AAPOR's online journal Survey Practice continues to be one of my favorite places to go for quick updates on the latest methodological research. To wit, the December issue has a nice little piece by Mario Callegaro on mobile users who perhaps unexpectedly show up to take an online survey designed for administration via a laptop or desktop computer. Frequent conference goers have heard lots of papers about the problems of designing for mobile—simple screen, straightforward questions, less than five minutes, etc.). Mario's piece reminds us that the more immediate problem may be dealing with respondents who are ahead of the curve and already expect to do surveys on their smartphone or iPad. He describes how to detect these folks, presents a series of screen captures that show just how badly different mobile devices can scramble a Web survey, and reports on an experiment that shows significantly higher termination rates for mobile versus computer respondents. He concludes with a range of options from ignoring them (as long as they remain a very small percentage of respondents--say, less than 2 percent) to modifying the survey design to accommodate the different screen sizes. It's good advice worth checking out first hand.
We do a lot of very long and complex Web surveys sometimes including multiple DCM exercises, all of which make designing for mobile impossible. Our preference has been to detect the mobile users at the outset and turn them away with a message that says, in essence, you need to do this on a real computer. We've never tracked how many of those people actually come back later but we probably should.