求教：我想用一个二元选择模型

感觉还是少些

[此贴子已经被作者于2007-6-11 18:25:29编辑过]

应该再增加一些

hi, you can use logit with the panel data. It is better but still the sample is not big enough because more parameters should be estimated.

And I think as a policy analysis, three independent variables seem too few, you can always think of other factors affect y or both xs and y. But anyway, you can try this as a good practise.