ABSTRACT
Strictly speaking, the validity of the null hypothesis statistical test is not affected by its power; however, running studies with low power not only reduces researchers’ likelihood of publication, but also imperils the potential contribution of individual studies and decreases the reliability of findings at the discipline level. Thus, it is the responsibility of researchers to not only conduct studies with “appropriate” power, but also to report power levels along with the rationale (defence) for its inputs. In the service of this objective, this chapter discusses the concept of power, its importance in designing meaningful studies and the various considerations involved in planning for adequate power levels. In addition, the chapter reports effect sizes and power levels along with researchers’ awareness and attention to these considerations for experimental Behavioural Accounting Research published in Behavioural Research in Accounting (BRIA) during the 2010–2014 period. Finally, as statisticians have long emphasized, the scientific or practical value of a finding cannot be properly evaluated without assessing the effect size obtained and its degree of statistical accuracy; consequently, it is recommended that the confidence interval for the effect size obtained be assessed for its scientific and/or practical implications and routinely reported.
