Hello! I am in the process of writing my dissertation proposal and need to figure out sample size (I will be using G*Power). To measure my DV, I am using a survey that has 4 subscales each with a mean score. When I run power analysis to determine sample size and indicate that I am going to do a multiple regression (5 IVs), is that sample size estimation for my study or per regression/DV subscale? Thank you in advance.
Hello Sabrina,
If you intend to use the same IVs for each subscale score, apply the same target effect size (e,g,, some minimal R-squared you'd like to be able to detect, should it exist), with the same target power and alpha level, then the results of plugging these numbers into G*Power (or any comparable computation) should give the lower bound for sample size needed that would apply to any and all four of your subscales.
It might be worth asking whether, in your study, it would make more sense to consider using multivariate regression or canonical correlation instead of running 4 separate regression models.
Good luck with your work.
Power calculations apply per analysis, not per study. Good luck with your research.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size