Pearson/Spearman correlation table (corsp) formatting error and transfer to Excel

[Alina Sigel]

Dear Statalist members,

I am using Stata 14.2 on a Mac.

I computed a correlation matrix with Pearson's correlation coefficients in the lower triangle and Spearman's rank correlations above the diagonal using the user-written command corsp by Daniel Klein.

Additionally, I chose the option to report p-values below each correlation coefficient.

My first problem: After having chosen to report p-values below each correlation coefficient, the vertical variables' names no longer fit the table. (See attached screenshot)

My second problem: I am not sure how about how I can transfer the full correlation table to Exel (I need to report in in Word later on). I am currently using the command esttab to transfer regression results to Excel.

Any help is highly appreciated!

Alina

[daniel klein]

For the record, corsp is (probably) from SSC.

The first problem is a bug in corsp (thanks for reporting) that I have fixed. I will send the update to Kit Baum so it should be up on SSC in a couple of days. For now, use the (old) option pvalues instead of sig. In the example above

corsp TREAT MVln ROA  BTM LEVERAGE FOLLOW , pvalues

This will also help with the second problem, as it will cause corsp to also return the combined matrix of coefficients and p-values in r(RP). This matrix is hidden in Stata version 12 and higher so you will need to type

return list , all

to see it.

I cannot comment on how to export this matrix to Excel with esttab (where is it from?). You might want to have a look at the putexcel command. I have never used it, so I cannot comment on this option either.

Best
Daniel