楼主: 雪泡xue
2684 16

[面板数据求助] 关于0-1因变量的面板想用Logit [推广有奖]

  • 0关注
  • 0粉丝

大专生

68%

还不是VIP/贵宾

-

威望
0
论坛币
551 个
通用积分
0
学术水平
0 点
热心指数
0 点
信用等级
0 点
经验
819 点
帖子
65
精华
0
在线时间
52 小时
注册时间
2010-9-4
最后登录
2021-12-20

楼主
雪泡xue 在职认证  发表于 2014-2-16 11:13:42 |AI写论文
5论坛币
面板数据想用logit,不想用xtlogit,老师建议设置行业虚拟变量缓解遗漏变量问题,可是带着行业虚拟变量运行后说是存在共线性问题,好多数据都被删掉了,这是为什么,怎么处理?

最佳答案

liuchang1989 查看完整内容

加了 你用fama-french 12 industry
关键词:logit Log 因变量 xtlogit 虚拟变量 因变量

沙发
liuchang1989 发表于 2014-2-16 11:13:43
雪泡xue 发表于 2014-2-22 14:15
那你加行业的dummy 了吗
加了 你用fama-french 12 industry

藤椅
jjjj6666 发表于 2014-2-16 11:37:25
in this case, it seems better to use random effect instead of fixed effect as suggested.  Other people may have better suggestions.

板凳
雪泡xue 在职认证  发表于 2014-2-16 11:51:44
jjjj6666 发表于 2014-2-16 11:37
in this case, it seems better to use random effect instead of fixed effect as suggested.  Other peop ...
How do you know  which type of effect I use? I donnot add fe or re.

报纸
雪泡xue 在职认证  发表于 2014-2-16 11:57:41
jjjj6666 发表于 2014-2-16 11:37
in this case, it seems better to use random effect instead of fixed effect as suggested.  Other peop ...
You mean that I should not use dummy variables?

地板
jjjj6666 发表于 2014-2-16 22:31:55
yes, it loos like the dummy variables caused the problem.  The dummy variable is used in the model as fixed effects, you can try random effect for the industry instead of putting dummy variables

7
雪泡xue 在职认证  发表于 2014-2-17 16:00:34
jjjj6666 发表于 2014-2-16 22:31
yes, it loos like the dummy variables caused the problem.  The dummy variable is used in the model a ...
How can I use random effect when I use logit.

8
jjjj6666 发表于 2014-2-17 22:29:06
I think you have to use xtlogit, but it seems you didn't want to use this approach?

9
liuchang1989 发表于 2014-2-17 22:45:43
jjjj6666 发表于 2014-2-16 11:37
in this case, it seems better to use random effect instead of fixed effect as suggested.  Other peop ...
His problem is the same as mine days ago,LOL

10
liuchang1989 发表于 2014-2-17 22:54:04

您需要登录后才可以回帖 登录 | 我要注册

本版微信群
加好友,备注jltj
拉您入交流群
GMT+8, 2025-12-9 07:08