Mcmc diagnostics converging problem


I notice that in my program the MCMC of most parameters seem ok (the lines are close together, there isn’t much and the values seem to converge). The multivariate diagnostics also seem ok.

But some of the MCMC graphs for a few parameters do not seem to converge to the same value.

Since the multivariate diagnostics are ok, does that mean I can ignore the problems with those parameters?


No usually not. If you know that some parameters have not converged, you cannot trust the other parameters. They are not independent.