两水平方差成分模型在长沙市高校大学生健康素养影响因素分析中的应用  

Analysis of health literacy influencing factors among college students in Changsha by two-level variance component model

作　　者：牟劲松[1] 李洪飞[1] 谭爱春[1] 欧阳江[1] 赵淑英[1] 陈欢[1] 

机构地区：[1]长沙医学院公共卫生学院,湖南长沙410219

出　　处：《中国健康教育》2016年第5期408-413,共6页Chinese Journal of Health Education

基　　金：湖南省教育厅科学研究项目(15C0143)

摘　　要：目的探讨长沙市在校大学生健康素养水平的影响因素,比较传统多元回归模型与两水平方差成分模型的分析结果。方法采用分层整群随机抽样的方法,使用自行设计的大学生健康素养调查问卷对调查对象进行调查;采用Mlwi N 2.03软件对资料进行两水平方差成分模型拟合。结果健康素养在班级水平上存在聚集性(健康知识、行为、技能素养的W值分别为940.860、74.462、45.388,ρ值分别为1.000、0.424、0.347);两水平方差成分模型获得大学生健康素养影响因素包括:学校、性别、专业、父亲文化程度、母亲文化程度、年龄。结论专科、男性、非医学专业、父母文化程度为不识字/少识字、年龄较大的大学生健康素养较低,多水平模型拟合含有层次结构数据时比传统模型的拟合结果更理想。Objective To explore the influencing factors of health literacy among college students in Changsha, and then compare the results between the traditional multiple linear regression model and two-level variance component model. Methods The investigation participants were selected by stratified cluster random sampling and surveyed with a self-designed questionnaire of health literacy. Data were analyzed by a two-level random intercept muhinomial logistic regression model. Results There were class aggregation in health literacy （the W value of health knowledge, behavior and skill literacy separately were 940. 860, 74. 462, 45. 388, p value were 1. 000, 0. 424, 0. 347, respectively）. The two-level variance model analysis suggested that the intensity sequence of factors were school type, gender, major, parents＇ education level and age. Conclusion The health literacy level of college students who were junior college, male, non-medical major, older, and their parents＇ education degree were illiterate/semi-literate were relatively lower. Multilevel statistical model fitting with hierarchical structure data is more ideal than multiple linear regression model.

关 键 词：健康素养 大学生 影响因素 两水平方差成分模型 

分 类 号：R193[医药卫生—卫生事业管理]