本研究调查北京某大学学生的抑郁现状，并验证前人研究所指出的导致抑郁的原因。综合考虑学习压力、就业压力、家庭经济、与父母交流情况、人际关系和社会支持6个潜在变量，建立含有序分类变量的结构方程模型。本文采用Gibbs抽样方法得到模型参数的贝叶斯估计，并通过残差分析和后验预测p值对模型拟合优度进行评价。之后考虑引入潜变量间的交互作用项改进模型，通过偏差信息准则进行模型间的比较。 由潜变量结构方程的参数证实了学习压力、就业压力、家庭经济状况不好以及人际关系处理不当、对社会支持使用较少都会加重抑郁程度。由测量方程因子载荷矩阵的估计值可知，就业压力对该校学生抑郁程度影响最大，其次是人际关系，学习压力和就业压力的共同作用，接下来是学习压力，社会支持，与父母的情感交流，人际关系和社会支持的共同作用，最后是家庭经济状况。 考虑学校环境、家庭环境以及人际关系等因素的可调试性，本文针对性地提出以下建议：第一，教师要引导学生创建互助、和谐的集体；第二，大学生要适当调整自己的就业期望值；第三，大学生应注重自身业余爱好的培养；第四，激发学生对所学专业的热爱；第五，着重关注贫困生的心理健康问题，并提供适当的困难补助。

The purpose of this paper is to explore the prevalence of depression among undergraduates and its related factors, so as to provide effective information for preventing depression. Taking six potential variables such as academic stress, employment pressure, domestic economic situation, family communication, interpersonal relationship and social support into account, we establish structural equation modeling with ordered categorical variables. Gibbs sampling is used to obtain Bayes estimation of structural parameters. A goodness-of-fit statistics given via sampling is proposed to assess the plausibility of the posited model. And through the predictive results of test sample to verify the validity of model. Results reveal that six potential variables above were significant predictors of depression. Employment pressure has the greatest effect on undergraduates’ depression, followed by interpersonal relationships, study stress, social support, family communication, the last one is domestic economic situation. From the structural equation, we get comprehensive evaluation system of depression in college students. Considering the controllability of these factors, we put forward some suggestions of prevention and controlling of college students' depression, hoping to provide theoretical knowledge for further research.