The USAID Promoting Governance, Accountability, Transparency and Integrity program (PROGATI) has been working to reduce corruption in Bangladesh since 2007. In 2008 PROGATI funded a national-level survey of citizens’ perceptions of corruption across various sectors. A mid-term survey was conducted in 2011 and an endline survey in 2012, the findings of which are detailed in this report, including comparison with 2008 and 2011 findings. PROGATI contracted Bangladeshi research firm Mitra and Associates to implement the survey.

Mitra collected quantitative data from 3510 households using the same questionnaire used in the 2008 and 2011 surveys. Qualitative data from focus group discussions (FGD) and key informant interviews (KII) provided additional information on citizens’ perceptions. 

Methodology

Both quantitative and qualitative data were collected in the survey, using Household Interviews (HIs), Focus Group Discussions (FGDs) and Key Informant interviews (KIIs). Household Interviews (HIs) were conducted with respondents from a nationally representative sample of 3510 households. In each household, one respondent over the age of 17was interviewed. The sample was drawn in 12 strata, stratifying the country by rural/urban areas according to six divisions. Four FGD sessions were conducted in every division—two sessions with women and two with men – for a total of 24 FGDs, 12 female and 12 male. Key informant interviews with the local influential persons were conducted in each of the 24 areas where FGD sessions were conducted. In total, 120 key informants were interviewed in the survey, including at least 5 from an area.  Data collection for the endline survey was carried out from 1 May 2012 to 24 May 2012. 

The 2011 National Survey of Bangladesh Citizens for Promoting Democratic Institutions and Practices (PRODIP) Program 

The purpose of the survey was to probe:

• citizens’ understanding of democracy, 
• their perceptions on the role and performance of Member of Parliaments (MP) and parliament, 
• their engagement with MPs and the flow of information between citizens and parliament. 

The result of the survey was used in informing the PRODIP project’s activities, strategies, and areas of opportunities. The result was also used to inform ongoing public information outreach and advocacy.

The survey was carried out with a nationally representative sample of Bangladesh population age 18 or older constituting the survey universe. The sample was drawn in two stages, with the households (basic sampling units) being selected in the second or final stage. The first-stage sampling unit was a cluster of 200-300 households. A cluster was formed with one or more Mauzas or part of a Mauza in rural areas. In urban areas it was made up of one or more Mohallahs or part of a Mohallah. Following the sampling strategies used in the Bangladesh Demographic and Health Survey (BDHS), 30 households was included into the sample from a cluster. More than 30 households were assumed to result in intolerable intra-cluster correlations. Thus, the sample was made up of 180 clusters with 146 from rural areas and 34 from urban areas, including 57 from the division of Dhaka, 31 from Chittagong, 25 from Rajshahi, 23 from Khulna, 22 from Rangpur, 12 from Barisal and 10 from Sylhet. The clusters were selected independently for each stratum, with PPES (Probability Proportional to Size) method of selection. The sampling frame for selection of the clusters for a stratum was created using the list of Mauzas/Villages by households, given in the 2001 Census Community Series Report for the districts making up the stratum.  

In addition to the national sample, a special sample was separately drawn from the coastal, haor and hill regions of the country. The data from the special sample was analysed separately, without combining them with the data of the national sample. The list of upazilas constituting a region was supplied by The Asia Foundation. From the listed upazilas, the special sample was drawn in three stages. At the first stage, two upazilas will be randomly picked up in each region from among the upazilas not included in the national sample. At the second, seven clusters were randomly selected from the selected upazilas in a region, selecting 4 from their rural areas and 3 from their urban areas.  At the third stage, the households, the basic sampling units, were selected drawing 30 households from a selected cluster. Thus, the special sample was made up of 630 households, including 210 households (120 rural and 90 urban) from each region.

Finally, the national sample and the special sample, together produced a total size of 6030 (=5400 + 630) households drawn from 201 (=180 + 21) clusters. 

The tasks were carried out including:

• Development of sample design (including determination of sample size, 
• procedures of selecting clusters, 
• procedures of selecting households, 
• procedures of selecting respondents within households); 
• Formatting/Modification of English questionnaires; 
• Translation of questionnaires into Bangla; 
• Pre-testing and finalization of questionnaires; 
• Recruitment of enumerators; 
• Orientation/training of enumerators; 
• Data collection (administering of survey); 
• Data processing (data entry, cleaning and compilation); 
• Tabulation as per the tabulation plan supplied by the client, 
• the single tables were generated on all variables of data sets for checking range and consistency of each variable as well as quality control of data sets; Data was analyzed developing statistical estimates such as percentages, proportions, means, medians, and using such statistical tools as frequency distributions, cross tabulations, etc. and submission of clean data sets.

The survey was awarded by PRAGATI and funded by USAID