What steps would you take to account for selection bias in a dataset collected from voluntary student surveys at Stanford University?
