基于潜变量得分的多水平多变量回归模型在职业紧张评价中的应用  

Multilevel Multivariate Regression Model Based on Latent Variable Score and its Application on Estimation of Occupational Stress

作　　者：张岩波[1] 刘桂芬[1] 郑建中[1] 徐秀娟[1] 药红梅[1] 

机构地区：[1]山西医科大学,030001

出　　处：《中国卫生统计》2008年第2期162-164,共3页Chinese Journal of Health Statistics

基　　金：国家自然科学基金资助项目(30200236)

摘　　要：目的介绍基于潜变量得分的多水平多反应变量回归模型在职业紧张评价中的应用。方法为克服测量误差的存在,以职业紧张量表14个分项的潜变量得分,将之作为中间结果引入多水平多反应变量回归模型。结果职业任务各分项不同程度地引起职业紧张,而个体应变能力是减轻职业紧张行之有效的方式。随机系数反映这些影响在不同科室存在着不同。结论采用基于潜变量得分的多水平多反应变量回归模型既可有效降低测量误差,又得以合理地解释。尤其对于系统结构数据,多元线性模型的多水平理论比多水平潜变量分析方法更成熟可信。Objective To introduce the application of multilevel multivariate regression model based on latent variable score on estimation of occupational stress. Methods Latent scores of 14 OSI - R items were used to avoid measurement error and were substituted into multilevel multivariate regression model as the middle outcome of confirmatory factor analysis. Results Result of multilevel multivariate regression model showed that Occupational Roles and Personal Resources influenced Personal Strain with different degree. Random coefficient implied that those effects varied in different clinical department. Conclusion Multilevel multivariate regression model with latent variable score could eliminate measurement error and the results were more reasonable.

关 键 词：职业紧张 多水平模型 多反应变量回归模型 

分 类 号：R13[医药卫生—劳动卫生]