THE RATING tables (pp. 191-193) giving an approximate evaluation of psi test results have already been mentioned. These obviate the necessity of making statistical calculations. The present chapter will describe, for the benefit of those who may wish to know bow to make a more exact evaluation of their results, the statistical methods most widely used in parapsychology.
The methods needed to evaluate results from the tests explained in the preceding chapter will be described in some detail. Other statistical methods have also been used in parapsychology, and the more important of these will be mentioned, with references to the literature in which full descriptions may be found. The statistical methods are considered here solely from the point of view of their application; that is, no effort will be made to examine the theory or give the derivation of any of the procedures. In the main the methods are well-known procedures which are discussed in the statistical textbooks.* In some instances, however, experimenters in parapsychology have raised questions in connection with test results that could not be evaluated by the available methods. These questions have led to advances in statistics to provide the new methods required.
A statistical test of significance is simply a means of evaluating a result in relation to what might be expected on a purely chance basis. For the evaluation to serve its purpose, it is of the greatest importance to make sure that no nonchance factors are allowed____________________