed for < 4 months, with 87% of women (170) exclusively breastfeeding for ≥ 4months. These figures are very similar within each district, however, there is large variation in the length of time beyond 4 months, especially in Bantey Ampil where 33% of mothers exclusively breastfeed for more than 1 year. Figure 38 35

shows the responses to the question ‘what type of weaning food did you give to the baby between 6 months – 2 years?’ The responses are shown as a percentage of all responses and it is clear that bor-bor, rice, and fish are the most common foods given to babies between 6 months - 2 years (41%, 21%, and 12% of responses respectively). Figure 38:

Figure 39:

percetnage of responses

Types of w eaning food given to baby betw een m onths and 2 years; perecentage of responses (n = 193)

50

41 21 8

6

12

6

1

1

5

0 er th O in g r th de No Po w ilk M gs Eg sh Fi s t u i le Fr tab ge Ve bor rBo e

c Ri

The responses to the question ‘do you use birthspacing in your household?’ are shown in figure 39. 32% of the sample (299 households) use a form of birthspacing, and at district level the percentage is much higher at 47% in Bantey Ampil, and lower rates in the east strata (19% and 17% in Anlong Veng and Trapeang Prasat respectively). Two methods of birthspacing are most frequently used, the pill (accounting for 63% of responses) and injection (34% of responses).

6.2 Capacity of communities to monitor and improve family health For communities to have the capacity to monitor and improve family health they need the knowledge to prevent illness and to be able to know how to act when they become ill, and to be able to access reliable health services. To understand the level of capacity the household survey asked questions about: • •

the presence of village health volunteers (VHV) or village health support groups (VHSG) in villages which are seen as key mechanism to provide health education and a link between people and health services awareness and use of health services

In addition section 6.3 below provides specific indicators for health education and health seeking behaviour, including boiling drinking water, covering water jars, and defecation.

6.2.1 Village Health Volunteers / Village Health Support Groups Figure 40:

36

Figure 40 shows that 63% of households (570) responded positively to the question ‘do you have a VHV or VHSG in your village?’ The percentage of positive responses was higher in Anlong Veng (76%) and lowest in Trapeang Prasat (51%). A following question was asked ‘what are the names of the VHV or VHSG?’ which revealed that 7% (41 respondents) of the households responding positively to the previous question could not name any of the members. Figure 41 below shows the responses to ‘what role do the VHV / VHSG carry out in your village?’

Figure 41: 'What do the VHV/VHSG do in your village?' (n = 566)

4

Don't know

14

P rovide health education

1

Help people with diseases P rovide first aid

1 1 0

Sell medicines P rovide medicines Sell condoms

2

20

P rovide information about HC/RH

8

Support new mothers Support pregnant women

21

28

P rovide vaccinations

0

5

10

15

20

25

30

Percentage of responses

It is clear that the main roles for the VHV/VHSG are providing vaccinations (28% of responses), although this is interpreted to mean supporting in the delivery of vaccinations. The next most frequently cited roles are supporting pregnant women (21% of responses), and providing information about the health centre or referral hospital (20%) and health education (14%). 6.2.2 Awareness and use of health services Households were asked where they go when they need to seek medical treatment, generating 625 responses from 490 households, and the results are shown in figure 42. Health centres received the most responses (30%), followed by private clinics (23%) and the provincial hospital in Samroang (16%). 25% of respondent’s first response (assuming this shows their first priority in terms of health seeking behaviour) identified the HC and 25% responded private clinics as the health facility they would visit if they were seriously ill. A further 20% identified the RH in Samroang and 13% identified a hospital outside the province. The lowest percentage for visiting HC was in Samroang (14%), but it also had the highest percentage visiting the RH (39%). Chong Kal had the highest percentage using private clinics (30%) and a corresponding low percentage of people using the HC (17%). The east districts had a total of 1% of respondents visiting the RH in Samroang as their first response. 37

Figure 42: 'When a household member is seriously ill where do they go to seek medical treatment?'; percentage of responses (n = 490) 30

Percentage of responses

30 23

20

16 12

10

7

6

4 1

0

P rovincial

P rovincial

P rivate

Hospital

Hospital (outside

Clinic/ hospital

(Samroang)

province)

Kru Khmer

Health Centre

Local P harmacy

Market

0 TBA

Other

Households were then asked if they know where the health centre is and if a household member had been to the health centre in the last year. The results are provided in figures 43 and 44 below. 82% of households (770) know where the health centre is. The results are slightly higher in the east strata with 90% responding positively in Trapeang Prasat and 83% in Anlong Veng. The lowest results are found in Chong Kal where 77% of households in the sample know where the health centre is. 58% of households (356) that know where the health centre is, have a household member who has visited the centre in the last year. The results shown in figure 44 are highest in Bantey Ampil (69% of households who know where the health centre is) and lowest in Chong Kal (43% of households who know where the health centre is). Figure 43:

Figure 44:

The 356 households who have members who visited a health centre in the last year were asked ‘how did you rate the service?’ and the results are provided in figure 45 below. A very small number found the service to be excellent (1%), 71% found the service to be ‘good’’, 21% found it to be average and 7% found the service to be ‘bad’. The main variations by district are in Bantey Ampil where 87% found the service to be ‘good’, and in Trapeang Prasat 38

where 17% found the service to be ‘bad’. Households were asked for their reasons for their responses and the positive responses can be summarised as (unless indicated responses are evenly distributed by district): • • • • •

good attention (99 responses, with over 50% of these coming from Bantey Ampil) friendly (35 responses) take care (26 responses) good relationships (15 responses) don’t take money (14 responses)

