渴望大神学霸指点指点

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

求大神指点模型是否可识别，数据的文件应该是什么样的

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：可识别 模型

Hi buddy,
I wish we can figure it out together.

Based on the limited information, I assume that it is difficult to build a quantitative model.
The reason is following,
1. it is hard to measure the following criteria:
e15, e16,e17,e18
2. it is also hard to get complete e19 information
3. you may access to e20 and e21, but the cost may beyond its benefit

Another issue,
the dependent variable you mentioned is more like a fraud.
If you want to build a model anticipate it, there is a famous model.
The independent variables are:
financial pressure to XXX
opportunity to XXX
rationalization of unprofessional conduct (e.g. XXX).

Facing three variables, your load will be much smaller.
And the data may be easier to collect.

e15,e16 and e17 mainly describe CEO or business owner's personal interest.
You may combine them as one.
It will be much simpler.

Another issue,
the words you use is very colorful, e.g. "群众\曝光"
Academic world prefer to neutral style.
"Whistle-blower, public monitoring" will be much professional.

The last but not the least,
I don't think regulation is one decision variable.
Because they are there no matter your choose to nondisclosure or disclosure your accident.

I am a student too,
just want to discuss with you about your model.
I hope it helps.
Good luck.

cpamodeler 发表于 2014-6-26 03:41
Hi buddy,
I wish we can figure it out together.

十分感谢，你提出的意见都很宝贵，如果抛开信息的获取难度来看，如果数据是全面的，模型能运行吗？还有，请问你有没有Amos的因果模型的分析数据和实例源文件，我想看一看，自己操作下，第一次接触这个软件，十分感谢

Thank you for your update buddy. :)

With sufficient data, you can definitely build your model.
You may screen your decision variables with their p-values.

I am sorry, I am not familiar with Amos.
As far as I know, spreadsheet can get the job done.

With identical data, normally speaking, different software will generate identical result.
Good luck buddy. :)   

cpamodeler 发表于 2014-6-26 11:30
Thank you for your update buddy. :)

With sufficient data, you can definitely build your model.

thank you very much!!!