Hi there,
I conducted a propensity score matching analysis on multiply imputed datasets following Ling et al. 2021's guidance.
Psmatching was done within each individual datasets using teffects psmatch, I then stack the results of the analysis from each dataset in an output dataset consisting of regression coeff, A-I std error (which is deemed more robust than the vce stderror), p-values and sample size.
I now want to pool these findings using the Rubin's rule (using this manual: https://bookdown.org/mwheymans/bookm...m-and-p-values). I have done so for the regression coeff and standard errors. I have formulas to derive the Wald statistics and adjusted df using the manual, but stuck on how to derive the p-values.
Can anyone help at all?
Best,
Steve
I conducted a propensity score matching analysis on multiply imputed datasets following Ling et al. 2021's guidance.
Psmatching was done within each individual datasets using teffects psmatch, I then stack the results of the analysis from each dataset in an output dataset consisting of regression coeff, A-I std error (which is deemed more robust than the vce stderror), p-values and sample size.
I now want to pool these findings using the Rubin's rule (using this manual: https://bookdown.org/mwheymans/bookm...m-and-p-values). I have done so for the regression coeff and standard errors. I have formulas to derive the Wald statistics and adjusted df using the manual, but stuck on how to derive the p-values.
Can anyone help at all?
Best,
Steve
