There's more...

What happens if there is substantial autocorrelation for a chain? Or posed in different terms: the effective sample size is small/or the dependence factor is too large? It could be caused by a poor choice of starting values (we could try with different starting values), or improper tuning of the MCMC algorithm (we could increase warm-up). It is always a good idea to increase the number of iterations as well.

Get R Statistics Cookbook now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.