Randomizers Reduce Dropoff

Question randomizers are a tool used to vary where a question shows up in the survey. The primary purpose of these randomizers is to avoid question bias (where one question affects the results of another question), test survey responses, and prevent survey habituation. But while these are the most common uses, they're not the only ones, and one of the ways you can use question randomization is to control the effects of survey dropout rate.

Limiting Questions The first method is to randomize which questions from a larger pool show in a shorter survey. For example, if you have 35 questions but you notice that there is a high dropout rate right around 15 questions, you can "cut" your survey down to 15 questions, and simply randomize which of the 35 will show at any given time. You'll need a much larger sample, but eventually you'll be giving out shorter surveys to customers while still having them answer, as a whole, a much larger dataset.

Collecting Data Anyway Imagine you have a survey with a 100% dropout rate after 15 questions, but you have 100 questions overall and you refuse to cut it down. In other words, every respondent will drop out before reaching the end of your survey. A less than ethical way of still receiving answers is to your questions is to randomize all of them.

The users will see different questions for the first 15, you'll still collect data, but the users won't be finishing the survey. This can be used no matter your dropout rate, as a way of increasing the size of your sample for each question even if the users drop out. Testing Dropout Rate Finally, you can also see if question order is what creates the dropout rate in the first place.

Imagine you are not randomizing questions and you notice that people are dropping off at about question 15 on average. While you may believe it's because your survey is too long, the truth is that it may be that questions 14 or 15 are too tough, personal, or otherwise hard to answer. Randomizing it allows you to test why dropout rate is occurring, which in turn can help you control it.

All of these are ways that randomizing the questions in your survey can help you control dropout rate and, ultimately, negate it altogether.