Still confused about this one. I guess it depends how you read that last line:
"Which of the following potential flaws is most likely to invalidate this study?"
If you read it as which flaw is most likely to be present, then I guess selection bias is most likely.
If you read it as which flaw, if present, is most likely to invalidate the study, then that type I error would be 100% likely to sink the results, guaranteed. Guess they meant the first one... Because yes p=0.01 so it's not likely they had a type 1 error.
Beef with this q.
submitted by ∗bwdc(697)
A cohort study (as opposed to a randomized controlled trial) is ripe for selection bias, which occurs when the treatment and control groups are not truly comparable. Matching for some factors (age, gender) doesn’t mean you’ve controlled for all possible confounders. That’s what randomization does!