Incorporating Statistical Power into the Test of Significance Procedure: A Methodological and Empirical Inquiry.

Statistical power represents the probability of rejecting the null hypothesis when it is false and some specific alternative hypothesis is true. Despite the crucial role it plays in providing for meaningful tests and obtaining reliable inferences, the concept of statistical power rarely surfaces in accounting-either methodological or empirical. The purpose of this article is to redress this imbalance in arguing for its formal inclusion in the test of significance procedure. Specifically, the article examines what power is, the serious consequences of failing to formally incorporate power in the inference process, and the twofold role of power analyses in classical statistical tests. In addition, a survey of management accounting research is undertaken for the purposes of calculating the statistical power of the average study and determining researchers' awareness and attention to power considerations. The results indicate that researchers display little understanding or appreciation for the importance of power in conducting their research; moreover, they have paid a dear price for this: the power of the average study is unacceptably low. Also, the obtained level of significance p continues to be the predominant measure used by many researchers to interpret the meaningfulness and overall importance of their results. Taken in total, these findings suggest a situation which can be expected to impede the development of the discipline. Finally, a redress to the low power problem facing many behavioral researchers is examined, along with an application of the approach.