Does It Matter If Non-Powerful Significance Tests Are Used in Dissertation Research?
- Heping Deng
Abstract
This study examines the statistical power levels presented in the dissertations completed in the field of educational leadership or educational administration. Eighty out of 221 reviewed dissertations were analyzed and overall statistical power levels were calculated for 2,629 significance tests. The statistical power levels demonstrated in the dissertations were satisfactory for detecting Cohen's large effect (d=0.80) and medium effect (d=0.50) but quite low for small effect (d=0.20). Therefore, the authors of analyzed dissertations had a very low probability of finding true significance when looking for Cohen's small effect. Accessed 19,703 times on https://pareonline.net from September 20, 2005 to December 31, 2019. For downloads from January 1, 2020 forward, please click on the PlumX Metrics link to the right.
Keywords: Dissertation research, poer, effect size
How to Cite:
Deng, H., (2005) “Does It Matter If Non-Powerful Significance Tests Are Used in Dissertation Research?”, Practical Assessment, Research, and Evaluation 10(1): 16. doi: https://doi.org/10.7275/qakz-t063
Downloads:
Download PDF
View PDF