ABSTRACT
In many areas of psychological research, conclusions about the data are based on a test of significance. Typically if the significance level is 5% or less, it is concluded that a real effect has been observed. This chapter is concerned with certain problems with the logic of this procedure. In particular, the significance level is not related to the probability of a Type I error (i.e., the probability that a conclusion that a real effect has been observed is erroneous) in the way that it is normally assumed to be.
