[R-meta] R-square (change) as effect size

[Hanel, Paul H P]

Hello,

Is it possible to run a meta-analysis using R-square (or R) as an effect size in the same way as you run a meta-analysis using Pearson's r?

Specifically, I am interested in which of two sets of psychological constructs is better in predicting a range of outcomes such as well-being or self-reported behaviour. Set 1 consists of five variables (the so-called Big-5 personality traits) whereas set 2 consists of 10 variables (Schwartz's 10 value types). I have the adjusted R-squares from linear regressions for both sets of variables as well as the adjusted R-square change. For example, does set 1 explain more variance in well-being than set 2?

Thanks
Paul



	[[alternative HTML version deleted]]