嵌套logit模型在Stata软件中的拟合

FITTING OF NESTED LOGIT MODEL IN STATA SOFTWARE

[目的]掌握嵌套logit模型的原理及其在Stata软件中的拟合过程。[方法]运用实例来介绍该模型的拟合过程。[结果]在Stata软件中的拟合过程中,需注意数据结构的正确性、构建模型中的各层次及解释变量、检验IIA假设,最后采用nlogit模块进行嵌套logit模型的拟合。[结论]嵌套logit模型可以解决反应变量的各个类别之间的相关性问题,正确分析因素对反应变量的影响效应。

[Objective] To understand the principle and fitting process of nested logit model. [Methods] The fitting process of this model wasintroduced via an example. [Results] The validity of data structure, establishment of each level and explanatory variables and test of IIA hypothesis must be before fitting explanatory variables and test of IIA hypothesis must be finished the nested logit model via nlogit module during the overall fitting process in Stata software. [Conclusion] Nested logit model can deal with the problem of dependencies of each categories; which helps to correctly analyze the effects of factors on dependent variable.