The negative responses tend to be the opposite of the above with 28 feeding back that staff do not care (the majority from Samroang), 22 responding that staff do not pay attention (50% of these from Chong Kal), and 14 that they have to pay a lot of money. Figure 45:

Figure 46:

Figure 46 shows that 32% of households in the survey used a hospital in the last year. This was highest in Samroang where 45% of household used the hospital and lowest in Trapeang Prasat where 13% of households used the services of a hospital in the last year. It is important to note that he question was intended to ask specifically about usage of the Referral Hospital in Samroang, but due to an error in the questionnaire, the results include all hospitals (expected to include hospitals in Seam Reap and Thailand).

6.3 Improved health and quality of life through access to potable water supply, sanitation, and hygiene education 6.3.1 Access to health education

39

Figure 47:

43% of adult household members (401) have received health education messages in the last six months. Figure 47 displays the results by district, with the higher percentage in Anlong Veng (61% of adult household members) and the lowest in Bantey Ampil (34% of adult household members). The 401 households provided 482 responses to the question ‘from whom did you receive the message?’ with 39% of responses being from HC staff, 36% by NGO staff, and 13% by VHV/VHSG.

6.3.2 Access to drinking water Figure 48 below shows that in the dry season, the majority of households take their drinking water from a bore hole (24% - 226 households) or traditional well (23% - 216 households). In addition, lakes and river/canal/drainage channels are the source for >10% of households (16% - 150 households and 11% - 103 households respectively). During the rainy season the primary water source is rainwater (45% - 423 households) and traditional wells (19% 179 households). The remaining rainy season drinking water source each account for ≤ 5% of household drinking water sources. Figure 48: ‘what is the main source of drinking water for your household?’ (n = 940) 1 3 0

bought - truck rainw ater river, canal, drainage channel lake community pond

45

11

4 5

16

3

8

3

8

12

community pond/filtered borehole ring w ell and pump traditional w ell

45

open ring w ell

0

5

10

24

15

15

19

23

20

25

Dry season Rainy season 30

35

40

45

50

Percentage of households

Figure 49 below shows percentage of households accessing ‘safe’ drinking water sources in the dry and rainy seasons. For the purposes of this analysis, the following drinking water sources are considered to be ‘safe’ water sources; boreholes, ring wells with pumps, filtered community ponds, rainwater, and bottled water.

40

Figure 49: safe drinking water sources by district Dry season

Rainy season

100% 75%

80% 60%

56% 54%

60%

55%

78% 69%

40%

40%

36%

28%

20%

65%

9%

al

as at

on g

al K

pi l

Ve ng

0%

There is a marked difference between the percentage of households accessing safe drinking water sources in the dry (36%) and rainy season (65%). The extent of this variation is caused largely by only 9% of households in Bantey Ampil accessing safe drinking water sources in the dry season. The highest percentage is found in Trapeang Prasat at 78% of households. The highest percentage of houses accessing safe drinking water sources in the rainy season are found in Bantey Ampil (75%) and Trapeang Prasat (69%) and the lowest in Chong Kal (55%).

To t

Pr

pe an g

m ra

Tr a

Sa

ho ng

C

m

A

ey

an t

B

A

nl on g

Trapeang Prasat has the highest percentage of households accessing safe drinking water sources throughout the year (dry and rainy season both >69% of households). Figures 50 and 51 below show the distance to water sources in the dry and rainy seasons. The percentages for distance from households to dry season drinking water sources are highly variable by district. In the east strata a large number of households gain their drinking water within 100m of the house (65% in Anlong Veng and 61% in Trapeang Prasat). This is significantly higher than the districts in the west strata where the percentage gaining drinking water within 100m ranges between 39% in Samroang to 19% in Bantey Ampil. The east strata has only ≤10% of households gaining their drinking water over 500m from the house whereas the figure is >26% of households in each of the three western districts. Figure 50:

Figure 51:

The distance to rainy season drinking water sources are much less variable by district than the dry season distances. 81% of households (760) get their drinking water within 100 meters from their house. This is due to the high percentage of households drinking rainwater during the rainy season (45%) which is generally collected from

41

nearby houses with metal roofs. Figure 52: Figure 52 shows that 56% of households identified that they have a Water Point Committee (WPC) in their village. The highest percentage is in Samroang (69% - 159 households), and the lowest is in Bantey Ampil (42% -122 households). However, the results for the following question ‘can you name a member of the WPC?’ (see figure 50) suggest that the committee’s are inactive or not well-known in some villages. This appears to be especially the case in Samroang where 39 households who identified a WPC in their village could not name a member (25%). The highest percentage of households able to identify WPC and name at least one member is in Trapeang Prasat – 61% (92 households). The lowest percentage is in Bantey Ampil – 35% (101 households). The role of WPCs according to responses provided by respondents is shown in figure 53. 30% of responses were safe water education, 21% to help clean the water point and 19% to maintain the water point. 20% of respondents (100 households) did not know what the WPC does, and a further 5% (20 households) thought they did nothing. Figure 53:

