The effects of survey modality on adolescents' responses to alcohol use items

Abstract

Background: We examined differences in response to self-reported alcohol use items by survey mode, whether self-report differences were the result of modality effects or self-selection, and whether these differences varied across the treatment and control arms of a preventative intervention trial. Methods: Data from an existing alcohol prevention trial were used to estimate the effect of survey modality on adolescent's self-reported alcohol use at ages 17 to 18. Estimates were derived from regression models controlling for self-reported alcohol use during 8th grade, measured using a single survey modality, as well as time invariant selection factors. Results: No statistically significant survey modality effects were found. No differential effects of survey modality were observed by assigned intervention group. Conclusions: We provide initial evidence that adolescent alcohol prevention trials may use multiple survey modalities when necessary to increase response rates without harming interpretation of intervention effects.