Originally Posted by virus646
True, but the reality is that it will always be hard to have good studies on this kind of subject. It's a small field with researches not always done on trained individuals. If there are both studies and real world experience backing up a claim, I'm usually inclined to accept such claim.
Very hard to generalize in this topic imo.
Those issues could be addressed by conducting studies with enough participants (and with a reasonable selection strategy) though, allowing both generalization about average effects and mapping of individual differences , but I get the feeling that there isnt much funding for doing that type of studies....
Fuji, there are huge amounts of online shit about R as you probably know, but you might like this book, uses R and WinBugs (which I dont know though) for multi level regression models http://www.stat.columbia.edu/~gelman/arm/