6.3.3. Storage and boiling of drinking water Figure 54 shows that 34% of households always boil their drinking water (319 households), and that there is marked variation by strata. The percentage of households always boiling their drinking water in the districts within the eastern strata are much higher (59% in Trapeang Prasat and 50% in Anlong Veng) than the districts in the western strata. 862 provided responses to the question ‘why is it necessary to boil drinking water?’ and the responses are provided in figure 55. Killing microbes accounted for 37% of responses (648 households) and preventing diarrhea accounted for 23% (409 households). It is interesting to note that preventing malaria or dengue accounted for 13% of all responses. In terms of the percentage of respondents who think that boiling water can prevent malaria or dengue, this varied from 31% of household respondents in Trapeang Prasat, 28% in Bantey Ampil, 23% in Samroang, 22% in Anlong Veng and 13% in Chong Kal. 42

Figure 54: Figure 55: 'Why is it necessary for families to boil their drinking water?' (n = 862)

percentage of responses

40

37

30 23

20 13

12

11

10 2

0

Kill

Prevents

Prevent

micr obes diar rhoea malar ia or dengue

0

Prevent

Improved

Good

For

gener al disease

taste

hygiene

visitor s

2 Habit

1

0

Don't know

Other

83% of households own water jars (780 households), with 41% owning one (385 households), 30% two (282 households) and 12% (112 households) three or more (see figure 56). The situation in each district is very similar to the total sample, with marginally more water jars being owned per households in Bantey Ampil (87%, 252 households). Figure 56:

Figure 57:

Figure 57 shows the proportion of water jars with proper covers as observed by interviewers. It reveals that 38% of households fully covered their water jars at the time of the survey (357 households). There are distinct difference at strata level, with the eastern districts having a higher percentage of households fully covering their water jars (59% in Anlong Veng and 50% in Trapeang Prasat), whereas in the western district the percentage is down to 23% of households in Bantey Ampil. Households were also asked if they thought it was necessary to cover their water jars. with 95% responding yes (887 households) with very little variation between districts. The 887 households were then asked why they thought it was necessary to cover their water jars. 2055 responses were recorded from 881 respondents with the main reasons being, to prevent dust/dirt (35% of responses, 719 households) and to prevent mosquitoes/larvae (29% of responses, 597 households). 43

6.3.3 Sanitation Interviewers were asked to observe if the household has a latrine and if it looked used. The results for observation of latrines is shown in Figure 58 with 10% of households (93) Figure 58: observed to have a latrine. Only 9 latrines appeared to be unused. Households were also asked ‘where do your household members usually defecate during the daytime?’ and the results are shown below in figure 59. 67% of households defecate in a field/forest away from the house, and 24% defecate in the house plot area. There are distinct variations by district with higher percentages defecating in the field/forest away from the house in Trapeang Prasat and Chong Kal (87% and 83% respectively). In addition, Bantey Ampil has the highest percentage defecating in the house plot area (39%). Figure 60 shows that 50% (443 households) of those without latrines bury the faeces in the soil. Chong Kal district varies from the others, having the lowest percentage to bury the faeces is in Chong Kal, where 20% do not bury it and 43% only bury it sometimes. Figure 59:

Figure 60:

'Where do your household m em bers usually defecate during the day?'; percentage of responses by district

Total 8 24 TP 4 9 23 AV 9 BA 3 39 CK 6 9 SM 17 21 0%

67 87 68 57 83 61 50%

100%

Ow n latrine In house plot area Field/forest aw ay from house

6.4 Data from the Needs Assessment The FGD provides data related to health from 16 discussion meetings (women’s and community leader’s groups) and is detailed below both to help provide explanations for the above results but also in the case of HIV/AIDS to provide additional data that was not possible to obtain from the household survey. Comparison with the needs assessment is also made throughout this section. 6.4.1 HIV/AIDS

