I visited the Environmental Protection Agency’s site for some work-related research and was asked to take a short survey when I finished my visit. I agreed, only to discover that the survey was a man-made disaster. See for yourself.
Good survey design is (understandably) not a legislative priority, but the environmental watchdog broke many unofficial laws when it created this monstrosity.
- When asking customers to rate various aspects of your site, lump all aspects together that have the same scale. (Here, questions 1-8 could be “Question 1”.)
- Only use a 10-point scale when absolutely necessary, i.e., when you can explain a meaningful difference between a 3 and a 4. (Here, questions 1-18 could use a 7- or even 5-point scale and be less cluttered and provide more meaningful responses.)
- Not every rating needs to be on a X-point scale. (Here, “How well does this site meet your expectations?” could have the following answer options: Exceeds my expectations, Meets my expectations, Falls short of my expectations.)
- Limit the number of answer options for any given question. Surveys should be as easy as possible for respondents to complete, and having long lists to read through is obnoxious. (Here, questions 26, 27, and 28 are are horrible and question 29 makes me want to redirect my tax dollars.)
- Resist the temptation to ask everything you’d like to know. I know you’re curious, but figure out what is crucial and actionable and sit on the rest. This means that if you don’t know what you’ll do with the information, don’t ask for it. (Here, it is absurd to ask a casual visitor to the EPA site 34 questions.)*
In going through the EPA survey, do you notice any other flagrant violations of good survey design?
* If you have more burning questions than can reasonably fit into a casual visitor survey, provide an optional field for respondents to provide their e-mail addresses if they’d be willing to participate in future/follow-up surveys.