Most biological scientists conduct experiments to look for effects, and test the results statistically. One of the commonly used test is Student's t test. However, this test concentrates on a very limited question. The authors assume that there is no effect in the experiment, and then estimate the possibility that they could have obtained these results. The logic of the Student's t test procedure, which is a null hypothesis significance test, is not very clear when applied to biological experiments, which are usually done to show effects. The Student's t test proposes that one is "not" going to find an effect. The authors suggest that an alternative approach which often suits biological data from experiments very well is to construct a model, for example using a linear or logistic regression to explain the data, and see how well the model fits. A simple example would be a dose-response curve. For scientists who are investigating mechanisms, this process is very satisfactory. The authors conclude that statistical tests are tools to be used carefully and with some prior thought. (Contains 3 figures.)