Local or Outsider Interviewer? An Experimental Evaluation

Mariano Sana, Vanderbilt University
Guy Stecklov, Hebrew University of Jerusalem
Alexander Weinreb, University of Texas at Austin

Matching of interviewers and respondents along demographic characteristics is a popular practice in survey research, further enhanced in survey endeavors where interviewers are recruited from the community being studied. The effects of employing local instead of outsider interviewers have never been experimentally evaluated. We report results from a survey experiment that we conducted in the Dominican Republic to measure these effects. Respondents were significantly more likely to report contraceptive use to outsiders, who also recorded significantly higher amounts of remittances received by the household as well as significantly higher outflows (monetary and in-kind) from the household. Respondents were also more likely to tell outsider interviewers that they had heard of presumably famous, but fictitious, people, and presented themselves as more tolerant of stigmatized groups when interviewed by outsiders. We explore possible mechanisms leading to these differences, and believe that misreporting affects both interviewer types, depending on the type of question.