Factors that influence correspondence between subject and collateral reports of subject substance use among college students

Collateral informants have been incorporated in substance use intervention research involving college students to independently verify subject self-reported substance use. Few studies have identified factors that influence correspondence between subject-collateral (S-C) reports specific to substance use behavior. This study examined the associations between S-C reports of subject substance use and identified factors that influence agreement between both sets of reports. Same sex S-C pairs (N = 300; 74.3% female) were recruited from undergraduate psychology courses. Each was asked to report on his/her own substance use as well as his/her friend's substance use. Data yielded moderate correlations between S-C pairs for 12 substance use variables (e.g., cocaine/hallucinogen use; r's .30 to .73), while discrepancy analyses revealed a tendency for subjects to report greater substance use relative to collateral reports. A regression analysis revealed that greater S-C agreement for frequency of subject drug use occurred when subjects' consumed less alcohol quantity (b = .07, p < .01), collaterals' engaged in less drug use (b = .11, p < .01), and when subjects and collaterals had more common occasions of consuming alcohol (b = -.06, p < .03) or drugs together (b = -.18, p < .01). S-C correspondence regarding drug use in this study suggests that college students provide reasonably accurate substance use data. In studies in which collaterals may be warranted, findings indicate that the ideal collateral for studies involving college students is someone who can observe the subject's drug use behavior, but do not consume drugs on a frequent basis.