44

HIV/AIDS was discussed with all the women’s groups in detail. All groups believe that HIV/AIDS is now decreasing in Cambodia due to the information that has been disseminated on the radio, TV, from NGOs and MoH. There certainly appeared to be a high level of understanding how the disease is transmitted amongst all the focus groups (although this was specifically tested with men). However, two groups also described the practical problems in putting their knowledge into practise. One women explained that men feels that using condoms is ‘not romantic, like eating fruit with a pill’, and a women in a separate group explained that men did not use condoms and they (the women) were too shy to ask them to do so. In two groups they explained that they did use condoms. In total three groups reported that men did not use condoms, with a woman in one of these groups explaining that women ‘don’t want HIV/AIDS training anymore because know all the issues, instead we need training for the men’. This is in contrast to the needs assessment that concludes that there is a low level of understanding of HIV/AIDS throughout the province. Whilst this in part may be true, it certainly does not appear to be the case amongst the women participants in the 8 Focus Groups. The FGs identified that people have died of AIDS in 3 villages (total of 7 people according to the respondent’s memories). In an additional 2 villages, the respondent’s suspected men had died of AIDS outside the village (because they work outside the village or because when they’re ill they return to the home of their families outside the province in the case of those who have recently settled) but could not be certain why they had died. In view of the number of deaths and the apparent reluctance of men in some villages to use condoms, the apparent widespread belief that HIV/AIDS is decreasing in Cambodia could be a potential danger to protect against the disease. This is especially true in a province such as OM which as the needs assessment identified there is reasons to suspect the risk from the disease is significant because soldiers constitute such a high percentage of the population and their apparent frequent use of prostitutes. The presence of temporary populations, such as contractors for road construction, traders from outside the province (especially at the borders), as well as migrant workers frequently moving away from home to work, are all anticipated to increase the risk of OM to the spread of HIV/AIDS. 6.4.2 Illness The FGD gathered data on trends in malaria, dengue and TB to better understand the extent to which these diseases form part of the severe fever and sever illness data recorded in the household survey. The situation with each disease is described below: a) Malaria; according to FGD data the prevalence of malaria varies between villages. 3 villages talked about people dying of malaria, in Tamann (5 people died in 1990), in Thnal Tteng (many people died in 1999), and most recently in Toul Sala (2 women died in 2003). Malaria is still prevalent in the east strata (having reduced since 2002 in Tamann and not being seen as a serious problem in the other 3 villages) Toul Tasek also reporting reductions since 2002. There was a general perception that malaria was decreasing even in the east strata, with the reasons for reduction being health education being provided through NGOs and MoH trainings. Although in Thnal Tteng and Phdeak Chour participants explained that it still happens because people are not using nets in the forest, drinking dirty water, not boiling their water, and also due to a reduction in forests which has consequently reduced the number of mosquitoes. Certainly the use of forests, including for chamkar rice production is much higher in the east strata (and incidentally also in Tamann) where the incidents of malaria is the highest. The needs assessment also identified that mosquito nets are often left at home when a family member goes to work in the forest for the children to use. b) Dengue; many villages reported high cases of dengue fever in the following specific years: 45

•2001 – Tamann (2 children died), Phoum Kor, Romcheck (3 children died) •2003 – Tamann, Phdeak Chroum, Prasat Lberkg, Pdkeak Chrum, Thnal Tteng, Toul •Toul Tasak also in 2004

Sala

However, all groups reported that this was currently not a problem due to a better understanding how to prevent it such as use of bed-nets and improved living standards. c) TB; high rates of TB were described, including deaths in three villages (Tamann and Toul Sala in 2002 and in Phoum Kor in 2003). In addition, two other villages reported individual cases of TB. There are no reported cases of TB since 2003, and in Tamann this was thought to be due to increased understanding amongst the people because of both the information sent by MoH and NGOs to the village and also an increased understanding amongst HC and RH staff of how to treat it. 6.4.3 Maternal and Child Mortality (MCM) 5 groups could recall deaths to mothers due to child birth and/or children in the first year after birth, which is summarised in the table 18 below. Table 18: MCM according to FGDs Village Tamann Prast Ibeuk Romcheck Phdeak Chrum

Toul Sala

Description provided in FGD •In last 4 years – can remember 1 women and 3 children •1 mother died in 1993 •in 2001 a mother and baby died •in 1997 a mother died •2003 one women and baby died •1993 were one woman died •1990 two women and children dead •2004 two new born baby died •2002 one delivery mother died with baby •There were high mortality and child dead rate in 1991; 4 mothers and baby died

It is not possible to compare these figures reliably with those gathered by the household survey, because the recall periods differ and participants have only reported the cases they are aware of. However, all groups certainly felt that the MCM rates had decreased in the last few years (only one trendline was completed for MCM and this showed sharp decline from 2003 onwards). Reasons for this change or previous occurrence were provided in 6 villages as follows: •4 •1 •1

due to presence of HC and improved TBA skills because before could not afford hospital fees and improved TBA skills due to lack of HC in the past

The role of HC appears important in at least 5 of the villages for the perceived reduction in MCM, however, its specific role is not clear in all of these villages. In Rom Chek and Prast Ibeuk participants specifically explained that HC are used for ANC and for delivery whereas in Toul Sala participants explained ‘they now knew about high risk and if there is a problem they can go to HC’.

46

The above is consistent with the needs assessment which explained that ‘Malterer found that reproductive health services have increased considerably in the whole province’. However, the assessment also described that in general women do not go to HC for examinations during pregnancy. The household survey data indicates that 50% of pregnant women visit HC before pregnancy (81% of those visiting a health facility, with little variation by district), and the use of HC by pregnant women is also verified in the FGDs as described above. The needs assessment also includes the results of a UNICEF baseline survey which found that almost no women fed their babies colostrum (only 3% started breastfeeding immediately after birth). The household survey indicates a much higher rate of 13% starting to breastfeed immediately, although the assumption that this means they feed their babies colostrums needs to be explored further. 6.4.4. Water Supply The findings of the household survey are consistent with the needs assessment with the exception of boiling drinking water. Data from both agrees that boiling drinking water in the east strata is more common than in the west, however, the whilst the needs assessment found that ‘in the western districts almost no-one boiled water’ the results from the household survey range from between 16% of households in Bantey Ampil and 36% in Samroang. 6.4.5 Sanitation The findings of the household survey are consistent with the needs assessment, although it appears that the number of latrines has increased in Bantey Ampil and Samroang (6% and 21% of households respectively) from the date of the ZOA baseline survey which identified that 2% of households in these districts had latrines. In addition, the household data varies considerably from the UNICEF baseline survey in the eastern districts described in the needs assessment which identified that 80% of households had their own latrine (according to families own responses), whereas the household survey based on observations recorded a figure of 5% of households in Trapeang Prasat and 11% in Anlong Veng. 6.4.6 The use of health services Both the household survey and the needs assessment identified that people use a variety of health services for serious illness (see section 6.2.1), and it certainly can not be assumed that all people will use the state provided health services. It is clear from the FGD data that the provision of state health services has increased dramatically in the last 5 years within the province, including the construction of HC, use of HC services during pregnancy (see MCH section above) and provision of training and information related to health seeking behaviour. The service provided by HC appears to vary between centres, as the following table, which summarises responses to questions form the FGDs about the use of health services, shows: Village/District Phoum Kor Chong Kal Tamann Samroang Prasat Lbergk – Bantey Ampil Toul Tasek Trapeang Prasat Phdeak Chrum -

