Hi everyone,
I have conducted a sensitivity power analysis for linear multiple regression, r2 increase on a dataset containing 738 participants, 5 tested predictors, 39 total predictors at alpha .05 and power .80. This produced a minimum effect size detectable of r2 of .018. I was wondering if I could use this information to calculate the minimum Sr2 possible to detect?
TIA
Hello Peter,
What do you mean by "Sr2" in your query?
What you have done thus far is ascertain that the smallest change in R-squared from the 34 IV model to the 39 IV model, given the sample size, alpha level and target power you've indicated, is .018.
If you're asking, what would be the smallest overall R-squared for your situation with the 39 IV model, it would be .018.
If not, could you kindly elaborate?
Good luck with your work.
Hi David,
Sorry, I should have elaborated on that. By Sr2, I mean semi-partial correlation squared. I have been told previously that it is possible to calculate a minimum detectable semi-parital correlation squared using power analyses but I cannot remember how.
Thankyou :)
- Badges
- Science topic
- Similar topics
- Filing