基于多项相关系数的结构方程模型在大学生自我学习期望影响因素研究中的应用  

A Structure Equation Model Based on The Polychoric Correlation Coefficient: An Application to Associated Factors of The College Students' Self-Learning Expectancies

作　　者：周映雪[1] 欧春泉[1] 赵智涛[2] 彭成华[2] 陆梦洁[1] 

机构地区：[1]南方医科大学公共卫生与热带医学学院生物统计学系,510515 [2]南方医科大学教务处

出　　处：《中国卫生统计》2013年第4期529-531,539,共4页Chinese Journal of Health Statistics

基　　金：广东省大学生创新实验计划项目(1212111029)

摘　　要：目的构建基于多项相关系数的结构方程模型,探讨大学生自我学习期望的影响机制,为有序分类数据的结构方程模型分析提供参考依据。方法本研究借助AMOS和R语言,对广州某重点医科大学新生入学调查数据进行基于多项相关系数的结构方程模型分析。结果在验证性因子分析中,基于Pearson相关的因子载荷均明显低于基于多项相关的因子载荷;学术自我效能、学术目标、社会目标对自我学习期望均有正向影响(P<0.05);学术自我效能、学术目标、社会目标、经济目标两两间正相关(P<0.05)。结论分析有序分类数据,基于多项相关系数的结构方程模型能获得更准确的模型。大学生提高学术自我效能、学术目标、社会目标有助于提高自我学习期望,而经济目标可能通过学术目标、社会目标间接影响自我学习期望。Objective To establish a structure equation model(SEM) based on the polychoric correlation coefficient for exploring the influence mechanism of the college students' self-learning expectancies.The study would provide reference for analyzing ordinal categorical data.Methods We used AMOS and R language to build SEM based on the polychoric correlation for analyzing survey data of precollege freshmen in a medical university in Guangzhou.Results Confirmatory factor sanalysis shows that factor loading based on Pearson correlation is significantly lower than that based on polychoric correlation.Academic self-efficacy,academic goals and social goals have positive impacts on self-learning expectancies.There is significant positive correlation among these factors.Conclusion Our study indicates that a SEM based on the polychoric correlation offers more accurate results when analyzing ordinal categorical data.For college students,raising academic self-efficacy,academic goals,and social goals could improve self-learning expectancies,and economic goals may influence self-learning expectancies indirectly through academic goals or social goals.

关 键 词：多项相关系数 结构方程模型 自我学习期望 

分 类 号：G642.4[文化科学—高等教育学]