Using Behavioral to Improve Honesty from Respondents
I recently participated in a Behavioral Economics class for market researchers, and one of the first things that was discussed was "how do we know that people are giving honest answers to surveys?" Good question... I guess we really don't know for sure! So why wouldn't people give honest answers? Basically, there are two primary reasons:
1) Social desirability
People may understate their alcohol consumption when the doctor asks at an annual physical (ok, I have), or they name more "highbrow" channels when asked what they typically watched. For the latter, I watched many focus groups where "what channels do you typically watch" was usually a warm-up question, and I was always amazed by the PBS responses! If everyone who claimed they watched PBS actually did, then I'd bet PBS wouldn't need pledge drives!
People will alter their information to their advantage. A good example here might be related to life insurance, where people may underestimate their health and wellness.
So what are some Behavioral Economic techniques that we can use to try to get people to be more honest?
- The first one is very simple: just ask them to be honest. Many of us do this in our survey and focus group introductions already by saying something along the lines of "there are no right or wrong answers, we just want your honest opinions."
- The second one is to give them a "no opinion/don't know" option. When I heard this, my first thought was, "Really?" There are plenty of times we ask people for their attitudes about something (usually through some sort of rating scale), but how many of us give them a "no opinion" option? I know, it was ingrained in my head during my research career that we need to force a rating out of people - if we give them a "no opinion" option, they'll just take the easy way out and choose that. But maybe we've been thinking about this wrong all along - what's so wrong if you really don't have an opinion? There's plenty of categories I'm not invested in that I don't have an opinion on, and maybe it's time to rethink the way I've been doing things.
For this second Behavioral Economics technique, a good first step may be to conduct research on research by including a "no opinion" option to one set of respondents vs. making respondents provide a rating/perception. It will be interesting to see if the results are impacted - stay tuned!