Export correlation matrix

[Sophie Klein]

Dear all,

I am trying to Export my pwcorr Matrix to word. However, all commands I found and tried so far do not work.


So far I tried:

pwcorr $ylist $xlist, star(.05)

esttab using Correlations.rtf, unstack compress b(2)

Please can anybody suggest some help on this?

Thank you very much.
Best, Sophie

[Weiwen Ng]

Sophie, see if this example does it?

[Rob Sukaldi]

If Weiwens solution doesn't help you, you can try the following:

estpost pwcorr $ylist $xlist, star(.05)
eststo correlation
esttab correlation using Correlations.rtf, unstack compress b(2)

I know you mentioned that esttab doesn't work, but I don't know whether you saved the correlation matrix.

[Weiwen Ng]

If Weiwens solution doesn't help you, you can try the following:

estpost pwcorr $ylist $xlist, star(.05)
eststo correlation
esttab correlation using Correlations.rtf, unstack compress b(2)

I know you mentioned that esttab doesn't work, but I don't know whether you saved the correlation matrix.

It's Ben Jann's solution, not mine. But your observation about saving the correlation matrix is spot on. To the original poster: the esttab and estout commands work with e(b) and e(V) matrices - these are matrices produced by estimation commands. e(b) refers to the parameter estimates, and e(V) to the variance-covariance matrix of the parameters. e(V) gets you the confidence intervals, p-values, and stars in Stata's own tables and in estout.

Stata normally posts the output from -pwcorr- to r(C). You have to get it into an e(b) matrix if you want to use -esttab-, which is what -estpost correlate- or -eststo- do.

[Sophie Klein]

Thank you everybody, that gave me a lot of insight!

I tried the estpost corr command, then store and then esttab, it works perfectly!

[Branon Xerxes]

If Weiwens solution doesn't help you, you can try the following:

estpost pwcorr $ylist $xlist, star(.05)
eststo correlation
esttab correlation using Correlations.rtf, unstack compress b(2)

I know you mentioned that esttab doesn't work, but I don't know whether you saved the correlation matrix.

I tried to replicate this for my own work. Here is the necessary code

eststo clear
estpost pwcorr tr1 tr2 tr3  tr4a tr4b tr4c 
eststo corrtr
esttab using "$texpath/tr_corrtable.tex", replace b(3) p lines not noobs nonum booktabs unstack compress ///
star (c 0.10 b 0.05 a 0.01) 
eststo clear

However after the second line, Stata returns the following

. eststo clear

. estpost pwcorr tr1 tr2 tr3  tr4a tr4b tr4c 
invalid subcommand
r(198);

end of do-file

r(198);

If I run the same without estpost , I would get the following output (which is what I am trying to export)

.  pwcorr tr1 tr2 tr3  tr4a tr4b tr4c

             |      tr1      tr2      tr3     tr4a     tr4b     tr4c
-------------+------------------------------------------------------
         tr1 |   1.0000 
         tr2 |  -0.0779   1.0000 
         tr3 |   0.6476  -0.0801   1.0000 
        tr4a |  -0.0532  -0.0047  -0.0940   1.0000 
        tr4b |  -0.0635   0.0104  -0.0714        .       1.0000 
        tr4c |  -0.0842   0.0472  -0.1476   0.1882   0.0466   1.0000

I should note that the error is not because there are missing values for (tr4a, tr4b) . This happens even when I restrict the table to only tr1, tr2 and tr3.
Thanks.

[Weiwen Ng]

I tried to replicate this for my own work. Here is the necessary code

eststo clear
estpost pwcorr tr1 tr2 tr3 tr4a tr4b tr4c
eststo corrtr
esttab using "$texpath/tr_corrtable.tex", replace b(3) p lines not noobs nonum booktabs unstack compress ///
star (c 0.10 b 0.05 a 0.01)
eststo clear

However after the second line, Stata returns the following

. eststo clear

. estpost pwcorr tr1 tr2 tr3 tr4a tr4b tr4c
invalid subcommand
r(198);

end of do-file

r(198);

...

