Thursday, December 29, 2011

Causal Relationship Between two Independent Variables: Is one a Confounding Variable?

Two of the independent variables in my study have a very high correlation, common sense. One is age, and the second is years driving license has been held for. Along with other independent variables, I'm planning on running a regression analysis. Due to the causal nature, I'm considering leaving out years driving license has been held for. Could years of license be considered a confounding variable, or is it called something else? Do I have a good reason for not including it?|||It's not really a causal variable or a confounding variable. It's an example of collinearity. You do not need to exclude it but might try standardizing ((X-mean_X)/SD_X) or other sensible transformations. Depending on your application it might make sense to use age at time of licensing instead of age, since combined with years licensed you would get total age.

No comments:

Post a Comment