0
$\begingroup$

I am conducting a meta analysis of pearsons r correlations and would like to double check my understanding of independent and non- independent effect sizes/correlations. Am I correct in understanding that all effect sizes/correlations for meta analysis should be independent , as in a single effect size for each article/study , which means that any article presenting more than one effect/correlation related to the same construct being measured should ALWAYS be averaged into one effect size/correlation via converting the r correlations into fishers z scores for averaging and analysis?

Thanks

$\endgroup$
  • $\begingroup$ Further to this , I also have several effects/correlations for the criterion measure in some studies ( job performance of task , contextual , managerial). As my meta analysis investigates how sorting constructs into these performance domains affects the correlations , I am confused as to why I would average them if I need to keep them as separate? $\endgroup$ – laboh Aug 4 '18 at 17:40
  • $\begingroup$ See my answer to a related question here: stats.stackexchange.com/questions/327186/… $\endgroup$ – jsakaluk Aug 4 '18 at 21:01
  • $\begingroup$ thanks jsajaluk. I am interested in creating a single effect to represent the range of non -independent effects using sample size-weighted composite correlations , in following methods of previous author. Although the author does not explain exactly how they did this? $\endgroup$ – laboh Aug 5 '18 at 10:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.