Question

Describe a study in which you can use both nonparametric and parametric statistics. A researcher has participants expecting to receive either painful or mild electrical shocks during a study wait in a quarter-full waiting room with other participants. He then measures whether they choose to sit next to other participants or if they sit far away on their own. What is the best test to run for this study and why?

Answer 1

Usually in survival analysis studies we can use both parametric and non-parametric statistics. For example, in trying to find the estimate of a survival function of a group of patients we can use both parametric and non-parametric test statistics. Kaplan-Meier estimate or Product-Limit estimate is the non-parametric procedure and Cox proportional hazard model is the parametric procedure.

For the given problem the best test to run for this study is a Binomial test because there are only 2 possible outcomes here: one is if they sit next to other participants and the other is if they sit far away on their own. Since we are testing for only these 2 possible outcomes here, a Binomial test will be appropriate.