Quality of baseline data: the case of PRIME project

Abstract

This study attempted to assess the quality of baseline data of the Primary Initiative in Mainstream Education project of BRAC Education Programme coliccted through census done by locally recruited interviewers. A sample survey of 900 households under census was done. Of these 900 households 100 were re-surveyed. Inter-observer agreement or concordance was used in assessing the quality of data. The measurement of interobserver agreement indicated that there were less than 10% cases in each variable in which a disagreement between observers of sample survey and re-survey was found. This further indicated an acceptable level of the quality of data in sample survey as well as resurvey. However, the study indicated that there were more than I 0% cases even more than 30% cases in each variable in which a disagreement between observers of census and sample survey was evident. This further indicated an unacceptable level of the quality of data in census. Thus, the study raises a question about the quality of data in census. Any interpretation of the census data with this unacceptable level of quality might be misleading. It is therefore, suggested that one should be careful in using the census data for a more realistic interpretation otherwise it would give misleading picture of the situation which further would mislead the intervention. The study also suggests, in any census in future it is to be ensured that proper, adequate and sufficient training is provided to interviewers with a well-explained manual. It further suggests that proper and adequate monitoring/supervision by the supervisors has to be done to ensure the quality of data and that the data are properly collected.