- #1

- 122

- 0

I'm using this this calculator to determine a rough sample size for a multiple regression (20 predictors).

http://www.stattools.net/SSizmreg_Pgm.php

I don't really understand the

**effect size**?

Could somebody tell me if you are using multiple regression (with 20 predictors) and you want the regression to have an R2 > 0.8, do you want a small effect size for example 0.2 or a large effect size for example 0.5?

Basically I want a accurate model, what effect size value should I use?

Thank you