Using Power Analyses to Examine Post Hoc Trends in Small Samples
Abstract
Many of the studies conducted in clinical research use small samples. This is especially true for small group research. These samples may present problems in the statistical analyses, as they reduce or limit the ability of the statistical tests of significance to detect differences between experimental conditions. This problem may befurther exacerbated when effect size is smalL This article discusses the application of a power analysis when such samples are used. The analysis is conducted after the study is completed and provides information concerning the actual power of the tests and important trends in the data that may have reached statistical significance if larger samples had been used