Comments from FGD ‘use private service, they care for and respect patients’ ‘feel unwelcome at HC, if you have no money the staff don’t care’ ‘When there was someone sick they went to get the treatment at the Health Center and only paid 5 bhts in case of serious illness the health center staff would refer the patients to the Samroang hospital.’ ‘When get sick go to Kok Mon HC 8m away, but in rainy season road is difficult to travel. HC staff are cordial and in good serving especially the poor with ID don’t need to pay. If severe illness send to be treated at Samroang hospital and required to spend much’. ‘when they got sick the villagers they go to Trapeang Prasath HC if severely sick they go to Seam Reap or Anlong Veng. The good attitude of Trapeang Prasath health centre staffs popular for the people’ ‘HC use also but have to pay’ ‘When they have disease, they go to Health Canter that far from village 4 Km, but we need to have pay 30000 Riel

47

Trapeang Prasat Thnal Tteng Anlong Veng Toul Sala Anlong Veng

to sleep at Anlong Veng Health centre. When some one ill seriously we take them to the Health center’ ‘For the Anlong Veng health center is not require the poor people pay money for Health center, if have letter from village heath support group or village chief’ ‘Women with problem referred to Seam Reap hospital by neighbors – via motopod or truck’. ‘Anlong Veng HC don’t care if have no money’ ‘when people get sick they go to health centre, sometime go to private clinic in ALV health centre poor people no need to pay for service fee if some one not so poor they need to pay some. They which one is poor or not so poor by VHSG write a letter for them when they go to health centre’.

The HC staff at Chong Kal received negative feedback from the Phoum Kor group which may explain why this district has the lowest use of HC services (see section 6.2.1 above). The attitude of staff is also commented on in Thnal Tteng, but this does appear (from the household survey results) to affect use of the HC services to the same extent as in Chong Kal, whereas the good attitude of staff at Trapeang Prasat HC is commented on by the participants from Toul Tasek. Many of the groups talk about cost (the needs assessment identified that cost appeared to be a key factor in people’s choice of health services) and in three groups exemption fees for the poor were described positively, although it is unclear if this actually exempts clients from all costs. The needs assessment also identified that VHSG have been established to play a key role in providing feedback to HC and training to people. The assessment discovered that some members ‘have used their positions to start private health practises and are learning to give injection’. The household survey also identified that the main role of VHSG according to the household respondents was giving injections (see section 6.2.1), although this is interpreted to mean supporting the delivery of injections as community facilitators.

48

7. Education 7.1 School attendance The data from the household survey shows that 416 boys (aged 6-12 years old) and 382 girls (aged 6-12 years old) attend government school. Comparing this to the population data in section 1.1 above this represents school enrolment rates of 87% for boys and 83% for girls aged 6-12 years old. The rapid rate of school construction in the province may help to explain the relatively high school attendance data, and it certainly is a factor in explaining why children older than the official enrolment age are entering the education system (see below). For children aged over 12 years old, 366 boys and 258 girls attend government school. The attendance rates are very similar when broken down by district for boys and girls in both age groups. These results can not be compared with the population data derived from the survey as this combines all people aged 13 years and above or the 2004 Commune database which measures school attendance by age group 6-14 years. There are also 47 children below the age of 6 who attend school. The data suggests that this is more likely to be due to children attending primary school early rather than the presence of pre-schools due to the wide geographical spread of the children. In age group 6-12 years old 131 boys (31% of boys attending school in this age group) and 95 girls (25% of girls attending school in this age group) started in the last year. Over 70% of children starting school in the last year were aged between 6-8 years old (33% were 6 years old, 24% were 7 years old, and 13% were 8 years old). There were 143 responses (15% of the sample), to the question ‘if they are older than 6 years old, why did they start school last year?’, with the main reasons for starting school described in table 19 below: Table 19: Main reasons for starting school Reason Count Percentage

Reduced cost

Child too small

26 18%

No school/ too far away

25 17%

19 13%

No school before 16 11%

Greater involvemen t in school 11 8%

Afraid

Need to help at home

9 6%

9 6%

It is difficult to draw firm conclusions for the data because it is a very small data set and in addition, the question proved difficult for the interviewers to use. Some of the responses do not all seem to fit the question – in view of this ‘child too small’ has been interpreted ‘as child was too small before’ and no school/too far away’ as before there was no small/or it was too far away’ (and could be integrated with its adjacent category to become the main reasons for starting school later than the official enrolment age). 386 households responded to the question; ‘If there are any children under 12 who do not attend school, why not?’ and this provided a total of 400 responses. The main reasons are described in table 20 below: Table 20: Main reasons for not attending school Reason Count Percentage

Child too small 236 59%

Need to help at home 81 20%

School too far away 27 7%

Other

Too expensive 26 7%

17 4%

49

The results for this question may contain errors because respondents might have included children below school enrolment age when considering that their child is too small.

