We have a series of surveys with extensive branching, skip logic, etc. When we export the data, missing values for questions that were not displayed are indistinguishable from missing values for questions that were displayed but skipped by the respondent. I know that an easy way to distinguish between them would be to force responses, but ethically (and per IRB regulations) we cannot force participants to respond to every question, and requesting a response but allowing them to skip after a reminder gives us better data than forcing a response but offering a "refuse" option for every question.
Is there a way to automatically populate non-displayed questions with "999" or some other code? Or to populate questions displayed but left empty by the respondent with a similar code? We are collecting a very large amount of data monthly from a large panel of study participants over several years, so anything we can do to simplify high frequency checks and quickly identify how much data is actually missing (vs. not applicable) would be very helpful.
Page 1 / 1
There may be some other way to handle but since you are using branching logic you could manually set embedded data for each question (or block) that was skipped.
Here's how the logic would look:
!
Here's how the responses look (not in the most recent repose it was shown but the question was left blank:
!
Here's how the logic would look:
!
Here's how the responses look (not in the most recent repose it was shown but the question was left blank:
!
Thanks @TomG!
Leave a Reply
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.