Comparing wo treatments or exposures

    One of the things we are always worried about when we compare two treatments or exposures is that our observations may result in a type I error (false positive result); that is, that we will reject the null hypothesis of no difference when there really is no difference. What are the two primary explanations for a false positive result?