Content area
Full Text
Abstract
Researchers have increasingly adopted a web survey for data collection. Previous studies have examined factors leading to a web survey's success. However, virtually no empirical work has examined the effects of the three levels of forced responses or the two styles of question items displayed on a web survey's response rate. The current study attempted to fill this void. Using a quasi-experiment approach, we obtained 778 unique responses to six comparable web questionnaires of identical content. The analysis confirmed that (1) there were statistically significant differences across the surveys with the 100%-, 50%- and 0%-forced responses, and (2) there is not a significant difference between the response rates between surveys with scrolling and those with paging styles. In addition to extending the theoretical insight into factors contributing to a web survey's response rate, the findings have offered recommendations to enhance the response rate in a web survey project.
Key words: Forced responses, Question display styles, Scrolling, Paging, Web survey response rate
JEL code: C80, C83, M15
Introduction
Online questionnaires are tools that social science researchers including those in business and management fields have adopted to gather data from samples through major web browsers (Lee & Yun, 2015). The increasing number of publications has addressed issues on how to implement a web survey using online questionnaires.
One of the issues is the execution of online questionnaire survey to increase a response rate. This is because a project in which a questionnaire is delivered either by mail or online is often known as having a low rate of response.
The rate is sometimes so low that the validity of the project might be in question.
Online questionnaires do have certain limitations and advantages (Reips, 2002, 2007). An online survey has two limitations that researchers must prepare a proper plan for in order to minimize prior to starting the data collection. First, an online survey always reaches only Internet users. If the project's target population taps those whose profiles are not largely shared with Internet user profiles, researchers may have to give up the online version. Second, given the nature of the Internet, the samples' responses may be different from those taking part in traditional paper-based questionnaires. Such responses include those from the same subject...