Implications of latent trajectory models for the study of developmental psychopathology
Implications of latent trajectory models for the study of developmental psychopathology
复制标题
DOI：
10.1017/s0954579403000300
复制
发表时间：
2003-06-01
影响因子：
3.3
通讯作者：
Willoughby, MT
中科院分区：
文献类型：
--
作者：
Curran, PJ;Willoughby, MT
发展精神病理学领域面临着双重挑战。一方面，我们必须建立跨学科的理论模型，充分反映正常和异常人类发展的复杂性。另一方面，为了保持一种可行的经验科学，我们必须使用充分捕捉这种复杂性的统计方法来严格评估这些理论。我们的统计模型与理论模型的不一致程度会削弱我们有效测试发展理论的能力。广泛的随机系数轨迹（或增长曲线）模型使我们能够以以前不可能的方式测试我们的理论。尽管有这些优点，但这些模型目前可以评估的问题类型仍然存在一定的限制。我们通过追求三个目标来探讨这些问题。首先，我们提供了一个概述的各种轨迹模型，可用于严格测试发展精神病理学的许多假设。其次，我们强调了哪些类型的研究问题是使用这些方法进行了很好的测试，哪些类型的问题目前还没有。第三，我们描述了未来统计发展的领域，并鼓励发展理论和定量方法之间的持续交流。
The field of developmental psychopathology is faced with a dual challenge. On the one hand, we must build interdisciplinary theoretical models that adequately reflect the complexity of normal and abnormal human development over time. On the other hand, to remain a viable empirical science, we must rigorously evaluate these theories using statistical methods that fully capture this complexity. The degree to which Our statistical models fail to correspond to our theoretical models undermines our ability to validly test developmental theory. The broad class of random coefficient trajectory (or growth curve) models allow us to test our theories in ways not previously possible. Despite these advantages, there remain certain limits with regard to the types of questions these models can currently evaluate. We explore these issues through the pursuit of three goals. First, we provide an overview of a variety of trajectory models that can be used for rigorously testing many hypotheses in developmental psychopathology. Second, we highlight what types of research questions are well tested using these methods and what types of questions currently are not. Third, we describe areas for future statistical development and encourage the ongoing interchange between developmental theory and quantitative methodology.