I would call it a failure of interpretation. at one extreme one can assume that your sample is unique, and results inapplicable to anyone else. At the other ext

I would call it a failure of interpretation. at one extreme one can assume that your sample is unique, and results inapplicable to anyone else. At the other extreme is the assumption that your sample is sufficiently representative so that your results apply to everyone. obviously in most instances it is somewhere between these, and that it "failed" is the basis of new inquiry