多水平模型在大学生预防艾滋病健康教育影响因素分析中的应用

摘要

Objective The purpose of this study is to methodologically identify the factors affecting on AIDS prevention among college students using multilevel model. Methods The investigation object was extracted from four universities in Yantai city by clustering random sampling method. According to the hierarchical structure of the data, multilevel statistical model was used to fit the data. Results The data on students' knowledge of AIDS had hierarchy. The data was anlysised respectively by two level logistic regression and single-level logistic regression, gender, grade,the source of students in the impact of variables on the knowledge score did not show statistically significant ( P > 0. 05). the knowledge and behavior scores showed a positive correlation ( r = 0. 006, P < 0. 05). Conclusion Multilevel model analysis considers the data hierarchy, the obtained result is more reliable. The knowledge score and AIDS risk behaviors is positive correlation,but this correlation is relatively limited, health education on AIDS should not only focus on the improvement of knowledge.%目的 探讨多水平模型在烟台市大学生艾滋病知信行影响因素中的应用.方法 采用整群随机抽样方法在烟台市4所高校抽取在校大学生作为调查对象,根据大学生艾滋病知识得分数据层次结构的特点,采用MLwiN 2.02软件进行多水平模型拟合.结果 大学生艾滋病知识数据存在聚集性,学校、学生两水平logistic回归和单水平logistic回归分析显示,性别、年级、生源地对大学生艾滋病知识得分的影响无统计学意义(P＞0.05).双变量二分类反应变量模型拟合显示:知识和行为得分在学生水平上存在正相关(r =0.006,P＜0.05).结论 多水平模型分析考虑了数据误差的层次性后,获得的结果更可靠；知识得分与艾滋病个人高危行为的相关关系比较有限,艾滋病的健康教育不应只注重知识的提高.