Effect of ACASI on Reporting of Abortion and Other Pregnancy Outcomes in the US National Survey of Family Growth.

Laura Lindberg,Rachel H Scott

doi:10.1111/sifp.12068

Abstract

Abortion is a behavior that is stigmatized and difficult to measure. To improve reporting of abortion and other sensitive behaviors in the United States, the National Survey of Family Growth (NSFG) supplements the interviewer administered face‐to‐face (FTF) interview with audio computer‐assisted self‐interviewing (ACASI). This paper estimates differential reporting of abortion and other pregnancy outcomes (miscarriage, live birth) in the NSFG (2002, 2006–2010, 2011–2015) between women's ACASI and FTF interviews. Examining reporting of less stigmatized pregnancy outcomes can help understand the relative contributions of stigma and survey‐level factors in reporting of abortions. More women reported abortions, miscarriages and births in the ACASI than the FTF interview. Differences in reporting were moderated by the length of recall. The ACASI elicited relatively more reporting of abortions and miscarriages among non‐white and low‐income women. Reporting ratios increased over time. ACASI is a tool that may work differently across time, for different measures, and with varying survey contexts.

Highlights

Demographic research is constrained by the challenges of measuring sensitive behaviors, such as abortion, in surveys
We identified the number of respondents reporting each pregnancy outcome occurring in the same time period in the audio computer-assisted selfinterviewing (ACASI) and the FTF interview, and calculate a ratio of these reports
Our analysis focused on the effect of ACASI in the National Survey of Family Growth (NSFG)

Summary

INTRODUCTION

Demographic research is constrained by the challenges of measuring sensitive behaviors, such as abortion, in surveys. 15 percent more women reported ever having an abortion in the ACASI than the FTF interview in unweighted analyses of the 2002 NSFG (Tourangeau, Rips, and Rasinski 2000) and in the 1995 survey, the ACASI resulted in 29 percent more women reporting an abortion than in the FTF interview (Peytchev 2012) These increases in reporting in ACASI have been interpreted as more complete and accurate than the FTF interview measures, they have still been incomplete as compared to external surveillance counts collected directly from abortion providers Jones and Kost 2007) In this analysis, we focus on patterns of abortion reporting by survey mode in the NSFG to directly test the assumption that ACASI improves reports of abortion by increasing. These findings may be relevant to the survey measurement of other sensitive behaviors such as sexual behavior, illicit drug use, or violence (Kalsbeek and Carol 2014; C. Kelly et al 2014; Turner et al 1998)

BACKGROUND

RESULTS

DISCUSSION

Strengths and Limitations