We're thrilled to have Caroline, author of Forms that Work and our forthcoming book Surveys that Work participate in our network of UX experts. Like what she has to say below? Then consider having her work with your organization; check her profile page for information on her consulting and and full-day courses.
RM: What's a common misconception people have about the design of forms?
CJ: That it's all about visual design, things like where to put the labels compared to the boxes. Whereas the questions are far more important. It's amazing how users can survive really terrible visual designs of forms—including horrible mistakes like putting the labels inside the boxes—provided they can understand the questions, find the answers easily, and consider that the questions are appropriate in the context of their goals.
In Luke Wroblewski's book Web Form Design: Filling in the Blanks, I contributed a piece, "People before pixels," that dives into this in greater detail—and there's even more in our book Forms that Work: Designing Web Forms For Usability.
RM: What's one thing you wish everyone knew about survey design?
CJ: You've got to test—and preferably, test and test again. The professional survey methodologists are obsessed with testing; they test the topics of the survey on stakeholders, subject matter experts, and data users, as well as on respondents. They test the questions with everybody, and extensively with respondents. They test the heck out of the questionnaire itself ("the instrument") in usability tests and pilot tests. And they iterate: make changes, test again. And again.
I realise I'm not going to persuade everyone to do that much testing for most surveys, but I'd dearly love to persuade everyone to do a bit of testing. "Write, send, and hope" is a recipe for poor quality data at best, and alienating your users at worst.
RM: Thanks Caroline!