Comparison of Survey and Interview Methods for Collecting Recent Caffeine and Alcohol Use Data in College Students

Abstract

Background: The present study collected benchmark data on alternative methods for collecting quantity and frequency information about caffeine, alcohol, and caffeine and alcohol combined (CAFF+ALC) use. Materials and Methods: Participants were N = 50 college students who completed computer-administered survey (COMPSURVEY) and Timeline Followback (TLFB) interviews in randomized order. COMPSURVEY and TLFB data were compared using t-tests, McNemar's tests, and Spearman's correlations. Results: For alcohol, COMPSURVEY underestimated quantity consumed compared to TLFB. Similar patterns were found for caffeine, with every participant having ≥1 COMPSURVEY-TLFB self-report difference and 65.5% having a CAFF+ALC inconsistency. Over half (56%) of participants had ≥1 caffeine misreport, with 46% misclassifying a beverage as containing/not containing caffeine. Moderate to strong relationships were found for caffeine and alcohol quantity and frequency of use, but associations for CAFF+ALC were weak. Conclusions: The largest inconsistencies were found for CAFF+ALC use, indicating that more research is needed to identify methods for collecting reliable and valid CAFF+ALC use data.