| Summary: | Interviewer design effects occur when data collected by the same interviewer is more similar than data collected by different interviewers. Design effects inflate survey variance and reduce the precision of estimates. Using household survey data collected via computer assisted personal interviewing (CAPI) in Sudan this paper employs a two-level mixed effects regression model to identify interviewer design effects for key variables. The study finds mean interviewer design effect values of 7 with a maximum of 16, implying a significant loss of precision. Recommendations to mitigate interviewer design effects include simplifying questions, sound survey implementation practices, and utilizing multi-way cluster robust standard errors to account for both area and interviewer clustering during data analysis
|
|---|