Statistical Power and Effect Size in Educational and Psychological Research

• Main Author: Faisal Ahmad Abdelfattah
• Format: Article
• Language: Arabic
• Published: King Saud University 2011-04-01
• Series: Journal of Islamic Studies
• Subjects: Statistical power; effect size; statistical significance; type II error
• Online Access: http://jes.ksu.edu.sa/sites/jes.ksu.edu.sa/files/0023-02-01.pdf

The study aimed at identifying statistical power and effect size in number of published research in Education and Psychology. In addition, discovering how much researchers are committed to provide results with enough requirements to calculate both statistical power and effect size. To achieve the study goals, 64 articles from Journal of Educational and Psychological Sciences and International Journal for Educational Research were analyzed for power and effect size levels using the G*Power 3 software. The results show approximate 6% of hypotheses were not provided with enough information to calculate statistical power and effect size. Of the hypotheses tested, one – quarter hypothesis did not meet the accepted level of power 0.80, which raises the occurrence of type II error. Additionally, Results from failed to reject null hypotheses show lower level of effect size. Conversely, rejected hypotheses were similarly distributed across low, medium, and high effect size levels. Based on study results, Researchers should consider power analysis and report effect sizes accompanied with corresponding practical significance for all hypotheses in education and psychology literature.