2
$\begingroup$

I was reading in the literature that in order to compare linear regression results with independent variables & categorical (variables), it is recommended to standardize the coefficients to a zero mean.

I am trying to do linear regression where Y in my model comprises birth rate, X1 the monthly temperature, and 11 dummy variables representing Jan-Nov. While running it in Microsoft Excel, I am not sure whether the coefficient values are standardized or not. Is there a way to check it and is it possible to normalize coeff. to a zero mean in Excel?

$\endgroup$
  • $\begingroup$ Although this question is asking for Excel solutions, there are underlying statistical issues that need to be addressed (& which, when addressed, will take care of the problem). This should be considered on topic here, IMO. $\endgroup$ – gung - Reinstate Monica Oct 25 '15 at 17:30
1
$\begingroup$

I don't think you have to standardize the coefficients- you have to standardize the regressors. In that sense, you have to take your regressors, de-mean them (subtract the mean), and then divide by the standard deviation. Now, you coefficient $\beta$ measures the marginal effect of $x$ on $y$ in standard deviation units, is therefore unitless, and can be compared across regressors.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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