Pattern analysis and statistical power.
Charles J. Krauskopf · Psychological Assessment · 1991
Various publications concerning pattern analysis with the Wechsler scales do an excellent job of instructing clinicians to avoid Type I errors while ignoring Type II error. Some recognize that one of the clinician's main interests is generating hypotheses, but still advocate such statistically conservative rules that many reasonable hypotheses are effectively hidden. Balancing power and alpha is recommended as a way of reducing Type II error to a more reasonable level (increasing power). Tables are provided to help in identifying the effect of reducing one type of error on the other. Nothing in this argument should be used as an excuse for relaxing alpha when it is possible to increase power by increasing n