How to combine the putexcel with significance symbol star

[Eric Li]

Dear all,

I would like combine the putexcel code with significance symbol star(*** ** *).

Does anyone know how to yield style B instead of style A in using putexcel code?


putexcel B4=(_b[X1])

putexcel B5=(_b[X1]/_se[X1])

[Kevin Crow (StataCorp)]

Eric,

You will have to create the number with the significance symbol yourself before using putexcel. For example,

sysuse auto, clear
regress mpg weight turn
local coeff = _b[turn]
local coeff : display %5.3f `coeff' . "***"
putexcel set test, replace
putexcel A1 = "Coeff"
putexcel B1 = "`coeff'"

[Wameq Raza]

Hi Kevin,
Is the only way to do this manually, i.e., individually for each coefficient? That seems really cumbersome! Is there a way to automate the process?
Thanks and best,
Wameq

[Attaullah Shah]

asdoc can be used to export regression tables in three different styles. The one that matches the Stata output is the detailed regression table. For example, the following one line of code will generate the given table.

* Install asdoc
ssc install asdoc

* Load example data set
sysuse auto, clear

* Estimate the regression and send it to MS word in one click
asdoc reg price mpg rep78 headroom trunk weight length turn , replace

You can read this blog post for more examples and variations http://fintechprofessor.com/2018/12/...ing-stars-etc/

[Weiwen Ng]

Not to take anything away from asdoc, but estout (SSC, written by Ben Jann, some info on his website) is also a good program. Exporting a matrix of correlation coefficients is easier to do in putexcel than in estout, and there are probably some other edge cases (especially for commands that don't write to e(b) and e(V)).

[Attaullah Shah]

Dear Weiwen Ng I have not used estoutmyself, I think estout produces a summarized table of regression. If that is true, then asdoc differs from estout in a sense that asdoc offers three different types of regression tables.
1. Detailed regression tables, like the one shown in the #2 above
2. Nested regression tables such as those produced by the estout and outreg2
3. Wide regression tables (good for regressions over groups)

some video examples are here
https://www.youtube.com/watch?v=rkUU7UiygBU
https://www.youtube.com/watch?v=61ks3cMPz3c