What is intreg? What is the theory behind it?

[Hutchins Yeo]

The UCLA IDRE website has an article on -intreg- to estimate models with interval dependent variables. It seems interesting, but I can't find any theoretical discussion of it. No wikipedia article. No sources when I google. It always just re-routes to the stata package. The R equivalent is survreg(), which is based on something in survival analysis. However survreg() and -intreg- seem different. For instance, the former estimates log-odds and the latter standard OLS coefficients.

I guess I'm trying to learn more about "interval regressions," but I can't find any info on it outside these two packages. So I'm wary to use them.

[Arthur Morris]

Maybe you’ve already looked into this, but the intreg package is quite well documented. I’d start with the following resources:

1. help intreg (I’m including this for completeness, not to be pedantic)
2. The full PDF documentation of the command has a solid discussion of the methods it uses. (Note that 1. includes links to 2.)

The PDF documentation of the command is very thorough and includes a number of recommendations for further reading on the topic of interval regressions.

[Arthur Morris]

The R equivalent is survreg(), which is based on something in survival analysis. However survreg() and -intreg- seem different. For instance, the former estimates log-odds and the latter standard OLS coefficients.

Note that intreg does not estimate “standard OLS coefficients,” see the “Methods and formulas” section of the PDF documentation.

[Nick Cox]

To add to the helpful posts of Arthur Morris our own FAQ gives this generic advice -- not even mentioning Wikipedia or Google....

3. What should I do before I post?

Before posting, consider other ways of finding information:

• the online help for Stata
• Stata's search command, which can tell you about all built-in Stata commands, all ado-files published in the Stata Journal, all FAQs on the Stata website, www.stata.com, and user-written Stata programs available on the Internet (if you have Stata 12 or earlier, you can use findit to search all these sources at once)
• the manuals, accessible in .pdf form to all (note that choosing "Help" in the Stata menu, followed by "PDF documentation", gets you there directly)
• past issues of the Stata Technical Bulletin or the Stata Journal
• contacting the authors of user-written ado-files (who usually have email addresses but are not always members of Statalist)
• contacting Stata Technical Services if it is really a question for them

[Richard Williams]

Here are my notes on it. They don't go much beyond what is already in the Stata documentation, although it does have an additional example.

https://www3.nd.edu/~rwilliam/xsoc73994/intreg2.pdf

[Hutchins Yeo]

Note that intreg does not estimate “standard OLS coefficients,” see the “Methods and formulas” section of the PDF documentation.

Oh ya, sorry. I meant it can be interpreted similarly (at least, according to the UCLA site). Also that's weird. The help file that comes up when I use -help intreg- cuts off before the Methods section. Hence my confusion. Thanks!

[Nick Cox]

Nothing weird: Methods and formulas are documented in the linked manual entry as Arthur Morris documented in #2 and #3.

[Hutchins Yeo]

Thanks!