[R] confint() error

[Tord Snall]

Dear all,
I have run the confint() function according to below and I get the
following error message:

> confint(stepAIC.glm.spe.var.konn2.abund, level=0.95)
Waiting for profiling to be done...
Error: missing value where logical needed
In addition: Warning message: 
NaNs produced in: sqrt((fm$deviance - OriginalDeviance)/DispersionParameter)

or

> confint(stepAIC.glm.spe.var.konn2.abund, parm="DIA", level=0.95)
Waiting for profiling to be done...
Error in object[[pm]] : subscript out of bounds

if only one covariate is analysed.

Profiling seem to be the problem. I would too very much like to have know
what the warning and error messages mean. 

All clues are welcome.

Thanks in advance!

Tord Snäll

Copy of the analysis:

standardiserat.data <- read.table("Inom_buffert (standardiserat).txt",
header=TRUE)
library(mass)
glm.spe.var.konn2.abund<-
glm(OSPEINC~DIA+FICKAD+RINGB+TUNNB+UTANB+DYING+DEAD+QUAL+SPRUCEBR+SOILMOIST+
SHADE+BARKDEPTH+PPULABUN+SPEABART2+SPEABFER2, family=binomial,
data=standardiserat.data)
stepAIC.glm.spe.var.konn2.abund <- stepAIC(glm.spe.var.konn2.abund,
direction="both", trace=FALSE, test="Chisq")

> drop1(stepAIC.glm.spe.var.konn2.abund, test="Chisq")
Single term deletions

Model:
OSPEINC ~ DIA + RINGB + TUNNB + QUAL + SPRUCEBR + SHADE + SPEABFER2
          Df Deviance    AIC    LRT   Pr(Chi)    
<none>         454.28 470.28                     
DIA        1   480.20 494.20  25.91 3.571e-07 ***
RINGB      1   460.01 474.01   5.73 0.0166762 *  
TUNNB      1   458.47 472.47   4.19 0.0407075 *  
QUAL       1   462.13 476.13   7.85 0.0050844 ** 
SPRUCEBR   1   457.07 471.07   2.78 0.0952700 .  
SHADE      1   465.66 479.66  11.38 0.0007443 ***
SPEABFER2  1   460.00 474.00   5.72 0.0167640 *  
---
Signif. codes:  0  `***'  0.001  `**'  0.01  `*'  0.05  `.'  0.1  ` '  1 


> summary(stepAIC.glm.spe.var.konn2.abund, test="Chisq")

Call:
glm(formula = OSPEINC ~ DIA + RINGB + TUNNB + QUAL + SPRUCEBR + 
    SHADE + SPEABFER2, family = binomial, data = standardiserat.data)

Deviance Residuals: 
     Min        1Q    Median        3Q       Max  
-1.60196  -0.75333  -0.56346  -0.03123   2.45180  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)  -1.5160     0.2627  -5.770 7.93e-09 ***
DIA           0.6128     0.1221   5.018 5.23e-07 ***
RINGB        -0.8115     1.1471  -0.707  0.47930    
TUNNB        -0.9344     1.1959  -0.781  0.43460    
QUAL         -0.3696     0.1421  -2.601  0.00930 ** 
SPRUCEBR     -0.2059     0.1335  -1.543  0.12291    
SHADE         0.4601     0.1417   3.247  0.00117 ** 
SPEABFER2     0.2711     0.1126   2.408  0.01602 *  
---
Signif. codes:  0  `***'  0.001  `**'  0.01  `*'  0.05  `.'  0.1  ` '  1 

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 512.21  on 466  degrees of freedom
Residual deviance: 454.28  on 459  degrees of freedom
AIC: 470.28

Number of Fisher Scoring iterations: 6

> confint(stepAIC.glm.spe.var.konn2.abund, level=0.95)
Waiting for profiling to be done...
Error: missing value where logical needed
In addition: Warning message: 
NaNs produced in: sqrt((fm$deviance - OriginalDeviance)/DispersionParameter) 


> profile.stepAIC.glm.spe.var.konn2.abund <-
profile(stepAIC.glm.spe.var.konn2.abund)

Error: missing value where logical needed
In addition: Warning message: 
NaNs produced in: sqrt((fm$deviance - OriginalDeviance)/DispersionParameter) 


> confint(stepAIC.glm.spe.var.konn2.abund, parm="DIA", level=0.95)

Waiting for profiling to be done...
Error in object[[pm]] : subscript out of bounds


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._