Thursday, February 16, 2006

Re: st: predictnl and ologit

At 10:15 PM 2/16/2006, Jeff Pitblado, StataCorp LP wrote: >giacomo <gc4@duke.edu> found a bug in -predictnl- when used after -ologit-: > > > I am estimating an ordered logit model and I would like to use predictnl > > to compute standard errors and confidence bounds for the estimated > > probabilities. In more detail, I first estimate: > > > > ologit y x1 x2, where y takes on three values (1,2,3) > > > > However, when I use -predictnl-, I get an error message: > > > > . predictnl phat=predict(outcome(#2)), se(phat_se) > > equation cut2 not found > > predict(outcome(#2)) invalid > >This bug has been fixed, and will be available in the next ado-file update.

And if you are absolutely desperate to get it working now - download the slow but sure -oglm-. -oglm- estimates the same model as -ologit- by default but also supports the probit, cloglog and loglog links. I double-checked -oglm- versus -oprobit- and the predictnl command produced the same result in both cases. Example:

. use "http://www.indiana.edu/~jslsoc/stata/spex_data/ordwarm2.dta" (77 & 89 General Social Survey)

. ologit warm yr89 male, nolog

Ordered logistic regression Number of obs = 2293 LR chi2(2) = 129.53 Prob > chi2 = 0.0000 Log likelihood = -2931.0051 Pseudo R2 = 0.0216

------------------------------------------------------------------------------ warm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- yr89 | .5897259 .0784044 7.52 0.000 .436056 .7433958 male | -.6349161 .0773504 -8.21 0.000 -.7865201 -.4833122 -------------+---------------------------------------------------------------- /cut1 | -2.036695 .0809737 -2.1954 -1.877989 /cut2 | -.2968644 .0661125 -.4264425 -.1672863 /cut3 | 1.499808 .0736113 1.355533 1.644084 ------------------------------------------------------------------------------

. predictnl phat=predict(outcome(#2)), se(phat_se) equation cut2 not found predict(outcome(#2)) invalid r(198);

. oglm warm yr89 male

Ordered Logistic Regression Number of obs = 2293 LR chi2(2) = 129.53 Prob > chi2 = 0.0000 Log likelihood = -2931.0051 Pseudo R2 = 0.0216

------------------------------------------------------------------------------ warm | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- yr89 | .5897256 .0784044 7.52 0.000 .4360558 .7433955 male | -.6349165 .0773504 -8.21 0.000 -.7865205 -.4833125 -------------+---------------------------------------------------------------- /cut1 | -2.036696 .0809737 -25.15 0.000 -2.195401 -1.87799 /cut2 | -.2968648 .0661125 -4.49 0.000 -.4264429 -.1672867 /cut3 | 1.499808 .0736113 20.37 0.000 1.355532 1.644084 ------------------------------------------------------------------------------

. predictnl phat=predict(outcome(#2)), se(phat_se)

. sum ph*

Variable | Obs Mean Std. Dev. Min Max -------------+-------------------------------------------------------- phat | 2293 .3145302 .0573749 .2243567 .3861832 phat_se | 2293 .0121664 .0004829 .0115066 .0130397

Or, if you happen to have -gologit2- installed, run it using the -pl- option, e.g.

gologit2 warm yr89 male, pl lrf

I'll have to add this to the documentation for my programs - "They work when the rest of Stata doesn't" :)

------------------------------------------- Richard Williams, Notre Dame Dept of Sociology OFFICE: (574)631-6668, (574)631-6463 FAX: (574)288-4373 HOME: (574)289-5227 EMAIL: Richard.A.Williams.5@ND.Edu WWW (personal): http://www.nd.edu/~rwilliam WWW (department): http://www.nd.edu/~soc

* * For searches and help try: * http://www.stata.com/support/faqs/res/findit.html * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/


Tag:


Links to this post:

Create a Link



<< Home

This page is powered by Blogger. Isn't yours?