This paper explores ‘gatekeeping’, the phenomenon where access to a sample person in the household is controlled by another person. Gatekeeping of female persons is especially an issue in societies governed by gendered social norms. It can increase survey error by reducing response rates and potentially increasing non-response bias, and can increase measurement error when gatekeepers insist on providing proxy responses. We contribute to the sparse literature on gatekeeping by using data from a telephone survey of migrants in India and focusing on the sample of female married-out migrants. We estimate the prevalence of gatekeeping, compare outcomes of calls made to gatekept versus non-gatekept cases, and model the likelihood of a proxy interview. We estimate that gatekeeping of females in telephone surveys is in the range of 56%–65%, with gatekept calls much more likely to result in proxy interviews. We find that older and more-educated women are less subject to gatekeeping and that male and married interviewers are associated with a greater likelihood of gatekeeping. We discuss implications of these results for survey practice.