Compared with restricted estimation of maximum likelihood (REML), MCMC was preferred to parameter inference of variance components, and was more robust for small sample data.

• 同时和基于似然的限制性最大似然估计法（REML）相比，MCMC方法在参数推断方面有更强的优越性，在小样本情况下可得到更稳健的参数估计。