0

I'm designing a web-based program for a university community. It's crucial that we get answers to some survey questions when people sign up, so we can evaluate the program's effectiveness. The more questions we can ask, the better -- but it's a voluntary program, and we don't want to ask so many questions that people quit during signup.

If the issue were making the questions clearer or easier, I could pilot test them with a small sample of people. But I don't know how to replicate, in a testing situation, the whole experience of wanting to sign up for a program vs. getting annoyed by too many questions. Is there an accepted way to do this?

Alternative options we've considered:

  • A colleague recommended doing quick A/B testing during the actual signup period. Unfortunately, we're planning a big universitywide rollout with many of the signups occurring in the first day, so we can't gather preliminary data that way.
  • We could give a subset of questions during signup and send the others in a second, optional survey. cons: this would reduce the response rate and be more trouble for both users and administrators.
  • We could make the less critical questions optional. cons: I'm not sure if users will realize that they can skip them, or if they'll just quit anyway.
  • I'm not sure if you need a pilot test for this. There's plenty of research about surveys about survey response rates and how the number of questions impacts it. How many questions and survey respondents are you anticipating? – nadyne Apr 29 '14 at 3:20
  • Between 10 and 20 questions, and maybe 1500 participants. – octern Apr 29 '14 at 6:04
  • Is this an actual survey, or is it a registration form? That is, if I don't fill out the survey, can I still register for the program? Your second alternative makes me think that it's a registration form. If that's the case, perhaps this previous question will help. ux.stackexchange.com/questions/55082/… – nadyne Apr 29 '14 at 18:10
  • Thanks, that's exactly what I was asking about... with the exception that the information we need is about how people were doing before they signed up for the program. If they wait to answer a question, their response is much less useful. Still, a short delay wouldn't hurt -- and once they're signed up with us, we can bug them for answers without as much of a risk of losing them. – octern Apr 29 '14 at 20:05
1

What you are looking to do sounds similar to a usability testing process, ie. you want to ensure user satisfaction and task completion. For one of my projects, we tempted our testers with promises of gift certificates for completion, which I considered to be an artificial, but similar desire to sign up.

Its important to make sure your survey questions are not just for curiosities sake and that they can provide real, quantifiable feedback about your process or service. Following this should distill your questions to a number that is both useful for you and understandable by your users.

|improve this answer|||||
  • Yes, we've already cut our questions to the bone and it's now a matter of deciding whether to effectively cut some stakeholders out of the project. I do like the idea of offering a similar incentive to testers. – octern May 8 '14 at 16:38
  • Motivation for tedium and possibly attached hardware is always a plus for your test subjects. :) – slyimperator May 8 '14 at 16:42
  • Other then motivation, another design goal for your registration should be the classic concept of "visibility of the system state" where your users are always aware of their position in a sequence of events. Progress bars, percentages, counts, all of these will help keep your users engaged. – slyimperator May 8 '14 at 16:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.