Induced Abortion in Rajasthan, India: Comparing Prevalence Estimates from Two Quantitative Methodologies

Batya Elul, Population Council

As with many sensitive issues, the measurement of abortion prevalence is fraught with methodological constraints, high levels of underreporting and great variability in estimates for a given population. In the absence of accurate abortion statistics, estimates have typically come from three sources: facility-based methods, self-reported survey methods, and indirect estimates. Each of these methods of estimation, however, has significant shortcomings. In this paper, we test an experimental approach of measuring abortion prevalence in Rajasthan, India in which 3266 women sampled at the community level reported on abortions occurring among up to five unnamed women in their social network, thus potentially increasing the survey sample size, as well as women's willingness to report abortions. As India lacks accurate abortion statistics to gauge the accuracy of estimates from this experimental method of measurement, referred to as anonymous third party reporting (ATPR), we compare these with those calculated using self-reported survey (SRS) data.