Potential Interviewer Bias in the 2005–06 National Family Health Survey in India

Abhishek Singh , International Institute for Population Sciences (IIPS)
Fred Arnold, ICF International
Ankita Shukla, International Institute for Population Sciences (IIPS)
Kaushalendra Kumar, International Institute for Population Sciences (IIPS)

There has been a lack of studies that have systematically assessed the level of interviewer bias in large-scale household surveys. The objective of the paper is to examine possible interviewer bias in the 2005-06 National Family Health Survey (NFHS-3) in India. The highest and lowest scorers are more likely to differ on sensitive questions than on non-sensitive questions in a one-way ANOVA and binary logistic regression. Of the 100 teams, the highest and lowest scorers are likely to differ in 87 teams on wife beating justified score. In comparison, the highest and lowest scorers are likely to differ in only 14 teams on average education. The north, south and west regions have the most disparity among interviewers within interviewing teams. Our analysis indicates the likelihood of interviewer bias related to some sensitive questions in NFHS-3, but there is no evidence of this type of interviewer bias for non-sensitive questions.

See paper

 Presented in Session 3. Population, Development, & the Environment; Data & Methods; Applied Demography