If you have very solid controls in place then a sample size of 32 is much more substantial IMHO than if you're dealing with real-world environmental factors whe

If you have very solid controls in place then a sample size of 32 is much more substantial IMHO than if you're dealing with real-world environmental factors where you have very little control.