Hi all, 

I am wondering about the details of how SDM computes the Egger test. I compared the results of the SDM-implemented Egger test with the Egger test in the metafor package in R, and while the overall conclusions were the same, the t and p values were different. Does SDM use only the last imputation for a given study, or average over all imputations? What kind of model is being specified and what is the predictor of the regression (standard error, sample size, etc.)? 

Thanks very much.

Anna

Dear Anna

SDM conducts a meta-regression of standard error across imputations and then applies Rubin's rules (see "metabias" at https://cran.r-project.org/web/packages/...)

With best wishes,

Joaquim