多阶段增长模型的方法比较

被引:9
作者
刘源 [1 ,2 ]
赵骞 [2 ]
刘红云 [3 ,4 ]
机构
[1] 香港中文大学教育心理系
[2] 北京师范大学心理学院
[3] 北京师范大学心理学院,应用实验心理北京市重点实验室
[4] 中国基础教育质量评价与提升协同创新中心
关键词
多阶段增长模型; 参数估计; 模型拟合; 一类错误;
D O I
暂无
中图分类号
B841 [心理学研究方法];
学科分类号
040201 ;
摘要
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(Δb=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。
引用
收藏
页码:415 / 422+450 +450
页数:9
相关论文
共 11 条
[1]  
Before and after third grade: Longitudinal evidence for the shifting role of socioeconomic status in reading growth[J] . Michael Kieffer. Reading and Writing . 2012 (7)
[2]   Analysis of multi-arm tumor growth trials in xenograft animals using phase change adaptive piecewise quadratic models [J].
Choudhury, Kingshuk Roy ;
Kasman, Ian ;
Plowman, Greg D. .
STATISTICS IN MEDICINE, 2010, 29 (23) :2399-2409
[3]  
Growth in Phonological, Orthographic, and Morphological Awareness in Grades 1 to 6[J] . Virginia W. Berninger,Robert D. Abbott,William Nagy,Joanne Carlisle. Journal of Psycholinguistic Research . 2010 (2)
[4]  
Growth in alcohol use in at-risk adolescent boys: Two-part random effects prediction models[J] . Deborah M. Capaldi,Mike Stoolmiller,Hyoun K. Kim,Karen Yoerger. Drug and Alcohol Dependence . 2009 (1)
[5]  
Exploring daily variations of drinking in the Swiss general population. A growth curve analysis[J] . Jean‐LucHeeb,GerhardGmel,JürgenRehm,MeichunMohler‐Kuo. Int. J. Methods Psychiatr. Res. . 2008 (1)
[6]   An introduction to latent growth curve modeling [J].
Duncan, TE ;
Duncan, SC .
BEHAVIOR THERAPY, 2004, 35 (02) :333-363
[7]  
Piecewise growth curve modeling approach for longitudinal prevention study[J] . Chih-Ping Chou,Dongyun Yang,Mary Ann Pentz,Yih-Ing Hser. Computational Statistics and Data Analysis . 2003 (2)
[8]   Exploring associations in developmental trends of adolescent substance use and risky sexual behavior in a high-risk population [J].
Duncan, SC ;
Strycker, LA ;
Duncan, TE .
JOURNAL OF BEHAVIORAL MEDICINE, 1999, 22 (01) :21-34
[9]  
Comparisons of two statistical approaches to study growth curves: The multilevel model and the latent curve analysis[J] . Chih-Ping Chou,Peter M. Bentler,Mary Ann Pentz. Structural Equation Modeling: A Multidisciplinary Journal . 1998 (3)
[10]   LATENT CURVE ANALYSIS [J].
MEREDITH, W ;
TISAK, J .
PSYCHOMETRIKA, 1990, 55 (01) :107-122