Missing Data In Surveys

While sexual violence-specific campus surveys may be preferable to general campus climate surveys, these specific surveys have their own flaws. Specifically, missing data is a widespread issue in sexual violence campus surveys. One study looking at this problem of missing data among 40 different surveys from 2010 to 2016 found that just under 25% of the total female student population, and 17% of the male student population, had responded to the surveys.

One of the problems with missing data in these surveys is that the data is likely to be missing due to one’s experience of sexual violence. Students may be unwilling to answer the survey because they want to avoid reliving a traumatic experience, or perhaps because they don’t have any emotional stake in the topic. The authors of the previous study suggest that sexual assault experience is likely different between survey responders and non-responders and that estimates of sexual violence prevalence are particularly sensitive to these differences, even when the differences are relatively small.

Both general campus climate surveys and sexual violence-specific surveys often conduct non- response bias analyses (NRBA), which can be done by comparing key responses between those who receive different incentives to participate in the survey, or alternatively by comparing responses between early and late responders. The problem with these analyses is that it is impossible for them to identify what the responses would be for true non-responders to the survey without follow-up data.

Chart with question mark representing missing dataAnother study looking at this problem of missing data found that not only was non-response an issue, but so was non-completion of sexual violence surveys, and in fact, there was a significant difference between those who completed more of the survey and those who completed less. The study also found that students who responded to the survey were not demographically representative of the student population at the target university, in terms of age, gender, minority status, international student status, and year in school. These results suggest that the data obtained from sexual violence surveys may not be providing accurate information about the true prevalence rates of sexual violence on campus.

Guide: PDF Version