Programming the statistical procedures from SAS

Correlation between variables

Accepted Solution Solved
Reply
Contributor
Posts: 69
Accepted Solution

Correlation between variables

Hello,

 

I am asking for your advice,

I have two independent variables with correlation of 0.5. The variables are as follows:

(1) Family ownership, which refers to the family % in ownership

(2) Founder, dummy variable with 1/0 if the manager is the founder

 

The correlation is 0.5 as in some firms this is the same person. If I want to check the "PURE" impact of each variable on the dependent variable, what you suggest me to do? (PURE means, for example, the real impact of being founder w/o the ownership too).

My first idea (that also checked) was to run the regression of founder impact only between similar ownship groups, but in this way I am loosing a lot of data that is not used and might help.

 

Can you please give any clue how to handle with it?

 

Thanks a lot! 


Accepted Solutions
Solution
‎05-04-2017 08:27 AM
Trusted Advisor
Posts: 1,671

Re: Correlation between variables


If I want to check the "PURE" impact of each variable on the dependent variable, what you suggest me to do? (PURE means, for example, the real impact of being founder w/o the ownership too).


In my opinion, this cannot be estimated from the correlated data you have. In other words, the question is meaningless, and any attempts to answer the question exactly as asked are a waste of time.

 

The best you can do is come up with a model that involves both variables predicting the dependent variable, and not try to determine any "pure" impact.

View solution in original post


All Replies
Solution
‎05-04-2017 08:27 AM
Trusted Advisor
Posts: 1,671

Re: Correlation between variables


If I want to check the "PURE" impact of each variable on the dependent variable, what you suggest me to do? (PURE means, for example, the real impact of being founder w/o the ownership too).


In my opinion, this cannot be estimated from the correlated data you have. In other words, the question is meaningless, and any attempts to answer the question exactly as asked are a waste of time.

 

The best you can do is come up with a model that involves both variables predicting the dependent variable, and not try to determine any "pure" impact.

☑ This topic is solved.

Need further help from the community? Please ask a new question.

Discussion stats
  • 1 reply
  • 112 views
  • 1 like
  • 2 in conversation