I've got some multiple regression results assessing the performance of a theoretical model. I want to compare my results with a couple of meta analyses of this theory, as that seems better than picking one or two (of many) studies rather randomly.
My question is this - Can you just compare an R (squared) against another? When a meta analysis reports an effects size in this way, and I've also got a number that looks roughly similar, is that it? Or are there other criteria I should be looking at?
Related question - I also want to make comparisons variable against variable, and will be comparing different r results. One meta analysis refers to r+ (and the plus sign is positioned slightly lower on the line than the r). I have no idea what this means or whether an r can safely be compared to an r+ without some explanation (which would be nice as they are very close!).
Please reply direct and if anyone else wants to read replies I'll happily share them. Many thanks
Caroline Wilson
Institute of Energy and Sustainable Development
De Montfort University
Email: [log in to unmask] <mailto:[log in to unmask]>
Web: www.iesd.dmu.ac.uk
|