Hi all SAS Experts,
When regressing nonexperimental data, I faced the Simpson Paradox. Simple description from this post:
Imagine
You run a linear regression with four numeric predictors (IV1, ..., IV4)
When only IV1 is included as a predictor the standardised beta is +.20
When you also include IV2 to IV4 the sign of the standardised regression coefficient of IV1 flips to -.25 (i.e., it's become negative)
In my case, when I run the linear regression:
y=b1x1+b2x2+b3x3.
b1 get the positive value. However, when I add the variable x4 to the regression, b1 becomes negative.
I look at a comment of this topic,one of the reason maybe due to the multicollinearity among variables.
I am wondering how to check whether Simpson's paradox in my case is caused by multicollinearity? How to generate the multicollinearity table among variables using SAS?
Many thanks and warm regards.
... View more