7.2 Role of the PTA 27% of the respondents have a PTA in their village and can name at least one member of the association. This varies by district from a high of 34% in Samroang to a low of 19% in Bantey Ampil (In addition, 32% in Trapeang Prasat, 28% in Anlong Veng, and 24% in Chong Kal). The PTAs appear to be relatively inactive at village level; of 304 responses to the question ‘what role does the PTA do?’, 34% of responses were ‘do not know’. The other main responses were ‘encourage students to go to school’ (17% of responses) and ‘school management’ (15% of responses).

7.3 Literacy rates To assess respondent’s literacy rates they were asked to read and write a sentence (‘I love Cambodia’). The results are presented in figures 61 and 62 below which reveal slightly contradictory results; 39% of respondents can read the sentence yet 72% can write the sentence. There is some variation by district but all within 10% of the total percentage figures. Figure 61:

Figure 62:

The literacy rates indicated by the above data show a slight increase from the previous ZOA baseline in Samroang and Bantey Ampil (January 2001) which discovered that 35% of people could read a sentence.

50

8. Women • •

See section 2.4 for women’s representation in VDC, and attendance and participation in village meetings See section 9.2.1 for domestic violence

51

9. Human rights and governance 9.1 Conflicts 96 households (10% of the sample) reported serious conflict in the last year within the household survey, and figure 63 describes the type of conflicts. Of the 96 cases of serious conflict 33 were land related cases, 31 were cases of domestic violence and there were 32 coded as Figure 63: others (these mainly relate to children such as conflict caused by punishing other people’s children, theft, animals eating crops, and being drunk). According to the data 31 households experienced serious conflict in the form of domestic violence (3% of all households sampled). The total number of counts are low and it is therefore difficult to make any firm judgments about difference between district, however, Bantey Ampil has the highest percentage of households with domestic violence (7% 20 households) and Anlong Veng has the highest percentage experiencing land conflict (7% - 8 households). 85% of the conflicts were with other village members or household members (48% and 37% of conflicts respectively). A very small number of conflicts were reported to be with the military or government (4 and 3 cases respectively). 45% of the 96 serious conflicts identified by the household survey were not resolved by anyone (this may also include conflicts that as yet remain unresolved), and 32% were resolved by the Village Chief. 7 cases were resolved by the Commune Council and 6 by the police. This data (together with FGD data described in section 9.2 below) identify the Village Chief as the main person villagers turn to if they need assistance in resolving conflict.

9.2 Data from the FGDs 9.2.1 Domestic violence According to the FGD data there are cases of domestic violence in all the villages although this is not particularly seen as a big problem in most of these villages (this could mean that cases were not reported in the household survey because it is not perceived to be a ‘serious conflict’ by some respondents). Only 2 villages suggested resolving domestic violence in the proposed solutions to conflicts in the villages, which again indicates that it is not seen as a big problem throughout all villages. Estimates of the number of households affected by domestic violence range from 3-5 households (estimates were provided in 5 of the villages) which is similar to the estimates derived from the household survey (see section 9.1 above) and also the needs assessment which estimated 3 cases per village. The causes of domestic violence provided by the FGD participants include drunkenness, lack of income, gambling, and jealousy. As described in the needs assessment, there did not appear to be any mechanism for resolving domestic violence, but in contrast to the needs assessment no examples were provided of victims reporting abuses to the VC or other authorities. Participants explained that it is usually left to be resolved within the household (although in two villages parents in law and neighbours were also reported to get involved in 52

resolving the conflict). One woman described the difficulties in reporting the problem, explaining that when she suffers domestic violence she can not go to police due to threats from her husband. 9.2.2 Conflict between neighbours Conflict between neighbours does not appear to be a significant problem according to most of the FGD participants. It was reported in all villages, but was described as ‘not happening much’, ‘if it does happen it is not acute’; and ‘have but not serious’. The main causes suggested are speaking badly about other families, livestock eating vegetables/crops, and land disputes. All conflicts appear to be resolved either by both parties or with the support of the VC, especially in the case of land disputes. In two FGD it was explained that the VC has a list of boundaries and will use this to measure boundaries on the ground to resolve the conflict. Only one case of land grab was highlighted, with chamkar rice fields being taken away from a participant from Anlong Veng. 9.2.3 Conflict involving youth In all villages this was described in being the most serious conflict, even resulting in 1 death in Tamann village. This appears to be an issue throughout all the villages when parties or ceremonies are taking place; groups of youth come from other villages (described as ‘big brother’ by FGD participants) and fight with youth from the village, which often starts during dancing or because of a girl. Only one village thought they had resolved this type of conflict, although the women’s group appeared to disagree. In other cases, the FG participants thought the police were even too scared to intervene in fear of revenge attacks from the youth (it was explained that the police no longer have weapons whereas the gangs often do). FGD participants suggested several strategies for reducing this type of conflict, including: •Parents educating their children •Permission for ceremonies to be sought from local authorities to help prevent •Involvement of police and military in controlling the ceremonies •Formation of youth association with support of NGO and involvement of

the conflict

youth in preparing ceremonies and listening to expert speakers during the ceremony •Communicating to people during the ceremonies by local leaders to help prevent conflict

