I know I can use the b() option in the community-contributed command esttab to control the number of decimal points displayed in the rows (i.e. the regressors):
estimates clear
eststo clear
sysuse auto, clear
eststo w1: regress price mpg trunk length
eststo w2: regress turn mpg trunk length
eststo w3: regress displacement mpg trunk length
esttab, b(1 3 5)
All of the decimal points in the first row (mpg) have one decimal point. The second and third rows have three and five decimal points.
However, instead of controlling the decimal points by row, I want to control the number of decimal points by column (model).
Is there a way to do this?
For example, I want all of the regressors in the first column to have one decimal point, all of the regressors in the second column to have three decimal points, and all of the regressors in the third column to have five decimal points.
You cannot do this directly but a workaround is to use a matrix.
A simple example is the following:
matrix A = ( -173.70800, -0.06556, -1.77658 \ ///
-0.85469, -0.05926, 0.06587 \ ///
21.40414, 0.16548, 3.06799 \ ///
5853.99300, 10.76202, -342.34697 )
matrix rownames A = mpg trunk length _cons
esttab matrix(A, fmt(1 3 5)), gaps mlabel(none) ///
collabels("price" "turn" "displacement")
---------------------------------------------------
price turn displacement
---------------------------------------------------
mpg -173.7 -0.066 -1.77658
trunk -0.9 -0.059 0.06587
length 21.4 0.165 3.06799
_cons 5854.0 10.762 -342.34697
---------------------------------------------------