Interviewer Effects in the Elicitation of Sexual and Drug Injection Partners

Abstract

Interviewer effects, or variations in interviewee responses associated with interviewers, are not uncommon in survey research. Such effects are more likely when interview questions are open-ended. Recent research shows moderate interviewer effects in the number of personal network members elicited, with intraclass correlations ranging between 0.13 and 0.15 after adjustment for respondent and interviewer characteristics. It is crucial that network elicitation be as complete as possible, because most network measures are sensitive to missing data. Sexual and drug injection networks shape the transmission of sexually transmitted diseases, HIV, and other infectious agents. The elicitation of sexual and drug injection partners involves asking sensitive open-ended questions, which may increase the likelihood of interviewer effects. Therefore, we assessed the magnitude of interviewer effects in the elicitation of such partners.

The results from five data sets suggest no or relatively small interviewer effects on the number of partners elicited (intraclass correlations = 0.06, median = -0.01). We found similar results when we adjusted for numerous covariates and when we examined interviewer effects on the number of partners that interviewees estimated (prior to listing partners individually) they had. Moreover, there is no consistent interaction between interviewer and interviewee sex on the number of sexual partners elicited, indicating that, on average, male and female interviewers are equally effective with interviewees of either sex.

Keywords: social networks, measurement, interviewing techniques, survey methods, infectious disease

JEL Classification: C42, C90, I18