Contemporary Analysis of Simulation-Based Research Data: P Values, Statistical Power, and Effect Size
Many quantitative researchers proceed from the assumption that statistical significance as represented by the p-value is enough to explain their results. In reality, though, p-values form only a part (albeit an important one) of the logic of hypothesis testing. The purpose of this chapter is to explain this logical flow by addressing core concepts such as the null hypothesis, alpha and beta error, and statistical power. Particular attention should be paid to the concept of effect size, which is a quantitative means of expressing the magnitude of an observed effect. Only when a p-value is correctly interpreted in the context of the power and effect size of a study can the results be given the most appropriate interpretation and meaningful conclusions be derived.
KeywordsEffect size Statistical power Estimating sample size Type I and II errors Hypothesis testing Effect size calculators Research planning Pre- and post-study power calculations Statistical tests and related effect size formulas
- 3.Cumming G. Understanding the new statistics: effect sizes, confidence intervals, and meta-analysis. New York: Taylor & Francis Group, L.L.C; 2012.Google Scholar
- 4.Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale: Lawrence Erlbaum Associates; 1988.Google Scholar
- 9.Grissom RJ, Kim JJ. Effect sizes for research: univariate and multivariate applications. 2nd ed. New York: Taylor & Francis Group, L.L.C.; 2012.Google Scholar
- 13.http://www.gpower.hhu.de/ [Web site for the G∗Power application].