After applying MCMC multiple imputations (10 iterations) I performed linear regression analysis where I obtain so strange values of pooled estimates (for instance: p - value = 678944.45). There is no inaccurate pooled results in ANOVA or in unadjusted linear models, just in the adjusted one. Moreover, estimations of p - values and CIs are also correct in each of the 10 iterations, problem occurred just in pooled estimates.
I tried to perform imputations several times (different number of imputations, different number of draws, different variables) but wrong pooled estimates always occurred in the same models.
Does anyone have the same problem? Any help would be much appreciated.