Survey Length vs Abandonment

Often on this blog we discuss the idea that survey length is correlated with a higher dropout rate. And indeed it is, as numerous studies have shown that the longer your survey, the more likely the recipient quits the survey before it's completed. This is a problem. Not only are you failing to get responses, but you also increase the likelihood that those that dropped out have something in common.

Still, dropouts are a fairly common element of surveys. On average, about 10% of all respondents will drop out before the survey is completed even if it takes only a few minutes. If you already have respondents dropping out, how much more are you willing to put up with? If the answer is "quite a few more" then long surveys suddenly become more appropriate.

In fact, surveys that take roughly a half hour (considered to be a very long survey), still only have a dropout rate between 15 and 25%. You are still likely to get at minimum 75% of the started surveys completed.

Furthermore, there are ways to decrease the dropout rate, including: Incentives that improve people's willingness to complete the survey. Keep the individual posted about how they're doing. Decrease the time necessary to answer each question by reducing its simplicity.

Make sure they know the length of the survey in time up front before they begin. Survey dropout rate has the potential to decrease if you put any of the above ideas into effect, depending on your sample and the type of survey you're running.

In addition, while researchers don't often like to admit this, you do still receive partial data when someone completes part of a survey, all without giving up the incentive. In some ways, an incomplete survey can be advantageous – although the missing data may make it harder to do your analysis.

Overall, the key thing to understand is that while we do need to warn you about the perils of a long survey, the reality is that a lot of researchers can get away with added length if they're not concerned about the risks that occur with regard to dropout and the likelihood of a representative sample. Survey dropout is a problem, but a long survey still tends to get a fairly decent completion rate regardless of the overall length of the survey.