追踪研究方法在国内心理研究中的应用述评  

Review of the Application of Longitudinal Study Method in Psychological Researches in China

作　　者：唐文清[1,2] 方杰[3] 蒋香梅[1] 张敏强[1] 

机构地区：[1]华南师范大学心理学院、心理应用研究中心,广州510631 [2]广西大学教育学院,南宁530004 [3]广东财经大学人文与传播学院,广州510320

出　　处：《心理发展与教育》2014年第2期216-224,共9页Psychological Development and Education

基　　金：全国教育科学“十二五”规划教育部重点课题(GFA111009);广东省哲学社会科学“十二五”规划2013年度一般项目(GD13CXL01);2012年度教育部人文社会科学研究青年基金项目(12YJC190016);广西大学科研基金资助项目(XGS100019);广州市基础教育学业质量监测系统(二期)(GZIT2013-ZB0465)

摘　　要：以1982~2012年中国期刊网收录的88例追踪研究为对象,从应用现状、设计特征、数据处理三方面分析和评估追踪研究方法在国内心理研究的应用情况及存在的问题。结果显示,2005年之前追踪研究方法应用增长缓慢,2005年开始呈显著增长趋势,研究对象以未成年及成年早期群体为主。主要采用固定样本追踪设计,大部分研究测量2-3次、样本量在10~300之间、持续时间在3年内。61例有缺失的研究中,38例用删除法处理缺失;主要运用HLM、方差分析、t检验和SEM分析追踪数据。相当部分研究存在测量次数少、样本量较小、持续时间短、被试缺失严重及数据处理方法相对陈旧问题。追踪研究方法的应用应注意,根据理论模型和研究有效性要求确定设计类型和设计特征,根据数据特征选择缺失处理方法和追踪数据分析方法。To explore current situation and problems of application of longitudinal study method in psychological researches in China, this article reviewed 88 studies published in Chinese psychological journals or Ph.D. thesis and master＇s thesis, obtained from the methodological literature in CNKI database on longitudinal study method, since 1982 to 2012. The article focus on longitudinal study methods in application, design features and data analysis. Results of sample description revealed that, application of longitudinal method in psychological study grew slowly before 2005, then increased in a continuous growth trend, researches mainly explored development of people from birth to early adulthood. For design features, researchers tend to select panel study, most of the longitudinal studies design in two or three time point measurement, with samples range from 10 to 300, and no more than three years study duration. Of the 61 studies reporting attrition, 38 studies used deletion as missing data technique;the most commonly used methods in longitudinal data analysis is hierarchical linear model, repeated measures analysis of variance, t test and structural equation model. Problems of using longitudinal method lie in small number of measurement and sample, short study duration, serious data missing and relatively old analysis method using. Attention should be paid in two folds when using longitudinal methods, type and features of design should be determined by theoretical model and research effectiveness requiring;statistical methods and missing data methods should be selected according to data characteristics.

关 键 词：追踪研究方法 设计特征 数据分析方法 文献述评 

分 类 号：B841[哲学宗教—基础心理学]