The medical literature. Let the reader beware.

Since two thirds of the studies appearing in the best medical journals contain unwarranted conclusions, it is important for the physician to be aware of pitfalls. Two common errors are committed; the first consists of confusing statistical significance with medical significance, and the second deals with drawing substantive conclusions from an accepted null hypothesis. In common parlance, significance pertains to importance and meaningfulness, whereas statistical significance specifies the probability that an observed effect could have been produced by chance variation. The null hypothesis is the hypothesis of no experimental effect or correlation. It can be accepted or rejected. The fact that a null hypothesis is accepted does not prove that it is true, ie, that there is no experimental effect or correlation.