gologit2 chooses my DV category of interest as the reference. How can I change this?

[Zach Goldberg]

Greetings,

I'm running Stata 15.1 on OSX. I reluctantly downloaded and switched over to gologit2 after one of my IVs in an ologit model failed the brant test. However, when I run the command it fails to output data for my category of interest. To clarify, my dependent variable contains 3 ordered categories: -1, 0 and 1. While the parameters for -1 and 0 are displayed, 1 is not. What can I do in this situation?

Additionally, I receive the following warning message when I run the command: [QUOTE][WARNING! 1 in-sample cases have an outcome with a predicted probability that is
less than 0./QUOTE]

Any help would be much appreciated. Thanks!

[Richard Williams]

See the gologit2 troubleshooting FAQ, which addresses both issues you raise:

https://www3.nd.edu/~rwilliam/gologit2/tsfaq.html

Very quickly, gologit2 output may look like mlogit output but there are fundamental differences. The first panel gives category 1 vs categories 2 and 3 combined. The second panel gives 1 and 2 versus 3. It would make no sense to have a panel for 1, 2, 3 versus nothing.

Also, some recently updated handouts on gologit2 are available near the bottom of

https://www3.nd.edu/~rwilliam/xsoc73994/index.html

[Zach Goldberg]

Hey Richard,

Thanks a lot for input here (and for all the hard work you put into designing the package). I'm accustomed to making tables for OLS and ologit models, but I'm not sure how to present the output displayed by gologit2. Have any examples you can show me? Thanks again!

[Richard Williams]

The troubleshooting FAQ (https://www3.nd.edu/~rwilliam/gologit2/tsfaq.html) has a section entitled "The output is hard to read and understand. What are some good ways to interpret and present the results?" The approach I used in Table 2 of my Journal of Mathematical Sociology paper is probably my favorite, but you can look at the other options.

The worst option is just to present the Stata output as is -- if I could figure out a nicer way to output the results, I would.