You've measured a variable in one group, and the means (or median) is not the same as expected by theory (or by the null hypothesis). Is that due to chance? Or does it tell you the mean (or median) of the values is really different from the hypothetical value?