CAIConsulting has offices – in Bishkek (Researches’ Department consists from 9 people- core staff) and Osh (3 people – core staff) and representatives with interviewers’ network in 6 areas of Kyrgyzstan (6 local supervisors and 23 interviewers) and regional offices in Tajikistan (Dushanbe), Uzbekistan (Tashkent) and representatives in Kazakhstan (Astana), Turkmenistan (Ashgabat). Supervisors are mainly people over 30 years old that are working and regularly paid more than 10 years on permanent basis under some long-term projects. The majority of interviewers are working at well-established working relationships and paid per interview. Interviews are conducted in any languages: Russian, Kyrgyz, Uzbek, Tajik and Turkmen.
Trainings are organized for specific projects – in situ or by gathering people in CAICo’s office in capital of country, sometimes with participation of the client. Trainings are organized once in 2-3 years, however, mainly the work is done by telephone – through detailed instructions.
Pilot testing is obligatory for each project without any exclusion. It is done by the core staff to adapt questionnaire, interviewing procedure, to understand timing and sampling problems. Very often pilot testing is done twice. The main principle of CAICo’s activity is to meet client’s needs. CAIConsulting offers high quality researches in social, socio-economic and political spheres. For each specific project, CAIConsulting’s specialists choose an optimal combination of qualitative and quantitative methods to assist clients in their work. There are various methods in CAIConsulting's working repertoire. Quantitative ones include mail, face-to-face, telephone and panel interviews. The interviewers’ network allows covering more 3,500 respondents all over country. Qualitative methods are presented by focus group discussion, educational methods, expert evaluation, participatory development approach (PRA) etc.
Sampling is constructed with use of official statistics (for example, population size by regions and localities, a number of enterprises in various sectors, proportions of organizations, etc.). Usually we 6 use systematic probability sampling with randomized starting point (taking then every kth respondent), stratified sampling, cluster sampling, quota sampling, etc. Mechanical sampling, such as snow-ball is often used to choose people for focus-group discussions. The sample size is defined according to formulas or with use of statistical calculator. CAIConsulting is taking care of rural samples, which is always under- and mis-respresented in surveys, because many companies survey just closest suburb areas. That is why our findings are more reliable and accurate.
Translation of the questionnaire. Questionnaires are translated into Russian, Kyrgyz, Uzbek, Tajik and Turkmen languages if it is needed. Although we do not do back translation into original languages, all versions are checked during pilot testing for better question formulation.
Quality control. Field works are conducted according to CAIConsulting’s internal instructions standards based on the “Researcher’s Code”. There is a system of permanent (or selective) control of the fieldwork, data entering and sampling control. Quality control is done by regional supervisors as well as central office staff at the different stage of the project implementation. Statistical analysis used is quite comprehensive, starting with simple descriptive statistics, such as analysis of variances and ending with regression and cluster analysis. The main tool is SPSS. Quality is maintained through the long-term working with supervisors (regional coordinators), checking of every 5th interview by supervisor, random checking of interviews by field supervisor by telephone (indicated in the filled questionnaire). The special procedure is done to prevent mistyping during data input, when every 5th questionnaire is checked by data input specialist: in case of mistakes all questionnaires are to be re-input. The double data entering is used under client’s special request. Actually, the company is trying to meet all expectations and requests coming from the client, in some disputable cases the procedure is discussed to solve problems.
Data entry. CAIConsulting is using SPSS, package for statistical processing. Data analysis includes descriptive statistics, frequencies, correlations, factor and regression analysis. Answers for open-ended questions are grouped into 5-7 categories groups; non-classified answers fall into category “other”. All categories are easily translated into English.