So, I didn't notice the issue before, but the documentation does indicate that -pwcorr- is not a valid sub-command. -estpost- is like a wrapper calling one of several admissible commands, and it will accept -correlate-, but not -pwcorr-. However, the documentation also says that -estpost correlate- does this:

estpost correlate varlist [if] [in] [weight] [, matrix nohalf print(#) bonferroni sidak listwise casewise quietly esample ] posts the pairwise correlations between the first variable in varlist and the remaining variables. Alternatively, if the matrix option is specified, all pairwise correlations among the variable in varlist are posted.

[Ziyun Deng]

Just wanted to say a big Thank You to everyone above. Your questions and answers are super helpful! This saves me so much time!

[Attaullah Shah]

You might also be interested in the one-liner solution of asdoc for exporting correlation matrix to Word. Watch this YouTube Video.

For those who are not yet familiar with asdoc, asdoc can be downloaded from SSC and can be used with almost all Stata commands. Here is a short blog post that shows how asdoc can be used with any Stata command http://fintechprofessor.com/2018/02/...basic-example/. You can also watch several YouTube videos that show the use of asdoc https://www.youtube.com/watch?v=zdI6...LwodAk2oqLYhr-

* For installation of the stable version 
ssc install asdoc

* For installation of the new beta version
net install asdoc, from(http://fintechprofessor.com) replace
help asdoc

[Attaullah Shah]

The new version of asdoc has been uploaded to the SSC. Details of the updates are given in this post
https://www.statalist.org/forums/for...ord-from-stata

[Ruggero Jappelli]

estpost correlate varlist, matrix listwise
est store storagename
esttab * using filename.tex, unstack not noobs compress

[Giuseppe Polito]

If Weiwens solution doesn't help you, you can try the following:

estpost pwcorr $ylist $xlist, star(.05) eststo correlation esttab correlation using Correlations.rtf, unstack compress b(2)

"estpost pwcorr" gives me an "invalid subcommand" error.

[Giuseppe Polito]

The new version of asdoc has been uploaded to the SSC. Details of the updates are given in this post
https://www.statalist.org/forums/for...ord-from-stata

asdoc writes documentation that is false.

In particular, the following code:

asdoc pwcorr var1 var2 var3, star(0.05)

writes a .docfile which contains the following:

*** p<0.01, ** p<0.05, * p<0.1

This is not true: 1 star indicates a 5% significance level, not 10%, and there are no ** nor ***

[Attaullah Shah]

I think you are referring the to the table note reported by asdoc at the bottom of the table. Since asdoc allows significance stars for three levels (Stata allows only one) with the pwcorr command, the table note always remain the same, that is

*** p<0.01, ** p<0.05, * p<0.1

The note was meant for use with the star(all) option (see example below). As you have reported, this might be misleading when someone wishes to report the significance stars for a single level, in which case only one star is reported. I shall fix the table note in a week or so.

Example of using option star(all) option.

sysuse nlsw88
asdoc pwcorr tenure  race, sig star(all) replace

[Attaullah Shah]

I found time today and fixed the table note. The new version of asdoc can be installed from my site. Copy and paste the following line in Stata and press enter.

net install asdoc, from(http://fintechprofessor.com) replace
asdoc pwcorr tenure  race, sig star(.05) replace

Please note that the above line has to be copied in full. After installation of the new version, then restart Stata.


asdocx is now available

---

A more powerful and flexible version of asdoc is now available. I call it asdocx. You may like to check the details here

https://fintechprofessor.com/asdocx


Please do remember to cite asdoc. To cite:

In-text citation
Tables were created using asdoc, a Stata program written by Shah (2018).

Bibliography
Shah, A. (2018). ASDOC: Stata module to create high-quality tables in MS Word from Stata output. Statistical Software Components S458466, Boston College Department of Economics.