[R] Generalized Estimating Functions

Frederico Zanqueta Poleto fred at poleto.com
Mon May 13 13:56:19 CEST 2002


Hi Thomas,

Thank you for your help.
You are right. I've tried the OME example and the working correlation was 
right.
Maybe it is a problem with Poisson, what do you think?

> desreg<-gee
(desovas~ger+pop+fec+ger*pop+ger*fec+pop*fec,id=unidade,data=desovas.ovos.in
viaveis,family=poisson,corstr="exchangeable")
[1] "Beginning Cgee S-function, @(#) geeformula.q 4.13 98/01/27"
[1] "running glm to get initial regression estimate"
[1]  0.93464328  0.02317831  0.29278443  0.21399704  0.11593571  
0.05833780 -0.11513896
> summary(desreg)

 GEE:  GENERALIZED LINEAR MODELS FOR DEPENDENT DATA
 gee S-function, version 4.13 modified 98/01/27 (1998) 

Model:
 Link:                      Logarithm 
 Variance to Mean Relation: Poisson 
 Correlation Structure:     Exchangeable 

Call:
gee(formula = desovas ~ ger + pop + fec + ger * pop + ger * fec + 
    pop * fec, id = unidade, data = desovas.ovos.inviaveis, family = 
poisson, 
    corstr = "exchangeable")

Summary of Residuals:
       Min         1Q     Median         3Q        Max 
-4.5893464 -1.5463047 -0.4124404  1.4106536  9.2329743 


Coefficients:
               Estimate Naive S.E.    Naive z Robust S.E.   Robust z
(Intercept)  0.93464317 0.06012985 15.5437480  0.05954487 15.6964517
ger          0.02317832 0.07001297  0.3310576  0.06915388  0.3351702
pop          0.29278452 0.06565443  4.4594785  0.06505729  4.5004104
fec          0.21399711 0.07252180  2.9507967  0.07150063  2.9929402
ger:pop      0.11593572 0.07355902  1.5760911  0.07318636  1.5841166
ger:fec      0.05833782 0.07000740  0.8333093  0.07054498  0.8269592
pop:fec     -0.11513904 0.07216442 -1.5955098  0.07149595 -1.6104274

Estimated Scale Parameter:  1.366923
Number of Iterations:  1

Working Correlation
      [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10] [,11] [,12]
 [1,]    1    0    0    0    0    0    0    0    0     0     0     0
 [2,]    0    0    0    0    0    0    0    0    0     0     0     0
 [3,]    0    0    0    0    0    0    0    0    0     0     0     0
 [4,]    0    0    0    0    0    0    0    0    0     0     0     0
 [5,]    0    0    0    0    0    0    0    0    0     0     0     0
 [6,]    0    0    0    0    0    0    0    0    0     0     0     0
 [7,]    0    0    0    0    0    0    0    0    0     0     0     0
 [8,]    0    0    0    0    0    0    0    0    0     0     0     0
 [9,]    0    0    0    0    0    0    0    0    0     0     0     0
[10,]    0    0    0    0    0    0    0    0    0     0     0     0
[11,]    0    0    0    0    0    0    0    0    0     0     0     0
[12,]    0    0    0    0    0    0    0    0    0     0     0     0


Best regards,
--
Frederico Zanqueta Poleto
fred at poleto.com
--
"It would be possible to describe everything scientifically, but it would 
make no sense; it would be without meaning, as if you described a 
Beethoven symphony as a variation of wave pressure." Albert Einstein


------------- Original message follows -------------

On Sun, 12 May 2002, Frederico Zanqueta Poleto wrote:

>   Hi,
>
> I'm trying to fit a marginal model via GEE but I'm getting strange
> results and few problems.
> If I set the working correlation as exchangeable I'm getting the same
> fitting when I set as independent. Comparing to SAS results it shouldn't
> happen.
> If I try to use another working correlation (like AR-M or stat_M_dep), R
> just exits without giving any error message.

We probably need an example.  The `OME' example in help(gee) gives
different results using exchangeable and independence working
correlations, so it isn't a universal problem.

> Another doubt is how R estimates the scale parameter for Poisson? I get
> 1.36 in R, 1.25 in SAS (estimation by the square root of DEVIANCE/DOF)
> and 1.17 (estimation by the square root of Pearson's Chi-Square/DOF).
> I'm using R 1.4.1 but all the problems apply for S-Plus 2000 too.

This is because 1.36=1.17^2, I should think.  The disperson parameter from
gee() is a variance scale factor not a standard deviation scale factor.

	-thomas

Thomas Lumley			Asst. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle



-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._



More information about the R-help mailing list