Interpreting regression coefficients in terms of standard deviation

[Samyam Shrestha]

Hi!

I feel it's more of an econometrics/statistics question, but I would also want to learn how I can do that on Stata.

I often see findings interpreted in terms of standard deviation, like in the following two examples:

"[A shock] led to the index value being 0.15 standard deviations lower than those of similar individuals who did not experience the shock."

or

"[A pollution shock] on exam days cause declines in average exam performance of roughly 2-6% of a standard deviation compared to exams taken during less polluted days."

Given my regression table in the following picture, how do I do interpret the results in that way (say, specifically column (8))?

I heard from someone that I need to do the following:

-0.210 / 1.161 * 100 = ~-18%

Where -0.210 is the coefficient in column (8) regression, and 1.161 is the standard deviation of PC1, the dependent variable.

And, therefore, interpreting the results in this way: one unit increase in MMI leads to 0.18 standard deviations lower?

Is this the right way of doing this?

Also, if this is not the right way, is there a code in Stata to do this the right way?

Thanks a lot!

[FernandoRios]

You can do that in two ways.
1. after regression, look into the "beta" coefficients options. Any introductory book on econometrics will show you how they are estimated and why they are useful
2. If you want to keep your original values as they are (possibly something you want for dummy variables), you can simply standardize the dependent variable.
HTH
Fernando.