Download presentation
Presentation is loading. Please wait.
Published byHoward Montgomery Modified over 9 years ago
1
“Significance” in Strategic Management Research Steve Werner Department of Management Bauer College of Business University of Houston
2
Overview The problems of statistical significance. The solutions. Conclusions.
3
The Problems of Statistical Significance 1. Statistical significance is not practical significance. 2. The conventional cut-offs are arbitrary and misleading. 3. Frequently occurring Type II errors.
4
Statistical Significance is not Practical Significance a. Practical significance is effect size and importance. b. Related but not the same. c. Statistical significance largely affected by n.
5
Conventional Cut-offs are Arbitrary and Misleading a. Convention is: i. p<.10 = marginally significant; ii. p<.05 = significant; iii. p<.01 = highly significant; iv. p<.001=very highly significant. b. p<.05 from Fisher (1925). c. Counters classical view of predetermining alpha. d. Simplifies hypothesis testing.
6
Frequently Occurring Type II Errors a. Studies conclude no relationship because findings not significant. b. Power issues. c. This occurs in 60% of all studies (Hunter, 1997).
7
The Solutions 1. Solutions to the practical significance problem. 2. Solutions to the conventional cut-offs problem. 3. Solutions to the frequently occurring Type II errors problem.
8
Solutions to the Practical Significance Problem a. Report confidence intervals. b. Report effect sizes. c. Avoiding using “significance”. d. Many journals still do not report effect sizes. e. SMJ: 98.5% of studies report effect sizes.
9
Solutions to the Arbitrary Cut-off Problem a. Go back to classical Neyman-Pearson hypothesis testing with researcher pre-determined alphas. b. Report actual p values (Fisher, 1956). c. Easier to accomplish now. d. SMJ: 13.4% (35/262) of studies report some actual p values.
10
Solutions to the Type II Error Problem a. Increase Power: i. Conduct power analyses; ii. Increase sample size; iii. Increase effect size. b. Relax alpha. c. SMJ: 65.6% (172/262) of studies use p<.10 as cut-off. d. Report actual p values.
12
Conclusion The problems of statistical significance. The solutions. Report actual p values. Consider p, power, effect size, sample size, and importance when evaluating research findings.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.