排序响应的异质交互模型研究

研究目标:扩展传统的交互模型,并估计家庭中幸福感相互影响在不同群体间的差异。研究方法:将异质特征加入传统的排序变量的社会交互模型,定量地描述外溢效应的异质性。研究发现:蒙特卡洛模拟结果表明,复合似然估计对包括异质内生系数在内的所有系数的估计具有渐进一致和渐进正态的良好性质,大样本下信息准则能够一致地选择出正确模型。应用该模型对CFPS2010的家庭数据的实证分析表明,幸福感在家庭内部存在相互影响,女性更容易受到其他家庭成员幸福感的影响,从而健康状况和教育水平对女性的幸福感影响倾向于更大。研究创新:将交互模型和影响的异质性引入到幸福感研究中。研究价值:模型为研究各领域外溢的异质性提供了分析工具。

A Study on Hetero Social Interaction Model for Ordered Response

Research Objectives:To expand conventional social interaction model,and estimate the difference of happiness spillover between various groups.Research Methods:To add hetero terms into conventional models with ordered responses,and quantitatively describe the het erogeneity of interaction.Research Findings:Pairwise composite likelihood estimate(pCL)are of good properties.According to MCMC results,all parameters,including the hetero one.are asymptotically consistent and normally distributed.Information criteria can help pick up correct model if the sample size is large enough.Using the model to study the gender differences of happiness spillover,the results show that females arc more sensitive than their male counterparts.As a result,health condition and level of education are more influential on females.Research Innovations:To com bine social interaction model and heterogeneous in fluences in happiness study.Research Value:This model provides an econometric tool to de scribe hetero spillover effect in various fields.