Standard deviations of variable after fixed effects are removed

[Charlotte Ferrai]

Good afternoon,
I am using panel data and I'm trying to calculate standard deviations for my independent variables after different types of fixed effects are removed. Is there an easy way to do that in Stata?
Many thanks,
Charlotte

[Clyde Schechter]

I am using panel data and I'm trying to calculate standard deviations for my independent variables after different types of fixed effects are removed.

It isn't clear to me what this means. Can you be more specific? Perhaps give a concrete example of what you want to do.

[Charlotte Ferrai]

As in this paper - p.462, far right column of Table 1
http://www.sas.rochester.edu/psc/cla...canGitting.pdf

[Carlo Lazzaro]

Charlotte:
just like Clyde, I'm not clear with what you're after: calculation or table formatting as in the paper you mentioned?

[Clyde Schechter]

Sorry, but the authors of that paper do not explain what they did, and I don't understand it. If this is important, I would suggest you contact the authors for more information.

[Charlotte Ferrai]

Clyde, Carlo, thank you for your answers. What I am trying to do is to calculate the standard deviation of a variable, after removing country- and region*year fixed effects. I have seen it done in a few academic papers but none of them explain how they computed the SDs.

[Clyde Schechter]

Charlotte, you are being as clear as the authors of that paper--which is the most anyone can ask of you, but which is not clear at all I do not understand what it means to remove effects from an independent variable in the model. The paper you linked to is the first I have ever seen that, and they do not explain what it means,nor how it's done. The difficulty is not with you, it is with them.

[Carlo Lazzaro]

I do share Clyde's point: I've never read/heard of anything similar (in health economics/econometrics, at least).

[Charlotte Ferrai]

What I am trying to do is to interpret results from a panel regression in terms of standard deviations: I worry that I cannot make claims about the effect of a one standard deviation increase in the independent variable if my calculations for these standard deviations do not remove the fixed effects..

[FernandoRios]

Hi Charlotte,
If I understand what you say correctly, one option could be to "detrend" your dependent variable.
In other words, create a new variable that is what is left after taking the fixed effects.
If you download the following program from the ssc library (ssc install regxfe)
there is a subprogram itercenter, that take the fixed effects from your variable. It, however, does not account for trends.
Hope that helps
Fernando