The frequency divide: implications for internet-based surveys.

Abstract

Those who use the internet more frequently are more likely to notice a request to participate in a survey than less frequent users. The frequency of internet use is thus likely to affect the likelihood of participation in internet-based surveys. If frequent and infrequent users are different in relevant features, this could influence survey estimates. This study aims to identify which demographic characteristics most differentiate frequent and infrequent users of the internet and whether those distinctions have an influence on substantive responses. The effect of internet usage frequency when conducting internet-based surveys on specific subgroups of the population is also examined. Results suggest that frequent and infrequent users are different both in demographic characteristics and substantive estimates. Differences in substantive estimates are also found when comparing frequent and infrequent users in the 15-24 years subgroup. Weighting can reduce the discrepancies found for most of the substantive estimates, but the differences between frequent and infrequent users remain statistically significant for some specific items.