A great exam question

Iain writes:

I [Iain] remember Dennis Cook used to have a multiple choice question in an exam for a regression class that asked simply “if in doubt do what?” with correct answer “take the log.”

I just want to know what the other options were in the multiple choice.

P.S. Iain told me the story after the following exchange. He wrote that a student in a regression class asked him the following question:

When I [Iain’s student] look at a regression model and try to determined which observation is high/low relative to the regression line I see two possibilities. An absolute difference that can be measured by residuals or possibly visually and a relative difference which is dependant on the residual relative to the size of the observation (or the fitted line at that X value). It is possible for a high X value, in an upward sloping regression, to have a large residual but because the Y values are large you really did not miss by much percentage wise and vice versa at the low end. How best to measure this and what do I do with it. I thought about standardized residuals but they standardizing to the residual not the initial values.

My reply:

It sounds like you’re talking about y being positive, in which case it might make sense to take the log first.