I want to produce quite a number of tables which compare original regression results in a paper with the results I generate by various estimation commands "side by side". I cannot reproduce the original results so I must copy them from the paper by hand or by converting the table in the .pdf file of the paper to a spreadsheet.
At the moment I manually copy the originals into a well known spreadsheet, and paste them into a
Code:
input variable o_var1 o_var2 ... rhs_var1 _beta _beta ... rhs_var1 _se _se ... rhs_var2 _beta _beta ... rhs_var2 _se _se .... end
I then construct a matrix in a loop for all the results I wish to display choosing first from these variables then from the estimates I make (estimates restore ...., and then choose the relevant estimation results), so that the resulting matrix alternates original and new estimates. Adding row and column names gives a matrix that can be "estout matrix()" or similar.
Neither elegant, nor "automatic" even if I "putdocs" or "putexcel", all of which need subsequent editing.
Can someone suggest a time efficient way to do this?
Could I use the new "collect" command to do this in a nicer way?
Thanks for your attention and suggestions.
Richard
0 Response to Publishable tables of regression results beside originals entered by hand
Post a Comment