How large is the sample? - How representative is it? - Was the data collected before or after the theory was formed? - Is the data noisy? How much of it is likely noise?