多阶段增长模型的方法比较

多阶段增长模型（Piecewise Growth Modeling,PGM）可以解决发展趋势中具有转折点的情形,并且相对其他复杂的曲线增长模型,解释更简单.已有的统计方法主要通过多层线性模型和潜变量增长模型对多阶段模型进行估计.通过模拟研究,用HLM6.0和Mplus6.0对上述两种模型分别进行估计,结果发现在参数估计的精度上,两种估计方法没有差异,只是在犯一类错误的概率上后者略小.进一步通过对错误模型的探讨发现,在样本量小（n=50）,斜率变化小（△b=0.2）时,用线性模型拟合数据而非PGM所犯错误概率较小,整体拟合更佳.但随着样本的增加和斜率变化的增加,错误模型的犯错概率明显增大.故在实际应用中,为了能更好拟合数据,研究者应根据数据本身的情况选择恰当的模型.

Methods Comparison of Piecewise Growth Modeling

Piecewise growth modeling（PGM）is the newly developed model dealing with the problems in detecting the growth trajectory of different periods in longitudinal study. This specific model can describe the development of behavior with crucial change, and can be estimated with the available statistic models, namely hierarchical linear modeling and latent growth modeling. Using simulation study, comparing these two models using HLM 5.0 and Mplus 6.0 respectively, the results indicated that the parameter estimations showed i- dentical between these methods, while the probability of type I error was a lit bit lower for the latter. Furthermore, with the comparison of the improper models,the linear growth modeling seemed superior to the others in condition of small sample size （N = 50 ） and weak slope change （Ab = 0. 2） with the better model fit indices and the minimum probability of error. However, the probability of error of im- proper models increased when sample size enlarged and slope change sharpened. In this case, researchers have to choose the proper model in practical studies according to the database, as well as the statistic considerations.