53

10. Co-ordination It is very difficult to generalise about the co-ordination of development activities in the province because only 4 interviews could take place in the time available and these were all with government officials. Therefore, although the following information may provide some insights it should be stressed that it is not a complete analysis and does not include the perspective from all government department or the NGO community. All line ministry departments interviewed appeared to have a good understanding of the development activities taking place relevant to their sector within the province, and could recall the organisations/location/type of projects. These interviewees each described 10+ working relationships with IOs/NGOs, most of whom sent reports to update the departments on their activities. There also appears to be some systems in place to assist co-ordination of development activities in the province including; Excomm, Prococom, and a number of health related meetings (monthly PHD/RH/HC staff management meeting, Health Centre Management Committee meetings). In addition, the establishment of agriculture and water/sanitation networks are also imminent. However, how these forums are used to promote government/CSOs engagement, beyond reporting on activities, is unclear. All interviewees thought that NGO representatives should participate within all the above forum, and one interviewee felt that NGOs should be using the meetings to a greater extent as an opportunity to raise problems, allowing solutions to be discussed within the meetings. The role of NGOs in the province according to the interviewees is principally to implement projects to support the government and to support the people, and also to provide financial support to the province. Advocacy was not mentioned by any of the interviewees as a role for NGOs. Most NGOs appear to be viewed as working partners of the government, and there is generally a very high level of appreciation for their work in the province; all interviewees felt they have very good working relationships with their partner NGOs. Concerns were raised by two interviewees related to a lack of awareness on the part of some NGOs of the context in which they are working and specifically: • • •

a lack of understanding of MoH national policy in the delivery of their projects examples of income generating projects not being linked to the market resulting in failure inappropriate approaches resulting in the lack of sustainability of projects (including lack of maintenance of water points)

With regard to the PDS, although most interviewees could confirm and explain how they had been involved in the process of the developing the strategy, there was uncertainly amongst most interviewees of its purpose. This uncertainly appeared to be because: • •

the presentation of the final PDS had not taken place at the time of interview each department generally maintains relationships/partnerships with >10 organisations, and the PDS is one of number of partnerships in which interviewees are involved.

Recommendations based on this data are provided in section C.

54

C. Recommendations C.1 Methodology As mentioned previously in section B, a paper capturing the lesson’s learnt from the survey has been compiled by the ZOA Advisor. The paper includes: • • • • • •

suggested changes to specific questions and a revised questionnaire ready for the mid-term evaluation in January 2007 the suggestion from the consultant to reduce the number of questions in the questionnaire including identification of specific questions to delete the use of a dedicated full-time translator to support questionnaire design, back-translation, testing and review, and data entry in future surveys the potential use of SPSS for data entry for future surveys (depending on time available to train data entry team) suggestions on reducing the time required for quality control more time included for the production of survey guidance notes for the interviewers, to provide the basis for training and a source of reference for interviewers in the field (including translated into Khmer)

In addition, to the above the consultant also makes the following recommendations for the use of FGDs in the next survey: • • •

to stagger the FGDs to follow the analysis of household survey data. This will enable the questioning route to be based on the result of the survey, identifying additional information needs or allowing the results to be interpreted/explained by a small number of focus groups greater emphasis to be placed on recording discussions within the FGDs through both additional training on recording and the use of tape recorders to capture the entire discussion in each group. to hold FGDs with young people. It is known, due to the demographics of Cambodia, that young people comprise a significant proportion of the rural population. In addition, young people are also seen as the main source of conflict in villages by the adults who were consulted in the FGDs. These two points provide a reminder that there is a need to listen to and understand the perspective of this age group generally in community development projects, and also within future FGDs.

C.2 Programme management cycle The consultant recommends: Logframe • the development of one logframe for the PDS. The logframe in annex E is an initial attempt to integrate the logframes, and further work will be required by the three organisations to complete this process. The process required could involve a joint meeting of staff to review the PDS based on the results of the survey and ensure that the programmes/strategies are clearly articulated in the logframe. The resulting logframe should provide an effective communication tool to summarise, and build understanding of, the PDS amongst all stakeholders (staff, government/NGO partners, and village representatives). • In addition, the logframe could separate aim/impact (related to actual changes in people’s health and/or socioeconomic status); objective/outcomes (changes in people’s/organisations behaviour and practise); and outputs 55

•

(things built/delivered). It could also be merged with the MandE framework to provide a more practical working document; table 21 below provides an example that could be used for the format. the logframe should be seen as a dynamic tool, reviewed and modified if required following each planning cycle using the information gathered for the MandE framework (see below)

MandE framework • An important step will be to work up an MandE framework to ensure that all information required for assessing progress in the PDS is being collected and there is a plan for using the information. Once the indicators have been finalised in the logframe, this will consist of deciding how (which tools), when and by who data will be collected and when the data will be analysed, documented and used e.g. for programme planning or advocacy purposes. The indicators and data collection procedures can be captured in a logframe format such as the example provided below. Guidance information on storage of information (i.e. where to file it once gathered), process for analysis, documentation, and use of the information could be added at the end of the table to complete the MandE framework. Table 21; suggested format for revised logframe Aim Impact Indicators: Aim: To stabilize and improve socioeconomic conditions of vulnerable communities through improved self help capacity, maternal and child health, food security, income generation and access to education

