Evaluating the list experiment method for reproductive health behavior in six U.S. states

The main objective of this study is to develop and evaluate the list experiment method to estimate the fraction of reproductive-age women who have experienced an abortion over their lifetimes. Abortion is a major component of reproductive health and autonomy, but is increasingly subject to access restrictions that will dramatically reduce utilization. Beyond general aggregate statistics, little is known about the characteristics, attitudes, and behaviors of those obtaining abortions at the state and national levels. A lack of validity of direct survey questions is a critical barrier to understanding attitudes and characteristics of abortion patients. Indirect survey question methods have the potential to overcome this barrier. In this study, we evaluate a form of the list experiment method, the Item Count Technique (ICT), leveraging data from six states with more and less restrictive abortion policies to estimate cumulative lifetime abortion incidence in a period (up to 2019) when abortion was legal across the United States. The six included states are located in the Northeast, Midwest, and South. We first evaluate whether ICT randomization and response assumptions are met. Next, we estimate lifetime abortion incidence overall and by sociodemographic characteristics and individuals’ attitudes to abortion. We conclude by evaluating list experiment estimates against the best available external counts from state health departments, the CDC’s annual compilations by state of residence, and the Guttmacher Institute’s periodic census of abortion providers. Methodologically, our study will contribute to existing literature by evaluating the validity of the ICT method for abortion to a broad social-geographic range U.S. states.