基于HGLMs模型的未决赔款准备金评估方法

本文将HGLMs模型用于未决赔款准备金的评估,应用R软件对未决赔款流量三角形中的增量赔款进行建模,得到未决赔款准备金的估计值和预测误差。这一方法的应用考虑了同一事故年赔款数据反复观测的纵向特征和不同事故年未观测到的特征所导致的异质性,且与贝叶斯方法得到的最佳估计结果非常接近。

影响学生参加数学课外补习的因素——以中国上海和日本大都市为例

选取PISA2012数据,采用logit模型和HGLM模型研究影响中国上海和日本大都市学生参加数学课外补习的因素。研究发现:影响两个国家(地区)学生参加数学课外补习的因素主要为学生层面的因素,具体而言,女性的数学课外补习参与率高于男性,独生子女学生数学课外补习参与率高于非独生子女学生;ESCS和家庭教育资源充裕程度对学生参加数学课外补习有正向预测作用;数学成绩对数学课外补习参与的影响不确定。

Using Hierarchical Generalized Linear Modeling to Examine Contextual Differential Item Functioning:A Validity Study of Teachers'Sense of Efficacy Scale

The purpose of the study was to further investigate the validity of the instrument used for col-lecting preservice teachers'perceptions of sel-fficacy adapting the three-level hierarchical generalized linear modeling(HGLM)model.To serve the purpose,the study used data collct-ed by the research team which elicited preservice teachers'self-efficacy beliefs using Teachers'Sense of Efficacy Scale(TSES).HGLM were used to analyze the data.Results of the HGLM analyses(at level-two)showed that one item in the scale displayed gender DIF.Another item became DIF item when the context variable was added to the level-two model.However,the effect of the context on the DIF item was not big.