Internet and telephonic IVR mixed-mode survey for longitudinal studies: choice, retention, and data equivalency.

PURPOSE This study examined data equivalency and loss to follow-up rates from Internet and interactive voice response (IVR) system surveys in a prospective-cohort study. METHODS 475 limited-service restaurant workers participating in the 12-week study were given a choice to report their weekly slipping experience by either IVR or Internet. Demographic differences, loss to follow-up, self-reported rates of slipping, and selection of first and last choices were compared. RESULTS Loss to follow-up rates were slightly higher for those choosing the IVR mode. Rates of slipping and selection of first and last choices were not significantly different between survey modes. Propensity to choose an Internet survey decreased with increasing age, and was the lowest among Spanish speakers (5%) and those with less than a high school education (14%). CONCLUSIONS Studies relying solely on Internet-based data collection may lead to selective exclusion of certain populations. Findings suggest that Internet and IVR may be combined as survey modalities within longitudinal studies.