[R-meta] extracting effect sizes from multiple regression

Michael Dewey ||@t@ @end|ng |rom dewey@myzen@co@uk
Wed Oct 12 14:09:36 CEST 2022

Dear Catia

Comments in-line

On 12/10/2022 11:14, Catia Oliveira wrote:
> Dear all,
> I hope this email finds you well.
> Is it possible to compare the findings from multiple regressions in a
> meta-analysis or should one only include the effect sizes from simple
> regression?

Yes, you can.

   If one can incorporate those from multiple regression, should
> we only compare those with effect sizes from studies that have used the
> same predictors,

That would be the most logical way to go since if the variable in which 
you are interested has been adjusted for different covariates then the 
coefficient is estimating different effects. Having said that if the set 
of covariates is similar across studies but with perhaps an extra one 
here and there then you might take the view that the coefficients are 
broadly comparable. If you had one set adjusted for age and another 
adjusted for age and sex you might include all of them with a 
categorical moderator for the different sets of covariates.

  or is there some way to extract the necessary information
> from a multiple regression?

Not without the raw data I think.


> Thank you!
> Best wishes
> 	[[alternative HTML version deleted]]
> _______________________________________________
> R-sig-meta-analysis mailing list @ R-sig-meta-analysis using r-project.org
> To manage your subscription to this mailing list, go to:
> https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis


More information about the R-sig-meta-analysis mailing list