Web Survey Bibliography

Title Targeting the bias – the impact of mass media attention on sample composition and representativeness of volunteer web-surveys
Year 2014
Access date 11.06.2014
Abstract

Relevance/Research Question: The increasing use of volunteer web-surveys has often been criticized as their results are not representative for the population of interest. While the discussion has mainly focused on different weighting techniques to correct for selection biases, this study addresses the question whether the bias of such samples might be corrected through specific mass media attention. Scientific literature, so far, has not addressed the different recruitment strategies of volunteer web-surveys in more detail. The knowledge of how to reach a large and heterogeneous group via web-marketing has, however, increased tremendously. Our study aims to explore whether selection bias in a volunteer web-survey might be tackled through media attention. Three sub-questions are addressed: How long does the effect of media attention last on average? Does the sample composition change in the periods of massive media attention? Which sample characteristics are subject to changes?
Methods/Data: The 2009-2013 data from the continuous, volunteer German WageIndicator web-survey (Lohnspiegel) will be used. The Lohnspiegel website has a ‘Salary Checker‘ where web visitors can calculate their monthly wages, based on the survey data. Since 2009, the overregional newspaper ‘Bild Zeitung’ has published every six months on its first page the top-hundred highest earning occupations. First, we analyse the change in the number of web visitors per day (prior and posterior to the event) via Google analytics. Second, we divide the sample between the period of increased and regular web visits and compare the two samples for socio-demographic and work-related characteristics.
Results: The preliminary findings show that in periods of mass media attention the number of visitors doubles and the number of questionnaire submissions increases fourfold. However, for the composition of the sample hardly any effects can be observed. This implies that groups so far underrepresented in the web-sample are hardly motivated to take part in a survey through mass media events.
Added Value: This topic has not been addressed in the scientific discussion about volunteer web-surveys. Therefore, it will offer new insights and an alternative approach to deal with selection biases in volunteer web-surveys.

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

Web survey bibliography - Tijdens, K. G. (27)