I am working with a survey data set and am curious about who abandoned it before finishing all the questions. I have information on how many questions each respondent answered (no skipping allowed) and I also know additional data about each respondent. Would it be appropriate to use survival analysis to determine what makes someone more or less likely to finish the survey? It seems obvious, but I haven't been able to find a good guide in the literature.
Thanks in advance.
Survival analysis is a time-dependent analysis that is typically used for follow-up data in prospective or retrospective studies. Since in this case we have a cross-sectional study where no follow-up was performed, this analysis is not appropriate. The correct approach is to use logistic regression, which explores predictors of a dichotomous event in a time-independent fashion. You can set finishing/no finishing of the survey as the dependent dichotomous variable and test the other variables as predictors in a multivariable model. The regression will identify independent variables associated with the probability of finishing the survey.