Which formula is best for determining the sample size for research work? Most especially if the sample population is infinite or not known?
Laverne A. Davis is correct. Your sample size is going to be determined by your design and the analysis that you'll want to conduct. Experimental studies will generally require less power (especially if you also gather covariates in the process), and analyses such as structural equation modeling and multilevel modeling can require over 1,000 cases depending on your model specifications. It also depends on the effect size of the IVs. If you think they're going to have large effects, then you'll need less power to find a significant result than an IV with a smaller effect. There are plenty of intuitive power calculators for various analyses--that should be a good start. Don't reinvent the wheel if an accessible software has this feature. Some researchers have even created Excel spreadsheets with power calculations in the formulas to make it very accessible.
It is recommended to use the largest amount of sample available for your study, especially if it is not experimental.
Alli Noah, I just sent you three excel files with power analyses for different types of studies.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size