Impact of mixed survey modes on physical activity and fruit/vegetable consumption: A longitudinal study

It is recommended that researchers who use mixed modal methods for data collection compare their impact on outcome measures. The purpose of this study was to examine the physical activity and fruit/vegetable consumption behaviors of a multiethnic sample of adults, comparing participants who continued a telephone survey and those who transitioned from a telephone to a web-based survey for a follow-up data collection point. This longitudinal study used a random sample of 700 Hawaii residents (63.3% Female; Mean age=47, SD=17.1). At baseline, participants completed a computer-assisted telephone interview assessing the stage, behavior, and decisional balance of both physical activity and fruit/vegetable consumption. For the three-month follow-up survey, participants were given the option of completing the survey either on the web or by phone. Repeated measures analysis of variance (ANOVA) was computed for related physical activity scales and fruit/vegetable consumption variables to compare the change in response across time between a web group and phone group. For both physical activity and fruit/vegetable consumption, all mode-by-time interactions were significant. The participants who preferred the telephone survey maintained their levels, whereas those who preferred the web survey reported a decrease in each variable. These results suggest that changing the mode of a survey may introduce a systematic bias in data and that researchers should proceed with caution when using mixed modes of data collection.