the hypothesis that there is not a real difference between the means of two data sets.