Can I Interview Her? Gatekeeping in a Telephone Survey of Female Migrants in India.

Sharan Sharma, University of Maryland
Bijay Chouhan, National Council of Applied Economic Research
O.P. Sharma, National Council of Applied Economic Research
Jaya Koti, NCAER
Sonalde Desai, University of Maryland

We study the phenomenon of gatekeeping in telephone surveys, where the selected respondent is accessible only via another person in the sample unit. While gatekeeping is an important phenomenon to study since it impacts non-response and measurement error, little literature examines it. We use a large-scale telephone survey of 46,147 migrants in India and focus on the gatekeeping of female married migrants. We contribute to the literature by estimating the prevalence of gatekeeping (59%). We find that gatekept cases have a lower response rate (30%) than non-gatekept cases (46%). We analyze the determinants of gatekeeping by using multilevel regression models with sample cases clustered by interviewers. In this manner, we can not only study geographic, household, and individual factors associated with gatekeeping but can also study interviewer effects. Since we collect background information on all our interviewers, we can examine interviewer-level variables that explain these effects. We then study the impact of gatekeeping on measurement error by comparing responses from gatekept (and proxy) cases with non-gatekept cases on select variables after adjusting for selection bias. Finally, we make recommendations for survey practice using interviewer observations on gatekeepers and call notes.

Keywords: Data and Methods, Multi-level modeling , Internal Migration and Urbanization

See extended abstract.