Welcome to the Listening Post!

Maximizing Survey Completion Rates

Tweet

 

You’ve improved the participation rate of your survey so that more invitees are clicking through to begin the survey (q.v., Improving Response Rates to Web Surveys). Unfortunately, respondents answer the first page of questions and exit as soon as they see the second page. Sigh…

How can you increase the number of respondents who complete your entire survey?

Extensive academic literature dating back to the late 1990s mostly proves the obvious when it comes to maximizing completion rates for web surveys. Here’s the advice:

• Shorten the survey - The longer the questionnaire, the higher the abandonment rate. Whether length is measured in pages or words or question blocks, longer questionnaires have lower completion rates. Ganassali 2008 found that abandonment spiked starting as soon as the second page. Check out these seven tips for shortening your survey.
• Decrease the number of matrix/grid questions - Matrix questions often prompt participants to abandon the survey (Brecko, Carstens; 2006). So rethink the matrix question.
• Reduce the number of open-ended questions - Many studies found that the more verbatim questions the higher the drop-out rate (Galesic 2006). Replace open-ended questions with closed-ended questions where possible.
• Put demographic questions at the end - Demographic questions in the middle of a questionnaire prompted respondents to exit: "respondents might assume this is the last part in the questionnaire" (since it frequently is). Keep any demographic questions not used in your screener at the end of the survey (Brecko, Carstens; 2006).
• Hide progress indicators on long surveys - A progress bar at the bottom of a long survey increases the likelihood of abandonment (Matzat, Snijders, van der Horst; 2009). [Despite this research, I advocate progress bars: respondents prefer them. Progress bars indicate you value and respect the respondent. In this case, just not enough to shorten your survey.]

Adding survey incentives does not appear to decrease the abandonment rate.

In “Dropouts on the Web: Effects of Interest and Burden Experienced During an Online Survey” (2006), Galesic proposes a simple model for predicting dropout behavior: “The higher [respondents’] overall interest in questions and the lower their overall experienced burden, the lower the risk of their dropping out.” In the survey tested, the burden experienced at the end of one block of questions was a function of the number of questions preceding the block, the amount of time required to complete the block and the number of open-ended questions the block contained.

While burden can be reduced to a formula, interest is not so easy to tame.  Want to maximize the completion rate? Make the survey interesting and enjoyable. Perhaps add a question in the middle that is more for entertainment purposes then research purposes. Engaging surveys are novel, clever and well-written, and—as a result—sustain respondents’ attention.