Hi,
We conducted a small preliminary research that included a multiple linear regression model. Then I tried to calculate its power using an online post-hoc power calculator for multiple regerssion (https://www.danielsoper.com/statcalc/calculator.aspx?id=9)
When entering our data in the calculator (no. of predictors = 2, R2= 0.504, alpha=0.05 and number of subjects=25), I got a statistical power of 0.991. I'm quite surprised with that value, as it seems too large for me, especially with the relative small sample size.
Am I doing anything wrong, or is that value is correct and possible?
Thanks,
Uri,
Why are u worried about power value of 0.991?
Pl remember that ' Power analysis is the name given to the process for determining the sample size for a research study'.
The technical definition of power is that it is the probability of detecting a “true” effect when it exists.
Uri, the reason is that R^2 is very large. Cohen's effect size f^2 is generally calculated as R^2/(1- R^2), which is in your case 0.504 /(1- 0.504 ) = 1.02. This is a very large effect size, and therefore power is almost 1.0. You would of course get the same result using G*Power.
Kind regards, Karin
Thanks, Karin Schermelleh-Engel! I was not aware that you could calculate effect size by R2.
Best
John-Kåre
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size