do-file does not even run

[yoli park]

Recently I am working on the paper of Ehrmann et al. (2011). For replication I need to run their do-file which you can find in the supporting information on the following website.
But when I opened the do-file in the stata, I found that I could not run the file and got errors like: command new is unrecognized, command library is unrecognized...
So I really want to know why this file does not run and how I can deal with it. I really appreciate your help.

https://onlinelibrary.wiley.com/doi/....1002/jae.1173

[Marcos Almeida]

By reading the ‘Read me first’ file, it seems the do file is not in Stata. Rather, it is in Gauss program. By the way, the command - library - is a ubiquitous one in R, not in Stata. In short, only a Stata do-file will, well, do something in Stata. Hopefully that helps.

[yoli park]

Dear Marcos,

Thank you for your reply. So I need to download a program named Gauss, but how can I run the do-file? Should I run this do-file in stata with the help of gauss program?

Yours sincerely,

Yoli Park

[Daniel Bela]

Hi,

there's nothing to add to Marcos' answer. I can only rephrase: The syntax file simply is not a Stata do-file. You can not run it in Stata. It is a program for the matrix computation language GAUSS, as the authors state in their "Read me" file.

They say that they used GAUSS version 6.0 to compute their results. To reproduce them, you will have to get your hands on a GAUSS license and run the syntax, or understand it and re-implement the very same procedures in Stata. The latter, I assume, will take you more time.

All other questions about what the calculated and how they did it should be, IMHO, addressed to the authors.

Kind Regards
Bela

[Nick Cox]

Marcos Almeida has already answered the question in #3. Whatever is written in Gauss must be run in Gauss. Stata does not translate or understand Gauss syntax. See https://www.aptech.com/industry-solu...tion/students/ for some options if Gauss is not available at your workplace.

The only alternative is to translate from Gauss to Stata. I can't even see the code mentioned in #1 so I can't comment further.

[Carlo Lazzaro]

Nick:
I think it's Marcos' #2 that should harvest your endorsement (thanks, anyway)!

[yoli park]

Dear all,

I warmly appreciate all your help.

I have already spent a couple of days to do the re-implementation. It seems indeed very hard. But I will still try my best.

You can find the code in a zip-file right under the Readme file. And to see the supporting information which contains the zip-file and Readme file, I think maybe you would need University VPN.

Yours sincerely,

Yoli Park

[Nick Cox]

My impression is that unless you pay for access, seeing that stuff depends on workplace purchase of a subscription to that journal.

In any case, please note our guidelines about giving references at https://www.statalist.org/forums/help#references