Hertzog C, Nesselroade J R
Child Dev. 1987 Feb;58(1):93-109.
The use of structural modeling techniques to fit change concepts, including developmental ones, to repeated-measurements data has been rather firmly but uncritically wedded to autoregressive model specifications. The uncritical application of an autoregressive specification to repeated measures does not take into account subtleties of conceptions of stability and change (e.g., the trait-state distinction) that are now recognized in the behavioral research literature. We review the basic distinction between trait and state and examine the implications of the different possibilities for modeling developmental phenomena. The arguments are illustrated with empirical examples.
使用结构建模技术将包括发展性概念在内的变化概念与重复测量数据进行拟合，这一做法相当坚定但不加批判地与自回归模型规范相结合。将自回归规范不加批判地应用于重复测量，没有考虑到行为研究文献中现已认识到的稳定性和变化概念的细微差别（例如特质-状态区分）。我们回顾了特质与状态之间的基本区别，并研究了对发展现象进行建模的不同可能性所带来的影响。文中通过实证例子对这些论点进行了说明。
