Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Odd "repeated time values" error in -glm-


From   "Clive Nicholas" <[email protected]>
To   [email protected]
Subject   st: Odd "repeated time values" error in -glm-
Date   Sat, 13 Jan 2007 19:54:13 -0000 (GMT)

I'm reposting this because, after checking the archive, this didn't reach
the list earlier today, even though my mailbox confirmed it as sent!

Talking of -glm-, as I have been for some of this week, I've come across a
frustrating error when attempting to fit one of my own models.
Essentially, I'm trying to get -glm- to fit OLS Newey-West models with
automatically logit-transformed estimates, which various parts of -help
glm- implies that it can do.

I've recreated the problem using one of Stata 9's toy datasets. Below,
-unemp- is the unemployment rate and -region- is a continuous variable with
no upper limit:

. webuse productivity, clear
(Public Capital Productivity)

. tsset state year
       panel variable:  state (strongly balanced)
        time variable:  year, 1970 to 1986

. glm unemp state public, link(logit) vce(hac nwest 1) t(year) robust eform
repeated time values in sample
r(451);

It fails miserably. Neither deleting -robust- nor normalizing -unemp- made
any difference; the -t()- option must be specified as -vce(hac ...)- is
also specified, as per -help glm-. But I don't know why it fails, because
when I use Baum/Schaffer/Stillman's -ivreg2-, from SSC, to fit pretty much
the same model:

. g logunemp=ln((unemp/100)/(1-(unemp/100)))

. ivreg2 logunemp state public, bw(2) robust small eform(OR)

OLS estimation
--------------

Statistics robust to heteroskedasticity and autocorrelation
  kernel=Bartlett; bandwidth=2
  time variable (t):  year
  group variable (i): state

                                                    Number of obs =      816
                                                    F(  2,   813) =    20.14
                                                    Prob > F      =   0.0000
Total (centered) SS     =  102.2836463              Centered R2   =   0.0434
Total (uncentered) SS   =  6067.674132              Uncentered R2 =   0.9839
Residual SS             =   97.8449067              Root MSE      =    .3469

----------------------------------------------------------------------------
           |               Robust
  logunemp |         OR   Std. Err.      t    P>|t|     [95% Conf. Interval]
-----------+----------------------------------------------------------------
     state |   .9987938   .0011866    -1.02   0.310     .9964672    1.001126
    public |   1.000003   4.15e-07     6.11   0.000     1.000002    1.000003
----------------------------------------------------------------------------
Included instruments: state public
----------------------------------------------------------------------------

Now how is it that -ivreg2- can fit pretty much the same model as I've
asked -glm- to, and yet -glm- chokes?

CLIVE NICHOLAS        |t: 0(044)7903 397793
Politics              |e: [email protected]
Newcastle University  |http://www.ncl.ac.uk/geps

Whereever you go and whatever you do, just remember this. No matter how
many like you, admire you, love you or adore you, the number of people
turning up to your funeral will be largely determined by local weather
conditions.

*
*   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/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index