Statistical Significance Tests: Scientific Ritualism or Scientific Method?

Abstract

The misinterpretation and misuse of significance tests is a serious problem in social work research. In this paper, the problem is defined in terms of errors of interpretation and technical errors. Examples of these errors are drawn from social work literature. It is recommended that statistical significance tests be de-emphasized or replaced by other procedures when requirements of the tests are not met, and that professional journals should better scrutinize articles that utilize such tests.