Three Approaches to Improving the Use of Face-to-Face Interviews to Measure Abortion

Guttmacher Institute/IUSSP their responses. The researchers found that many respondents had interpreted questions differently from the intended meaning and that nonsampling error was much greater than standard sampling error. They concluded that using qualitative research techniques to supplement and complement quantitative techniques can improve data validity and reduce nonsampling error. In separate studies Anderson and coauthors (1994) and Huntington and coauthors (1993) reported success in collecting abortion data with a filter question. In Anderson’s study, women were asked if they had ever had a pregnancy that did not result in a live birth. In Huntington’s, respondents were asked if they had ever had an unwanted pregnancy. Respondents who answered affirmatively to the filter question were then asked what they did when faced with the pregnancy. The filter question in those surveys was meant to reduce the stigma associated with induced abortion and to make reporting easier for the respondent. In a later study, Huntington and colleagues (1996) recognized the difficulty of obtaining precise estimates of induced abortion from direct survey questions and evaluated a survey technique for measuring induced abortion within the context of unwanted pregnancy. The technique was incorporated into multiple Demographic and Health Surveys (DHS). Results from countries in Africa, Asia and South America indicated that “no single, universally sensitive context for discussion of induced abortion may exist.” Huntington and coauthors concluded that methods to successfully estimate rates of induced abortion varied by setting. They recommended conducting qualitative background research to find a way to introduce abortion questions in a relatively nonthreatening, nonstigmatizing context. This chapter presents three examples of attempts to modify face-to-face surveys in three settings— Bangladesh, India and Eastern Europe. These modifications to improve the reliability of face-to-face interviews in measuring abortion include the following: