ABOUT THE CAMBODIA SOCIO-ECONOMICS SURVEY

The sampling design in the CSES survey is a three-stage design. In stage one a sample of villages is selected, in stage two an Enumeration Area (EA) is selected from each village selected in stage one, and in stage three a sample of households is selected from each EA selected in stage two. The sampling designs used in the three stages were:

Stage 1: A stratified systematic πps (systematic sampling with probabilities proportional to size) sample of villages was selected. Strata were defined by provinces and the urban/rural classification of villages. The size measure used in the systematic πps sampling was the number of households in the village according to the population census 1998.

Stage 2. One EA was selected by Simple Random Sampling (SRS), in each village selected in stage 1.

Stage 3. In each selected EA a sample of households was selected by systematic sampling.

The design described above was used for the CSES 2004 survey ⁽²⁾.  In 2007, a subsample of the villages, or EAs, in the 2004 sample was selected by SRS. The villages and EAs surveyed in 2007 were thus included in the sample in both years. In each selected EA a sample of households was selected by systematic sampling. The selected households in 2007 are not necessarily the same as those included in the sample in 2004.

The selection of households in stage three was done in field by first listing the households in the selected EA, and then selecting a systematic sample of households. Selected households were observed during one calendar month. The allocation of the households over the months in 2007 was done so that each village in the 2007 sample was observed in the same calendar month as in 2004.

The sample size in 2007 was 360 villages or 3,600 households, compared to the sample for the 2004 survey of 720 villages or 12,000 households.

Some provinces were excluded, due to cost and other reasons, in the sample for 2007. The estimates are however, adjusted for the under coverage error caused by excluding those provinces.

Totals and ratios such as means or proportions are estimated for the total population or for subgroups of the population, i.e. domains of study. The domains are defined by, for instance, age groups or sex. In the CSES 2007 the sample size is not large enough for a detailed breakdown on

e.g. provinces. Means and proportions are estimated by first estimating totals and then calculating the ratio of two estimated totals.

In order to estimate population totals or totals in domains from a sample, weights are needed. An estimator of a population total of a variable is the sum of the weighted variable values for the observed sample units.

The weights are determined by the sampling design, design weights, and adjusted for nonresponse and other imperfections such as under coverage or, adjusted to improve the precision of estimates. In CSES 2007, the design weights were adjusted using preliminary data on the number of persons and households from the population census 2008 and also the Neupert population projections⁽³⁾.

Questionnaires

The fieldwork started in October 2006 and finished in the end of December 2007. The results in this report are based on data from the calendar year 2007, i.e. only 12 of the 15 survey months are included in the estimation.

Supervisors and enumerators were recruited by NIS and trained for the field work. The training took place at NIS in Phnom Penh. A comprehensive field manual was used during the training and the field work.

Each fieldwork team consisted of one supervisor and three enumerators. For each selected village one enumerator was assigned as responsible and carried out interviews of ten households in the village.

Altogether 30 enumerators and 10 supervisors, divided into 10 teams carried out the fieldwork at the same time. Two groups of teams were formed and alternated monthly so that each interviewer and supervisor worked in the field every second month.

For a given month the team arrived in the village 2–3 days before the first day of the month for preparatory tasks like discussing with village authorities, filling in the Household listing form and sampled the households to be interviewed.

The supervisor was the leader of the team and was responsible for the coordination of the interviews, collaboration with local authorities, and checking of the questionnaires during the interview month. If errors were found in the responses the enumerator was required to re-interview. The supervisor was also responsible for the village questionnaire and the interview of the village leader or representative of the village leader.

Any survey of the CSES dimensions needs a comprehensive system of quality management and monitoring. The CSES management group within NIS therefore is using a careful monitoring scheme. The monitoring team included four NIS staff including top ranked NIS officers. The supervision took place during the last two weeks of the interview month.

Before going to the villages the teams were briefed and introduced to adjustments of the interviewing procedure that had to be made as a result of monitoring activities and feed-back from the data processing.

Comparisons of the results from the 2007 CSES with previous surveys, CSES 1993/94, 1996, 1997 and 1999, are not recommended due to differences in the survey design.

The CSES 2004 was conducted from November 2003 to January 2005, and the 2007 survey from October 2006 to December 2007. The monthly sample sizes were 1,000 households in the 2004 survey and 300 households in the 2007 survey. Concerning CSES 2004 the, results presented in 2005⁽⁴⁾ were based on both the 12 months and the 15 months samples, whereas in this report all estimates presented are based on the 12 month sample, i.e. the calendar year 2007.

The weights used in the reports from CSES 2004 are adjusted by using the preliminary population projections⁽⁵⁾  which give over estimated population counts. The weights in CSES 2007 are adjusted by using the preliminary result from 2008 Population Census. A recalculation of the weights in CSES 2004 will be made for the coming analyses to obtain higher comparability between CSES 2004 and the surveys conducted from 2007 and onwards.

The Statistics Law Article 22 specifies matters of confidentiality. It explicitly says that all staff working with statistics within the Government of Cambodia “shall ensure confidentiality of all individual information obtained from respondents, except under special circumstances with the consent of the Minister of Planning. The information collected under this Law is to be used only for statistical purposes.”