Predictors of inconsistent responding in web surveys

Abstract

Purpose – The purpose of this paper is to investigate the antecedents of inconsistent responding in web surveys. Consistency of responses to personal information questions and scale items were compared, and the influence of perceived social support, social appearance anxiety, academic self-efficacy and social networking habits on inconsistent responding was examined. Design/methodology/approach – A gaming application on Facebook was used to collect data. A repeated-measures design was conducted with 806 respondents in two online survey administrations. After inconsistent responses provided by the same nicknames were identified, consistent and inconsistent respondents were compared with regard to their responding patterns and research variables. Findings – Findings revealed that 45.7 percent of participants misreported their personal information such as age, educational status and gender. Participants were relatively consistent in their responses to attitude scales. Perceived social support led to inconsistent responding whereas social appearance anxiety and academic self-efficacy was not influential in response patterns. A binary logistic regression revealed that perceived family support, number of Facebook friends and Facebook use duration successfully distinguished inconsistent respondents from consistent respondents. Research limitations/implications – The sample frame has several limitations insofar as the study only addressed a unique gaming application on Facebook. Thus, unique interactive characteristics of the current context may have altered the nature of responding. Practical implications – Practitioners should not rely on the personal information provided by online survey respondents to conduct parametric tests, whereas responses to online attitude scales seemed relatively consistent. Originality/value – The principal contribution of the paper is that findings have provided insights into the current status of response patterns in online survey administrations. In addition, the paper highlights the importance of individual variables which influence the consistency of responses.