Two Approaches to an Evaluation of the Reliability of Psychological Investigations (Reliability of a Fact and Syndrome Analysis)

A psychologist faced with the problem of ensuring the necessary degree of reliability of a fact he has discovered usually proceeds along a scientifically well-established path. He defines the problem, isolates the relevant area of investigation, finds a procedure adequate to the problem, and collects the number of cases necessary to ensure the reliability of the discovered fact. The criterion of reliability in these cases is the amount of data the investigator has at hand and the amount of variability in the results obtained. It is generally accepted that when the facts at hand are completely homogeneous, a relatively small number of experiments is sufficient to ensure reliability, whereas if the variability of the obtained data is relatively great, the number of observations required to ensure the statistical reliability of a fact must be increased considerably.