mediation icon indicating copy to clipboard operation
mediation copied to clipboard

multiply imputed (MICE) data: pooled p-value calculation

Open makikoworldwide opened this issue 2 years ago • 0 comments

I am conducting a mediation analysis with multiply imputed 30 data sets using "amelidiate {mediation}".

variables: mediators= m,outcome=y, treatment=rep(e,30),covariates=z, control_val=a0, treat_val=a1,

medout <- mediations(datasets, treatment, mediators, outcome, covariates, families=c("binomial","gaussian"), control.value = control_val, treat.value = treat_val, interaction=TRUE, conf.level=.95, sims=1000, set.seed(1))

med_pool <- amelidiate(medout) summary(med_pool)

The package provides pooled CI for all estimates but does not provide pooled p-values. Would there any way to calculate pooled p-values?

Thank you so much!

makikoworldwide avatar Feb 24 '23 07:02 makikoworldwide