THIRD-PARTY PRESENCE EFFECT WITH PROPENSITY SCOREMATCHING.

This paper attempts to measure the effect of third-party presence during the interview on sensitive questions about women's empowerment in Morocco, a country with Muslim culture and traditions. Data are from the 2003 Morocco Demographics and Health Survey, which includes only female respondents who were asked two series of questions about their participation in the decision-making process at the household level and their attitudes toward wife beating. Like previous studies, we find that the occurrence of the third-party presence during the interview is not random. It is more likely to occur among women with particular characteristics (e.g., less educated, younger, less likely to be employed, and married). We then use propensity score matching to separate the third-party effect from personal characteristics. While we find the thirdparty effect to be statistically significant for some questions, the magnitude of the effect is relatively small in all questions. [ABSTRACT FROM AUTHOR]