Psychological Mechanisms Associated with the Accuracy of Drug Use Reporting: Social Interest

Little research is currently available that evaluates psychological mechanisms that may be associated with the accuracy of drug use reporting in epidemiologic surveys. Given that underreporting of drug use continues to be a major concern, epidemiologists should explore measures associated with this behavior so that they can be adjusted for in etiological modeling. In this study, we investigate the relationship between the construct of social interest and the biologically validated reporting of recent cocaine use in a community sample. Social interest emphasizes affirmative attitudes towards humanity in general and emphatic understanding of individuals in particular (Crandall, 1991). We hypothesized that increasing social interest would be correlated with greater agreement between drug use reports and biological assays. The sample for this study consists of 477 adults aged 18-40 residing in Chicago, Illinois who were interviewed using audio computer-assisted self-interviews (ACASI) in 2001-02. In contrast to our a priori hypothesis, the findings revealed that persons under reporting recent cocaine use scored higher on a measure of social interest (mean = 11.25, SE = 0.24) than did persons not under-reporting (mean = 8.90, SE = 0.33; F (1,32) = 50.5, p < .0000). Logistic regression analyses confirmed these differences (OR = 1.15, 95% CI = 1.00 - 1.32) when adjusting for sociodemographic variables, including age, gender, race/ethnicity and education. Implications of these findings for the conduct of epidemiologic surveys and for models of the survey response process will be discussed.