R help - Oct 2011 - Plotting GEE confidence bands using "predict"

Hello Fellow R
Users,I have
spent the last week trying to find a work around to this problem and I can't
seem to solve it. I simply want to plot my GEE model result with 95% confidence
bands. 


I am using the geepack package to run a basic GEE model involving
nestling weights, to a Gaussian distribution, with "exchangeable"
error
structure. I am examining how nestling weight varies as a function of distance
from a plot boundary.



The response variable (WEIGHT) and the explanatory
variable (DISTANCE) are both continuous. The clustering factor (MOTHER) is
entered into my model to account for similarity of nestlings from the same nest
produced by a given mother.


My simplified code is
as follows:


summary(model1<-geeglm(WEIGHT~DISTANCE,
id=MOTHER,data=df,corstr="exchangeable")))
#I've included part of the
model output
here



<Coefficients:
<  
                  Estimate Std.err
  Wald         Pr(>|W|)  
 
<(Intercept)     15.8702  0.4416  
1291.8  < 2e-16 ***
<Initiationdate  -0.0664
 0.0157   17.9          2.4e-05
***
<
<Estimated Scale
Parameters:
<
<            Estimate
Std.err
<(Intercept)     5.78  
 2.46






plot(df$DISTANCE,df$WEIGHT)


abline(model1)


x<-seq(min(df$DISTANCE),max(df$DISTANCE),l=1000)


y<-predict(model1,data.frame(DISTANCE=x))


Everything is fine up until this last line of code, when I get the following
error message:



"Warning message:
In predict.lm(object, newdata, se.fit, scale = 1, type = ifelse(type ==  :
  calling predict.lm(<fake-lm-object>) ..."



This isn't a geepack problem because I get the same error message using the
gee
package as well. The above simplified code is how I usually create and plot 95%
CI bands for a linear model, followed by:


matlines(x,y) 


I have read through the package PDFs and searched the web and archives of
various listservs without success. Any help would be most appreciative. Seems
like there should be a simple work-around. Thank you,
Jason






	[[alternative HTML version deleted]]

Hi Jason,

I would go for Zelig package to get simulated values and confidence
intervals. It can handle gee model.

Shige

On Mon, Oct 17, 2011 at 9:38 AM, JASON M. HILL <jmh656 at psu.edu>
wrote:
> Hello Fellow R
> Users,I have
> spent the last week trying to find a work around to this problem and I
can't
> seem to solve it. I simply want to plot my GEE model result with 95%
confidence
> bands.
>
>
> I am using the geepack package to run a basic GEE model involving
> nestling weights, to a Gaussian distribution, with "exchangeable"
error
> structure. I am examining how nestling weight varies as a function of
distance
> from a plot boundary.
>
>
>
> The response variable (WEIGHT) and the explanatory
> variable (DISTANCE) are both continuous. The clustering factor (MOTHER) is
> entered into my model to account for similarity of nestlings from the same
nest
> produced by a given mother.
>
>
> My simplified code is
> as follows:
>
>
> summary(model1<-geeglm(WEIGHT~DISTANCE,
> id=MOTHER,data=df,corstr="exchangeable")))
> #I've included part of the
> model output
> here
>
>
>
> <Coefficients:
> <
> ? ? ? ? ? ? ? ? ?Estimate Std.err
> ?Wald ? ? ? ? Pr(>|W|)
>
> <(Intercept) ? ? 15.8702 ?0.4416
> 1291.8 ?< 2e-16 ***
> <Initiationdate ?-0.0664
> ?0.0157 ? 17.9 ? ? ? ? ?2.4e-05
> ***
> <
> <Estimated Scale
> Parameters:
> <
> < ? ? ? ? ? ?Estimate
> Std.err
> <(Intercept) ? ? 5.78
> ?2.46
>
>
>
>
>
>
> plot(df$DISTANCE,df$WEIGHT)
>
>
> abline(model1)
>
>
> x<-seq(min(df$DISTANCE),max(df$DISTANCE),l=1000)
>
>
> y<-predict(model1,data.frame(DISTANCE=x))
>
>
> Everything is fine up until this last line of code, when I get the
following
> error message:
>
>
>
> "Warning message:
> In predict.lm(object, newdata, se.fit, scale = 1, type = ifelse(type == ?:
> ?calling predict.lm(<fake-lm-object>) ..."
>
>
>
> This isn't a geepack problem because I get the same error message using
the gee
> package as well. The above simplified code is how I usually create and plot
95%
> CI bands for a linear model, followed by:
>
>
> matlines(x,y)
>
>
> I have read through the package PDFs and searched the web and archives of
> various listservs without success. Any help would be most appreciative.
Seems
> like there should be a simple work-around. Thank you,
> Jason
>
>
>
>
>
>
> ? ? ? ?[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>