This is a convenience sample that would likely overestimate average daily exercise time for all college students.

The sample is reasonably representative of college students in general.

This sampling method introduces selection bias because gym visitors are more active than typical students.

The resulting sample statistic would have a systematic bias away from the true population parameter.