Survey of Conflict Prevention and Cooperation 2004
The project uses public opinion polling to gather and then analyze a sample that represents the entire population of each of four different countries of Central Asia: Kazakhstan, Kyrgyzstan, Tajikistan, and Uzbekistan.
Kind of Data
Sample survey data [ssd]
Data provided to the World Bank by The Brookings Institution on January 31, 2006
The project uses public opinion polling to gather and then analyze a sample that represents the entire population of the country.
Producers and sponsors
The Brookings Institution
The World Bank
For all four Central Asian countries in this survey, the sampling procedure is a three-stage stratified clustered one. Census data on the territorial dispersion of the population is used as the base to start the sampling methodology. The sampling procedure takes the total population of the country, considers geographic units within the country as either urban or rural, and then develops random procedures to select who to survey in three stages: first by randomly selected smaller geographic urban and units in each province (the primary sampling units or PSUs), second randomly chosing households within these units, and third, to randomly select which household member to interview in each household.
The sampling frame used to divide these four countries into smaller geographic units to randomly sample from differs slightly for each Central Asian country, based on differences in data availability on the population of the country and its dispersion. Subsequent sections explain the sampling methodology used and how this sampling frame differs in each country. Then all four countries have PSUs, random selection of households, and random sampling of individuals within households using the same methods.
Tajikistan has 4 provinces, with the city of Dushanbe then considered a separate fifth province. These provinces have 58 districts, with 17 cities and 7 settlements ("posyolok") of provincal submission. Districts incorporate rural settlements or villages, which are incorporated into rural districts ("djamoat dekhot" and "poselkovyi djamoat"). In total there are 23 cities (17 cities of provincal submission and 6 cities of district submission), 47 settlements (7 settlements of provincal submission and 40 settlements of district submission), 356 djamoat and 3,803 villages. The population of Tajikistan was 6,187,561 people, of whom 1,686,095 (27%) were urban, and 4,501,466 (73%) were rural as of January 20, 2000.
Several remote or inaccessible districts were excluded from the sample from since they are practically impossible to get to due to their remote location or absence of transportation. These are three districts in Sogd province, that have a population of 248,290 people, which is 0.1% o f the urban population of the country and 5.5% of the rural population - a total of 4.01% percent of the country.
The sampling frame for Tajikistan is based on the list of small territorial units (primary sampling units - PSUs) of three types:
- Villages - rural settlements subordinate to djamoats, each is a separate PSU.
- Parts of large rural settlements, divided into populations of between 2,504 and 4,835 inhabitants as separate PSUs.
- Parts of large urban settlements, divided into populations of between 2,450 and 4,903 inhabitants as separate PSUs.
- Like Kazakhstan and Kyrgyzstan, the sampling is three-stage stratified clustered sampling for Tajikistan. First, proportionate stratification is done by the population of provinces, with proportionate stratification by urban/rural population within provinces (except the city of Dushanbe which is all urban) and then a PPS-sampling of PSUs within these urban and rural strata. Second, sequential random sampling of households (Secondary Sampling Units - SSUs) is done in selected PSUs. Third, Kish grids are used to sample respondents within households.
For Tajikistan, 56 PSUs are randomly selected from the sampling frame, and between 7 people (for urban areas in Gorno-Badakhshan, which is a tiny proportion of the urban population of the country) and 29 respondent interviewed in each.
The sample distribution of the main demographic characteristics can be compared with census data from 1989 (with data from 2000 used instead in the nationality section). These data have changed substantially over fifteen years and the dramatic change in the economy, society, and polity with the civil war and other changes that have accompanied independence. The data are weighted, which somewhat reduces the typical disproportionate probability of selection of men and youth.
In comparison with the 2000 census nationality data, the number of Uzbeks has grown and the number of people of other nationalities (especially Russians) has appreciably diminished. This is due to high levels of unemployment and increased migration of Tajik men to Russia for work and, on the contrary, the settled way of life of many Uzbeks who have remained in agriculture. Second, census data overestimates the proportion of the titular nationality since belonging to this nation provides advantages in employment, careers, and education. In opinion polls, when no supporting documentation is required, respondents preferred to name their ethnicity as that which they actually identify themselves.
