Relevance & Research Question: We present the results of an analysis on spontaneous use of tablets and smartphones when completing web surveys. Since the surveys concerned were designed for a PC input interface and were not adapted for mobile devices, we call this unintended mobile responding. We investigate the most recent trend of mobile responding in online panels. We also analyze the factors predicting mobile responding using demographic and social behavior related background characteristics as well as the general progressiveness of the respondents.
Methods & Data: Our findings are based on the data of two Dutch online probability-based panels, the LISS panel and the CentERpanel. We logged the user agent strings in several waves in these panels from 2012 onwards. Based on this information, we detected the type of device the respondents used to access the surveys. The results were analyzed in combination with a broad set of background information which is known of the panel members.
Results: Our findings show that unintended, mobile access to online surveys has increased from a couple of percent up to 11-16 percent in the past two years, mainly due to an increase of tablet usage. We also not only find that tablet and smartphone respondents differ from PC respondents by several characteristics but that there are also important differences between these two groups.
Added Value: Based on our findings we strongly recommend testing web surveys on tablets, even when mobile responding is not aimed at. Developing the general web survey design to be more suited for touchscreens is suggested. If participation of young respondents is important, a design accommodating smartphone access should be considered.