The Impact of Visual Components on Online Survey Response
March 31, 2012
Although researchers are always looking for improved ways to design their studies in order to maximize cooperation and completion while minimizing item nonresponse or other sources of error, now seems to be a particularly good time to experiment with online survey design. Just in the last week articles on the Web from Lightspeed Research, Greg Heist at Gongos Research, and others have talked about the growing problem of declining response rates to our online surveys and called for shorter, simpler, and more engaging survey designs. Heist takes the idea of engagement a step further and contends that online research designs should strive to deliver “fun” to the respondent – “We’ve made the entire experience [of survey completion] about as much fun as a trip to the DMV.”
With survey length and grid questions contributing most to incompletes (see Lightspeed), it is reasonable to look for new ways to gain respondents’ attention and keep them attentively thoughtful throughout questionnaire completion. Online platforms provide any number of optional features for incorporating “eye candy” into online surveys, including face scales and graphic images, as well as a slew of functions within a rich text editor. All of which is great, as long as this visual stimulation doesn’t degrade the quality of the research; and, at the very least, researchers have some understanding how visual cues embedded in online survey designs impact response behavior.
A number of researchers have explored this issue. Mick Couper, Frederick Conrad, and Roger Tourangeau wrote about “Visual Context Effects in Web Surveys” and concluded from their experimentation that visual images do in fact have an effect on survey response. Achim Elfering and Simone Grebner found in their research that men and women interpret face scales differently – e.g., women were more likely to interpret a horizontal (“neutral”) line on a face as depicting “sad” while men were more likely to construe the neutral line as “happy.” And, Vera Toepoel looked at visual context effects related to two key components – verbal language and images (pictures) – and concluded that ”there is a hierarchy of features that respondents attend to, with verbal language taking precedence over visual cues like pictures.”
This is an exciting area in research design and it is hoped that more researchers, esp., those who cry most loudly for dumbing down our designs to “research snacks,” will initiate their own experimentation with online research design to better their own studies as well as contribute to the industry.