An Investigation of the Effects of Social Desirability on the Validity of Self-Reports of Cancer Screening Behaviors

Abstract

We evaluated 4 approaches to improving the reporting of disease prevention and screening behaviors. The conditions evaluated include (1) the mode in which data are collected, (2) asking the interview subject about her intention to obtain the procedure before asking whether the behavior occurred, (3) asking the interview subject about barriers that might keep respondents from getting the procedure before asking about whether she has received it, and (4) asking the interview subject about exceptions to the regularity with which she might report getting the examination. Data were collected in 2001 from a sample of women aged 50 and older in Champaign-Urbana, Illinois. After completing a telephone or audio computer-assisted self-interview (ACASI), respondents gave permission to abstract their medical records to validate self-reports of Papanicolaou tests, mammograms, and clinical-gynecologic examinations received during the past 3 years. Interviews and matching medical records were available for 588 respondents. Results indicated that first asking about future intentions may be an important design feature that warrants additional consideration. In addition, the use of ACASI may lead to lower quality reporting among women with little computer experience. This study represents the only research to date that reports experimental attempts to address the social desirability biases commonly found in the reporting of cancer screening behaviors.