How to export polychoric matrix results

[Odile Mackett]

Hello everyone

I have now searched far and wide for an answer to this question. I am using polychoric matrices in a paper that I am working on and want to include the results of the matrices in my paper (in a matrix format). I cannot find a way to export the matrices in a way that is easily formatted in a Word or Excel document (or even a text document for that matter).

Are there any suggestions on how I can do this, please? I usually use the estpost/estout commands when using the normal corr matrices, but it does not seem to work with polychoric correlation matrices.

Kind regards
Odile

[Jared Greathouse]

I've never heard of this, so you're gonna need to show a data example using dataex and show the precise syntax you've tried so far

[Mike Lacy]

One can assign to a regular Stata matrix the r(R) return matrix that the user-written command -polychoric- creates, and I'd presume that -estout- (which I don't happen to use) would allow export of it in some form like -estout matrix(YourMatrix)-. (Note that -estout- apparently doesn't like working with matrices in the return list, whether it is the r(R) from -polychoric- or the r(C) from -corr-.)

One could also use -putexcel-:

putexcel set YourOutputFile.xls
putexcel A1 = matrix(r(R)) //  -putexcel- is ok with a return list reference

[Odile Mackett]

Thank you so much Mike!

With putexcel, it comes out without the variable labels, but the correlations are correct.
With estout(matrix), everything comes out perfectly.

Thank you so much for your assistance!!

Kind regards
Odile

[Felix Kaysers]

Short intro to save the return matrix and exporting it using estout for all non-proficient Stata users like me:

*1. Download dataset
use https://www.stata-press.com/data/r17/nlsw88.dta
* 2. Run polychoric command
polychoric age race married industry occupation union
* 3. Assign the return matrix from polychoric to a Stata matrix
matrix correlations = r(R)
* 4. Use the estout package to export this matrix. The only problem is that the matrix fills every cell, so you will have to manually delete the correlations duplicates.
estout matrix(correlations)