Self-help capacity • % of households participating in village planning meetings • % of households carrying out community projects • estimate of amount of time given to community projects • % of households initiating self-help projects in the last year Health (Maternal and child health) • maternal mortality rate (number of women dying of childbirth per 100,000 of women of reproductive age per year) • infant mortality rate in the first year of life per 1000 • incidents of severe fever in the last 6 months • incidents of diarrhoea in the last 2 weeks Food security • % of households with shortage of food in the last year • Mean number of months of food shortages • % of households with access to rice banks in times of food shortage Income generation • household assets/house construction materials • households diversity of sources of income generation (not impact or it is reduced vulnerability) • Use of non-rice crops (vegetables and fruit) – not impact • % of households experiencing increasing income in the last year Access to education • Estimate of Primary school attendance rates • % of respondents literate

Objectives Outcome indicators (one row for each component) Community development outcomes

Method(s) for data Data collection – collection who/when

• •

Method(s) for data Data collection – collection who/when

% of households aware of VDC and can name at least one person As above for VDP

etc…..

56

C.3 Programme strategies The purpose of the baseline survey was not implicitly to review the PDS, however, the results do help to confirm that many of the strategies in the document are grounded in the situation within the province. In particular, the following strategies appear to be well matched with the priorities highlighted by the survey results: • • • • • •

• • • •

the emphasis on community development (1.1), with the baseline survey showing that VDC are not highly representative within their villages and the need to improve the transparency of the village planning process as a mechanism to lead and mobilise people towards their own development the lack of existing targeted support for marginalised people suggests that strategy 1.8 (direct targeting of particular groups) is a necessary component of the PDS the emphasis on rice banks (1.5) is essential due to the high number of households effected by food insecurity Cattle banks (1.4) targeted in the eastern districts they could provide an effective strategy to reduce the forest clearance for chamkar rice production Support for alternative and sustainable livelihood strategies such as community forests and community fisheries (3.13) meets the demand expressed in many FGDs, as well as the household survey which showed natural resources making a relatively small contribution towards household income within the sample population support to land titling (3.5) appears imperative to mitigate against an issue that appears likely to emerge in the future (due to low percentage of household with official land titles and the confusion surrounding this issue). In addition, although land conflict appeared to be a relatively uncommon problem according to the baseline survey results, the inclusion of land related conflict resolution strategies (3.11) appears to be important to anticipate the significance of this issue in the near future continued support to TBAs (6.5) and improvement of public health service (6.8-9) to continue the recent reductions in infant mortality working with the PHD to find ways to regulate private health services (6.1) is important due to the number of household using the private sector when they face serious illness the baseline results demonstrate (percentage of households boiling water, understanding reasons for boiling water, and covering water jars etc) that continued health promotion is required (6.7) with 36% and 65% of household accessing drinking water from safe sources in the dry and rainy seasons respectively the focus on water point development (6.13) appears a priority

The baseline results also suggest that a greater emphasis within the PDS could be placed on • • •

initiatives that will support the personal development of young people in the province such as the formation of youth groups to carry out community based activities initiatives focused on reducing the risk of HIV & AIDS in the province, especially targeted at men and migrant workers, a well as supporting the integration of HIV & AIDS considerations into other organisations projects within the province (such as road construction projects). consideration of alternative community development approaches – essentially to answer the question – ‘can/how can VDCs be used effectively for community development in Cambodia?’ – by looking at other examples within the country

In addition, the results suggest that to meet the objectives of the PDS there is the need for intensive, sustained and widespread support to villages throughout the province. It is currently unclear in the PDS how/where this support will be provided or prioritised, and so it is recommended that the results of the baseline survey are used to 57

develop/prioritise plans for the implementation of the PDS. In addition, although support to LNGOs and CBOs is contained within the PDS, it is felt that this could become a more explicit set of strategies to increase the availability of resources, ownership in the province, and the provision of long term support required. A continued focus on developing partnerships with INGOs and CBOs to work as part of the PDS is therefore recommended. If the PDS increases the number of partner organisations, it becomes increasingly imperative that a planning cycle is developed which will allow for the regular review of strategies/approaches and to give new partners a voice within this process. Furthermore, the baseline survey provides a potential opportunity to move towards participatory planning and evaluation approaches. If the results of the baseline survey are presented back to target villages, this will allow villages to both verify the results and use them as a basis for planning. The NGOs who facilitate this process can provide guidance and suggestions (using the results of the baseline survey) but ultimately hand-over decision making to the villages, asking them to determine what they would like to achieve in their village over a given timeframe. This could lead to the development of objectives and indicators for each village based on their own aspirations (these will probably fit within the broad MandE framework). Villages could also develop ways to monitor the achievement of their objectives which would move the PDS to adopt the use of participatory monitoring and evaluation. The feasibility of moving to such an approach will need to be considered by staff of Care, Malteser and ZOA.

58

Annexes: •

Annex A. ToR

•

Annex B. Questionnaire

•

Annex C. Focus group discussion guide questions

•

Annex D. List of villages sampled in household survey and FGDs

•

Annex E. Comparison of logframes

•

Annex F. Presentation of results and interpretation