Absolute coefficient of variation as a new variable

[OF Buyukkurt]

Dear Forum Users;

I have been trying to calculate the absolute coefficient of variation for two different variables and need to keep the results as new variables to be used in the regression.

The following command seemingly works yet I cannot store the outcome as a variable.


tabstat var, statistics(n mean cv) by(group_var1)


Could you please help me about how to keep the results as a new variable?

I hope my question makes sense to you.

[William Lisowski]

Since the coefficient of variation is the standard deviation divided by the mean, the following might start you in a useful direction.

sort group_var1
by group_var1: egen m = mean(var)
by group_var1: egen s = sd(var)
generate cv = s/m

[OF Buyukkurt]

It perfectly worked. Thank you very much, William.

[Nick Cox]

#2 and #3 seem to settle the matter, but what is the absolute coefficient of variation? The coefficient of variation isn't obviously useful or well defined unless all values of a variable are positive.

[OF Buyukkurt]

#2 and #3 seem to settle the matter, but what is the absolute coefficient of variation? The coefficient of variation isn't obviously useful or well defined unless all values of a variable are positive.

Sorry for the extremely late reply. I've employed the so-called "Eckel Model" to measure income smoothing. The model simply divides the absolute value of variance coefficient change in income from t-1 to t with the absolute value of variance coefficient change in sales from t-1 to t.

[Nick Cox]

That does not sound the same as the coefficient of variation to me.

[OF Buyukkurt]

That does not sound the same as the coefficient of variation to me.

It was my bad then since I couldn't express my question clearly. I tried my best but I think it happened because of the lack of comprehension of the model...
I thought I've completed the study but I just wonder how you think about the above code. Do you think it serves the purpose?