During the fieldwork, 88 cases of nonresponse were observed. The average response rate is about 94% (1,500 of 1,588 cases - due to using the sequential sampling of households the nonresponse had no effect on the final sample size). Generally, nonresponse was registered if a completed interview had not taken place, and an interviewer had made up to 3 callbacks. The response rate was 84.4% in urban areas and 98.9% in rural ones. In Dushanbe the response rate was 73.3%. Two-thirds (67.1%) of urban non-responses came from respondents not being at home; few emphatic refusals to participate were noted in Tajikistan.
According to the interviewers, the main (in the majority of cases) refusal was occupation (work). If in town its inhabitants spend the whole day at work, in village this is caused by cotton gathering season.
Most refusals were due to the households or respondent's straightforward refusals to give an interview. Like in Uzbekistan or in other countries, these refusals are partially generated by insufficient capability of interviewers to persuade household or respondent to agree for conversation. The same goes for other household members' refusal to contact a required respondent. Plain and direct refusals are characteristic of urban population. In the next surveys we are envisaging particularly scrupulous training for interviewers, who will work in towns.
Dates of Data Collection
Data Collection Mode
Field works were completed (latest questionnaires entering) on 9 October 2004. Check was started immediately upon the first questionnaires entering, i.e., since 21 September 2004.
Quality control comprises the following stages:
1. Check of correctness of sampling forms filling in.
2. Check of correctness of respondent's sampling.
3. Check of the correspondence between the number of Kish' card and the number of the sampling form.
4. Check of correctness of the questionnaires filled in.
5. Check of the correspondence between answers and questions in terms of logic.
Final control visits, apart from visits to supervisors, were rendered from 9 October to 17 October 2004.
After field works completion each interviewer, who participated in questioning, was checked. For this purpose a 100 percent check of filled in questionnaires, submitted to the headquarters' was made by supervisors, who worked in other regions and oblasts. If the check's results proved to be positive (the procedure of respondent's or household's sampling was not violated or questionnaire was filled out in pursuance with the interview taken etc.), the check of the given interviewer was considered completed. In the course of checks the following violations have been discovered:
In GBAO 1 case of the respondent's replacement was found out (sampling of the wrong respondent with Kish' card) and the checking person conducted a repeated interview.
In Hatlonskaya oblast 3 cases of the respondent's replacement were found out and in those three households repeated interviews were carried out.
In Dushanbe 4 cases of the respondent's replacement were revealed, and repeated interviews were conducted.
During check 1 case of the interview's falsification was discovered, after wards all interviews of this interviewer were checked through the method of control visits, and afterwards two more falsified interviews were found out. Repeated interviews were performed in these three households.
In Sogdiyskaya oblast 2 cases of the respondent's replacement were found out. Re-questioning was held by interviewers themselves.
After the beginning of field works, no particular problems were encountered in provinces. There were only some difficulties for interviewers to identify certain PSU and to specify the amount of households in them. It looked like the following:
1. In Tursunzade region the name of jamoat "Chapaevskiy" has been renamed into 10-solagii Istiklol.
2. In GBAO it was due to the problem to find a car, which would drive to PSU Udob. The driver agreed to transport only for a double price. During questioning 2 respondents demanded from the interviewer the payment for the time they had spent for an interview.
Data Collection Notes
The questionnaire piloting took place from 6 till 8 September in 60 households, which were selected in 3 communities - one in Dushanbe City, the second one in the village of Somonien and the third one in the village of Zainabobod of Rudaki region. Skilful interviewers and supervisors, who had the minimum of 2-3 years of relevant experience, participated in the questionnaire piloting.
Training for piloting survey was performed on 5 September 2004. During training the survey's goals and tasks were explained, a detailed analysis of the questionnaire was made, and also the working schedule and the schedule of piloting survey conduct were discussed. Training was performed by "Afkor" Centre Director.
On 10 September discussion of the piloting survey results was held with participation of all interviewers. All issues, which revealed respondents' difficulties or misunderstanding, were examined. In accordance with the piloting results, amendments were included both in Russian and Tajik variants of questionnaire.
TRAINING FOR SUPERVISORS AND INTERVIEWERS
In the course of preparation for questioning the following types of training were carried out:
- Principal training for Dushanbe interviewers and supervisors was conducted on 12-13 September 2004.
- Training in oblast centres for local intervieweres was performed depending on the field works commencement in every oblast (from 17 September to 19 September 2004).
(1) Training for supervisors and interviewers.
So that interviewers perceive goals and tasks of the survey at a more qualitative level, training was divided for 2 days, which were held in Dushanbe and conducted by the Client's Representative and Project-Manager from Uzbekistan. Training consisted of the following procedures:
- Explanation of the general survey strategy.
