English  Français

Social network-based abortion incidence estimation in Burkina Faso: Examining the impact of the network generating question

Suzanne Bell, Johns Hopkins Bloomberg School of Public Health
Georges Guiella, Institut Supérieur des Sciences de La Population de L’Université Joseph Ki-Zerbo
Selena Anjur-Dietrich, Johns Hopkins School Of Public Health
Elizabeth Nansubuga, Makerere University

Monitoring abortion rates is highly relevant for demographic and public health considerations, yet its estimation is fraught with uncertainty. Social network-based methods for estimating abortion incidence have shown promise in recent years, but their performance has been inconsistent across approach and context and the specific network generating question used. In this study we seek to examine the performance of social network-based reporting for estimating abortion incidence when using a network generating question that specifies mutual sharing of sensitive information (i.e. a “confidante” relationship) versus simply the closest female friend or relative (i.e., a “best friend” relationship). Specifically, we explore differences in the following social network-based reporting assumptions by network generating question: 1) full transmission of abortion-related information, 2) representativeness of the surrogate population, and; 3) reduced social desirability bias. We used data from a population-based survey of reproductive age women in Burkina Faso (n=6,709). Preliminary results suggest transmission bias is significantly reduced for the reporting of friends’/relatives’ abortions compared to their own, but that the surrogate sample differs from the respondent sample depending on the network generating used and that respondents appear more likely to share their abortion experience with their closest friend/relative compared to their closest confidante.

Keywords: Fertility and childbirth, Demographic and social surveys, Methodology, Social network methods

See extended abstract.

  Presented in Session P3.