No document available.
Abstract :
[en] Relevance: At the beginning of the era of online surveys, these were programmed to be answered using desktop PCs or notebooks. However, due to technical development such as the increasing role of mobile devices, studies on online survey research detect an increase of questionnaires that are answered on mobile devices (md). However, survey navigation on md is different compared to PC: it takes place on a smaller screen and usually involves a touch pad rather than a mouse and a keyboard. Due to these differences in questionnaire navigation, some of the traditional used web question formats are no longer convenient to be answered on a md. The most common formats are matrix questions. To deal with this development, so called mobile-friendly or responsive-designs were developed, which change the layout of specific questions that are not convenient on a md into an alternative mobile-friendly-design. In case of matrixes, these were separated into item-by-item questions which are suggested to be more comfortable to answer on a mobile device.
Research question: However, from a psychometric perspective the question whether these changes in question format produce comparable results is too often ignored. Therefore, this paper elucidates the following research question: Do different versions of responsive-designs actually produce equivalent response?
Data & Methods: Using the data of the first two waves of the Germ and Emigration and Remigration Panelstudy we can base our analysis on more than 19.000 cases (appox. 7.000 using different md). As GERPS makes use of a responsive design, we are able to investigate measurement invariance between different md and desktop device groups.
Results: As the data management is still in progress and will be finished in the end of October, we will be able to present first-hand information based on fresh data. However, first initial analyses reveal differences between md and desktop device versions. Added Value: Our study is one of the first that elucidates the equivalence of responsive design options. Thus, it enhances the perspective on the existence of possible new biases and error sources due to the increased use of md within web surveys.