Skip to content

again with polr

3 messages · Luca Braglia

#
hello everybody

solved the problem with summary, now I have another one

eg I estimate
now I've putted Hess=T in order to perform a summary on the
estimate. I missed this option last time
unfortunately when I do
I get
... infinite  or missing values. 

If I exclude I(eta^2) from the estimation I can do summaries: but
I need also I(eta^2) in order to compare this model with the same
specification estimated with Ols

Can you please help me or give me an hint?

If it can be useful, i hereby put an estimation with stata's
oprobit function (varnames are a little bit different)
thank you
   Luca
#
Hello everybody

as I said, i'm not a ordered probit guru, but what happened seems
to me a little strange.

Here I put my dataset

http://bragliozzo.altervista.org/asd.dta

this is the code you can try

# -------------------
library(foreign)
library(MASS)
asd <- read.dta("asd.dta",
                convert.underscore = F)

base.op <- polr(
                as.ordered(lz2_1) ~
                ly +
                lnfiglimin +
                lnfiglimagg +
                lnadulti +
                eta +
                eta2 +
                uomo +
                laurea +
                saluteno +
                extra
                ,
                ##                 weights = w,
                data = asd,
                method = "probit",
                Hess = T
                )

summary(base.op)

# -------------------

The problem is: you can try _any_ combination of the explaining
variables, but when you insert eta2 (eta == age in years, eta2 ==
age^2) you can't do the summary. I don't see why: it's a variable
like others with no strange problems (at least it seems to me to
be so).

If I'm wrong please tell me, otherwise I'll submit a bug report

ok
On 15/03/08 - 10:29, Luca Braglia wrote: