Initial nonresponse and survey response mode biases in survey research.

Abstract

We evaluated survey response factors (particularly initial nonresponse and survey mode) that may be associated with bias in survey research. We examined prevention-related beliefs and outcomes for initial mail survey responders (n=209), follow-up mail survey responders (n=78), and follow-up telephone survey responders (n=74). The Pearson chi-square test and analysis of variance identified beliefs and behavioral outcomes associated with survey response mode. Follow-up options to the initial mail survey improved response rates (22.0-38.0 percent). Initial mail survey responders more strongly believed topical fluoride protects teeth from cavities than others (P=0.04). A significantly larger proportion of parents completing a follow-up telephone survey (30.8 percent) refused topical fluoride for their child than those completing mail surveys (10.3-10.4 percent) (P<0.0001). Multiple mode surveys with follow-up improve response rates. Initial nonresponse and survey response mode may be associated with biases in survey research.