Skip to content

APE: compar.gee( )

2 messages · juli g. pausas, Uwe Ligges

#
Dear all,
I don't understand the following behaviour: Running  compar.gee   (in 
library ape ) with and without the option 'data', it give me different 
results
Example:

.... Start R ....
[1] "gee.na" "mydata" "mytree"
Loading required package: gee 
[1] "Beginning Cgee S-function, @(#) geeformula.q 4.13 98/01/27"
[1] "running glm to get initial regression estimate"
[1] 7.9500000 0.5155172

Call:
  formula: alt ~ R

Number of observations:  37 

Model:
 Link:                      identity 
 Variance to Mean Relation: gaussian 

Summary of Residuals:
        Min          1Q      Median          3Q         Max 
-12.1267954  -9.4267954  -7.4267954  -0.4267954  20.7903982 


Coefficients:
            Estimate     S.E.        t Pr(T > |t|)
(Intercept) 9.209602 4.760274 1.934679  0.08798892
R1          3.217194 2.548273 1.262500  0.24130121

Estimated Scale Parameter:  86.39367
"Phylogenetic" df (dfP):  10.24615
[1] "Beginning Cgee S-function, @(#) geeformula.q 4.13 98/01/27"
[1] "running glm to get initial regression estimate"
[1] 7.9500000 0.5155172

Call:
  formula: alt ~ R

Number of observations:  37 

Model:
 Link:                      identity 
 Variance to Mean Relation: gaussian 

Summary of Residuals:
       Min         1Q     Median         3Q        Max 
-14.123399 -11.423399  -8.856398  -1.856398  20.143602 


Coefficients:
            Estimate     S.E.        t Pr(T > |t|)
(Intercept) 9.856398 5.137029 1.918696  0.09020828
R1          4.567001 2.351966 1.941781  0.08701965

Estimated Scale Parameter:  103.2713
"Phylogenetic" df (dfP):  10.24615
[1] ".GlobalEnv"      "mydata"          "package:gee"     "package:ape"    
 [5] "package:methods" "package:ctest"   "package:mva"     "package:modreg" 
 [9] "package:nls"     "package:ts"      "Autoloads"       "package:base"
(R 1.8.1 for Windows
ape 1.2)



Note that the results are different.
Am I doing something wrong? Which is the correct answer?
Thanks for any help

Juli
#
juli g. pausas wrote:

            
Note that ape 1.2-1 is the recent version.
I guess that one with the data argument:
Are you sure there is no object called "alt" or "R" in your workspace 
before attaching the data frame (it will mask the attached object)?

You might want to check that:
  identical(R, mydata$R)
  identical(alt, mydata$alt)

If those objects are identical, you might want to ask the maintainer of 
ape, Emmanuel Paradis <paradis at isem.univ-montp2.fr>.


Uwe Ligges