help interpret data to test hypotheses. They determine how likely it is that the actual results obtained in an investigation occurred just by chance rather than for the reason posited by the hypothesis. For example, if inferential statistics show that the results of an investigation would happen by chance only 5 percent of the time, then the hypothesis has a 95 percent chance of being correctly supported by the results. An example of a statistical hypothesis test is a t-test. It can be used to compare the mean value of the actual data with the expected value predicted by the hypothesis. Alternatively, a t-test can be used to compare the mean value of one group of data with the mean value of another group to determine whether the mean values are significantly different or just different by chance.