My population is large >10k, however when i calculate for sample size using G Power, using bigger effect size, my sample size becomes less than 100. Do i change the effect size to justify my large population?
I do not use the 10 times rule as many literature do no not recommend it.
Also what would be the best settings in G*POWER, for a study that that uses SEM PLS. 5 independent variables and one dependent variable.
Pls advice. thank you.
With G*Power you have several different ways of approaching power in your tests. In your case, it seems that you set as fixed the effect size, the alpha level (usually 5%) and the beta level (usually 80%). What G*Power then spits out is the minimum sample size to achieve this. However, if your sample size is already fixed, you can choose another approach, that is, finding out the effect size you'll be able to discover, given a fixed sample size, alpha, and beta. I think this is called sensitivity analysis. In your case, you will be able to discover very small effects with a low probability of false negatives (Power of 99% or something), which is really nice.
However, I am not sure about which settings specifically work for PLS-SEM. Very likely, you'll find some relevant information in Hair et al.,'s Primer on PLS-SEM Book. It's generally a great read if you're doing PLS-SEM.
Edit: This might be good start:
https://www.researchgate.net/post/What-is-Sample-Size-Recommendations-when-using-PLS-SEM
let me know if this helps…
josh
- Badges
- Science topic
- Similar topics
- Physics
- Condensed Matter Physics
- Solid State Physics
- PEAKS