Relevance & Research Question: Interview duration is a central variable in Web survey design. For example, previous research shows that survey breakoff rates increase the longer a Web survey takes to be completed. Given the crucial role of duration, this paper addresses the question of which factors determine the interview duration in Web surveys.
Several factors on different conceptual levels contribute to interview length. First, on the level of survey and questionnaire design, overall length of the questionnaire, i.e. the number of questions, use of graphical elements, complex designs, and software choice contribute to this central characteristic. Second, different types of questions, e.g. open-ended questions, closed questions, questions arranged in grids etc., influence the time a respondent needs to respond. Finally, on the level of the individual, personal characteristics of the respondents have an impact on response speed. For example, previous research shows that older and lower educated respondents answer questions at lower speed. Also, the respondent’s motivation might affect response speed.
Methods & Data: In this paper we apply a broad methodological approach. First, we bring together the levels of survey design and individual explanations of interview duration. Second, we use data from 18 Web surveys on political attitudes and behavior of German Internet users conducted between 2009 and 2012. For each of these surveys we collected information on interview duration, survey and questionnaire design as well as on question and respondent characteristics. Data structure and theoretical approach require application of multilevel modeling.
Results: Preliminary analyses show that all levels explain parts of the variance in interview duration but that not all levels are of equal importance. Particularly, duration is influenced by questionnaire design as well as question and respondent characteristics. Among the latter, age and education stand out. Also, types of questions and their frequency affect interview duration.
Added Value: Enhanced knowledge of the factors explaining interview duration allows us to make informed design decisions and to further improve the quality of Web surveys. The multilevel method enables us to fit all explanatory variables in a single model and discuss respective results under control of the other parameters.