- Explanation of respondents' sampling rules for the survey.
- Explanation of households' sampling rules for the survey according to the sample form.
- Communicative interviewing questions and the technique of the face-to-face interview conduct.
- A question by question analysis of the questionnaire.
- Conduct of trial interviews.
- Ways and methods to establish a contact
- Analysis of field works organisational issues in oblasts.
- Sampling of interviewers.
Practical lessons on performance of sampling at places of questioning, as well such practical lessons as interviews between interviewers were carried out.
(3) Training for interviewers in oblasts (provinces)
Training for interviewers in provinces was carried out by oblast (provincial) supervisors depending on the beginning of field works in every oblast. All together 47 interviewers, who had passed tests, and 5 supervisors participated in the survey. Among them 40 interviewers speak Tajik and Russian and 7 interviewers speak Russian, Tajik and Uzbek.
REGIONAL DISTRIBUTION OF INTERVIEWERS ACCORDING TO THE AMOUNT OF CONDUCTED INTERVIEWS
While distributing the interviews' amount between interviewers, we adhered to the principle that each interviewer should have no more than 30 interviews. Still in some cases this rule had to be broken. It is explained by an inadequate efficiency of some interviewers, and their work had to be given to experienced interviewers, who performed a repeated interview after control visits and checks.
PROBLEMS DURING QUESTIONING
All the interviewers noted in their diaries the respondents' fatigue by the end of interview, which took the respondents so much time. By that time they answered without interest and sometimes responded beside the point. In such cases interviewers made a pause or suggested meeting some other time.
Rather often during interviews relatives of an interviewee expressed their wish to participate in the conversation too. It refers to cases when the respondents were either wife, daughter-in-law or children. Interviewers had to explain in rather a long and detailed way to parents or husband rules and particularities of the given work's conduct. For example, parents of a 18 year old girl wanted to be present by all means during questioning, stating that "What if she says something wrong," that "She is yet too young" etc. The interviewer managed to persuade the parents to leave the room, saying, that they could leave the door ajar and hear all the conversation. The interview was carried out with success.
Other example: The respondent's husband did not want to leave the wife with an interviewer and stayed in the room, but he took seat far from the place of the interview's conduct, promising, that he would not interfere in the interview. The respondent was asking the husband's opinion virtually regarding each question, and then he left himself, realizing, that he would not be capable of keeping his promise. The interviewer began questioning anew and the questionnaire was filled in.
There were instances, when relatives refused to permit to have an interview with a respondent, especially if a respondent was a woman, explaining that she understands and knows nothing. Urban inhabitants manifested interest to know with which purpose and who need this questioning (i.e. which structure) etc. In regions, people often asked to help solving their local or personal problems.
To perform questioning, the following documents have been prepared (attached):
- Questionnaire (in Tajik and Russian languages).
- Sets of cards (in Tajik and Russian languages).
- Forms of the respondent' sampling and of households' visits records with Kish's cards (in Tajik and Russian languages).
- Forms of the households' sampling in selected points of questioning (in Russian language).
- Sampling instructions (in Tajik and Russian languages).
- Instructions on households and respondents' sampling (in Russian and Tajik languages).
- Examples how fill out sampling forms
- Covering letter to local authorities and ID cards for interviewers (in Russian and Tajik languages).
- The data and related materials will not be redistributed or sold to other individuals, institutions, or organizations without the written agreement of the Microdata Library.
- The data will be used for statistical and scientific research purposes only. They will be used solely for reporting of aggregated information, and not for investigation of specific individuals or organizations.
- No attempt will be made to re-identify respondents, and no use will be made of the identity of any person or establishment discovered inadvertently. Any such discovery would immediately be reported to the Microdata Library.
- No attempt will be made to produce links among datasets provided by the Microdata Library, or among data from the Microdata Library and other datasets that could identify individuals or organizations.
- Any books, articles, conference papers, theses, dissertations, reports, or other publications that employ data obtained from the Microdata Library will cite the source of data in accordance with the Citation Requirement provided with the dataset.
- An electronic copy of all reports and publications based on the requested data will be sent to the Microdata Library.
- The original collector of the data, the Microdata Library, and the relevant funding agencies bear no responsibility for use of the data or for interpretations or inferences based upon such uses.
Tajikistan Survey of Conflict Prevention and Cooperation 2004, Ref. KAZ_2004_SCPC_v01_M, dataset downloaded from microdata.worldbank.org on [date]