Evaluating the medical literature. Part II: Statistical analysis.

We have attempted to provide sufficient information to enable the reader to verify that the investigator has used an appropriate statistical test for the evaluation of his study data. We have not discussed the actual calculation of the tests presented. With the wide availability of computers and programmable calculators, it is safe to assume that the investigator has performed the necessary mathematics accurately. Instead of assessing the accuracy of these calculations, the reader should verify that the correct statistical test was chosen in the first place. The astute reader may be surprised at how frequently an incorrect test is used. Merely achieving statistical significance does not characterize the author's data as clinically important. Neither does it, in and of itself, prove one agent superior to another nor prove a cause-and-effect relationship between two variables. When appropriately interpreted, however, statistical analysis can be a very useful and powerful tool in helping to arrive at the "truth."