Major update to gologit2 now available

[Richard Williams]

Thanks to Kit Baum, a major update to gologit2 is now available on SSC. The new version finally supports factor variables and the svy: prefix, as well as more of the display options that have been added to Stata in recent years. This also means that gologit2 will now work correctly with the margins command and with Long & Freese’s spost13 commands (findit spost13_ado; also see their new book at http://www.stata.com/bookstore/regre...ent-variables/). gologit2 requires Stata 11.2 or higher. Those condemned to using older versions of Stata should download the earlier version of the program, which has been renamed gologit29 and only requires Stata 8.2.

Here is the description of the program:

gologit2 estimates generalized ordered logit models for ordinal dependent variables. A major strength of gologit2 is that it can also estimate three special cases of the generalized model: the proportional odds/parallel lines model, the partial proportional odds model, and the logistic regression model. Hence, gologit2 can estimate models that are less restrictive than the proportional odds /parallel lines models estimated by ologit (whose assumptions are often violated) but more parsimonious and interpretable than those estimated by a non-ordinal method, such as multinomial logistic regression (i.e. mlogit). The svy: prefix, as well as factor variables and post-estimation commands such as margins, are supported. Other key strengths of gologit2 include options for linear constraints, alternative model parameterizations, automated model fitting, alternative link functions (logit, probit, complementary log-log, log-log & cauchit), and the computation of estimated probabilities via the predict command. gologit2 works under Stata 11.2 or higher. Those with older versions of Stata should use gologit29 instead. gologit2 is inspired by Vincent Fu's gologit program and is backward compatible with both it and gologit29 but offers several additional powerful options.

More on gologit2 can be found at

http://www3.nd.edu/~rwilliam/gologit2/index.html

The 2006 Stata Journal article that introduced the program is available for free at

http://www.stata-journal.com/article...article=st0097

In a subsequent post I will give an example that illustrates the substantial advantages of new gologit2 over old gologit2.

[Richard Williams]

This illustrates the major advantages of new gologit2 over old gologit2.

Long & Freese (2014; pp. 372 – 377; http://www.stata.com/bookstore/regre...ent-variables/) give an example using (old) gologit2. As they note, you have to go through a lot of extra hoops because gologit2 did not support factor variables. Since this is no longer true I assume there will be a massive recall of the book, but for now you can just read this post. ;-)

Here is their example. It uses their mchange command (-findit spost13_ado-), which is basically a shell for margins and is much easier to use than margins, especially with multiple-outcome commands like gologit2, ologit, oprobit, and mlogit. (Even if you don’t want my programs, you should get theirs, as they are extremely useful.)

* Long & Freese, pp. 371-377, original example
use "http://www.indiana.edu/~jslsoc/stata/spex_data/gssclass4", clear
gen year1996 = (year==2) if year < .
gen year2012 = (year==3) if year < .
gen educ_hs = (educ==2) if educ < .
gen educ_col = (educ==3) if educ < .
gen agesq = age * age if age < .
gologit2 class female white year1996 year2012 ///
            educ_hs educ_col age agesq income, or
* The following is tempting but incorrect!!!
mchange white, amount(binary) dec(5) atmeans
* This is correct, but tedious, prone to error, & requires atmeans
mchange white, amount(binary) at(agesq = 2039) atmeans dec(5)

Here is the easy way to do it with new gologit2. Fewer commands are required and you are much less likely to make a mistake.

* Long & Freese updated with new gologit2, pp. 371-377
use "http://www.indiana.edu/~jslsoc/stata/spex_data/gssclass4", clear
gologit2 class i.female i.white i.year i.educ age c.age#c.age income, or
* The followjng numbers are right!!!
mchange white, amount(m) dec(5) atmeans
* Can easily get all the other marginal effects too, & without using atmeans
mchange, amount(m) dec(5)

A more in-depth analysis would also probably use some of gologit2’s other model-fitting options, such as autofit or pl.