Statistics is no bloodless algebra. It is a discipline at the interface of biology and mathematics. Mathematics is used to answer biological questions. Biological processes are full of variations, and statistics gives no certainties, only chances. What kind of chances: chances that your prior hypotheses are true or untrue. The human brain hypothesizes all the time. And we currently believe that hypotheses must be assessed with hard data.

When it comes to statistical data analyses, clinicians and clinical investigators soon get very nervous, and tend to leave their data to a statistician who runs the data through SAS of SPSS or any other software program to see if there are significant p-values. This practice is called data dredging and is the source of multiple type I errors of finding a difference where there is none.

The best defense against this practice is the use of simple tests. These tests, generally, provide the best power for confirmative research, because this research is based on sound arguments. Multiple variable tests are not always in place here, as they tend to enhance the risk of power loss, data dredging, and type I errors producing a host of irrelevant p-values. Also multiple variable tests, although interesting, are considered exploratory rather than confirmatory, in other words they, generally, prove nothing, and have to be confirmed.

The current book was written for various reasons:

The current book is very condensed, but this should be threshold lowering to readers. As a consequence, however, the theoretical background of the methods described are not sufficiently explained in the text. Extensive theoretical information is also given in the above mentioned books from the same authors.