Notice: the WebSM website has not been updated since the beginning of 2018.

Web Survey Bibliography

Title Nonresponse problem in surveys of online communities: An experiment
Year 2014
Access date 10.12.2014
Abstract

Purpose of the study: While surveying online communities of different types for quantitative insights has become commonplace, very little is known about response inducing techniques that would secure higher unit response rates. This study aimed to investigate whether and how the design of email invitation letters could improve response rates and reduce unit non-response bias. Design/methodology/approach: Using a full-factorial between groups design an experiment with three manipulations in the content of email invitations, related to authority of the sender, plea for help, and sense of community, was implemented in a list-based web survey of a randomly selected sample of 2500 registered members of the largest online heath community in Slovenia. The aggregated response rates were analyzed with a series of bivariate comparisons and a multivariate binary logistic regression model. Findings: The overall aggregate response rate for all experimental groups was 8.4%. Invitations with more elements yielded a marginally significantly higher response with the largest difference between invitations with two (9.8%) and without any (5.4%) elements. Bivariate analysis showed that authority and plea for help had a positive effect on response, whilst sense of community did not have any effect. The logistic regression confirmed the positive main effect of plea for help. In addition, a significant interaction between sense of community and plea for help was unveiled: in cases where both elements were included in an invitation a lower unit response rate was more likely to be expected. Further, evidence on non-response bias was observed with more active online community users responding earlier. Originality/value: This study is among the first to empirically examine and quantify the impact of the content of email invitations on unit response rates in list-based web surveys in online communities. Using an experimental design it shows that an informed selection of response inducing elements in email invitations could significantly reduce the unit non-response. It also suggests that
the effect magnitude of different elements in invitations can vary considerably. Research limitations/implications: This study examined three elements of email invitations as experimental factors, even though there are probably many other invitation elements that interfere with this phenomenon. The findings apply only to list-based web surveys, meaning that important (sub)groups (e.g., lurkers) are excluded from the analysis. Practical implications: An appropriate and theory-informed use of response inducing elements in email invitations could importantly help online community managers and scholars to obtain more participation in web surveys of online community members. As such techniques are not cost-intensive they could be further optimized/tailored in the future to account for the specifics of different types of online communities.

Year of publication2014
Bibliographic typeConferences, workshops, tutorials, presentations
Print

Web survey bibliography - Slovenia (48)