Öztürk, Zeynep2023-03-092023-03-092018Öztürk, Z. (2018). The Effect on Convergence Diagnostic Tests and Model Parameters of Thinning Rate. International Journal, 74(11/1),117-128.http://dx.doi.org/10.21506/j.ponte.2018.11.11https://hdl.handle.net/11494/4797The basic subjects for the most users of Bayesian approach can determine prior distributions, the number of samples, start point of sample, number of burn-in and thinning rates by running Markov chain. Prior distributions represent expert opinion before data collection and the sample size is prespecified due to limitation of resources. The other subjects are associated with the speed of convergence of the Markov chains and they are not really problems due to the advancement in computing power. In this paper, after establishing appropriate model and determining the appropriate prior distributions, we have researched whether different thinning rates are an effect on convergence diagnostic tests by keeping the number of samples and the number of burn in a constant. Consequently, the thinning rate has been found to have no effect on the output of the convergence diagnostics and on the parameters. It is important for autocorrelation and memory.eninfo:eu-repo/semantics/openAccessBayesian logistic random effect modelsMarkov chain monte carlo (MCMC)Thinning rateConvergence diagnostic testsArticle7411/111712810.21506/j.ponte.2018.11.11