2  Chains may also be slow to converge due to high autocorrelation, a phenomenon where adjacent samples in a chain are highly dependent on each other. Some researchers address autocorrelation by thinning the MCMC chain. We use the Mplus default of no thinning in our tutorial, but we encourage readers who are concerned about autocorrelation in their analyses to check Depaoli (2021) or Kruschke (2014).