4 ms·

Sample size seems small. Is this adequately powered given the effect size?

• bart_spoon 7y ago
  Power wouldn't be a concern here. Power is simply the probability that, if a statistically significant difference exists, the study correctly conclude there is a difference. Essentially, if a study comes back failing to reject the null hypothesis ("there is no benefit to hydroxychloroquine"), then you might be concerned about power ("perhaps there is a benefit and our sample size was too small to detect it"). There may be other statistical points to be raised about the results of this study, but given that the null was rejected ("there was evidence of an effect of HCQ"), power wouldn't be among them. Here the null hypothesis is that hydroxychloroquine isn't any better than not receiving anything. If the study