4
$\begingroup$

I came across this claim in the Statistical Analysis section of a clinical research article.

GEE was used to derive standardized regression coefficients, which in any one regression equation are measured on the same scale, with a mean of 0 and a standard deviation of 1. They are then directly comparable to one another, with the largest coefficient indicating which independent variable has the greatest influence on the dependent variable ([26])

[26] Neter J, Kutner MH, Nachtsheim CJ, Wasserman W. Applied Linear Statistical Models. Chicago: McGraw Hill/Irwin, 1996

I cannot find an easily way to link to this citation. Is the bolded part(about influence) necessarily true?

$\endgroup$

1 Answer 1

4
$\begingroup$

Yes, it is true. And, I can't think of an exception when that would not be the case. But, you have to watch out about your data set, and when you test your model against new data. Let's say variable A has a standardized coefficient of 1.0 and variable B has one of 0.75. So, variable A is more influential than variable B.

However, in your new data set variable B is much more volatile (higher standard deviation) and variable A is relatively more stable. In this case, variable B will have more impact than variable A. And, that is not because it is more influential, but because it is moving around much more than variable A within the new data you are testing your model with. However, this does not contradict the statement in bold.

Standardized coefficients are extremely useful because they do put all variables on the same scale (the unit is the standard deviation). And, they should be used a lot more often than they actually are.

$\endgroup$
1
  • $\begingroup$ Upvoted for the clarity and excellence of the answer. $\endgroup$ Sep 4, 2019 at 1:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.