Fault in anova.glm (PR#1807)

[Bill.Venables@cmis.csiro.au]

When anova.glm is given several models to compare it ignores any prescribed
dispersion parameter.

An example:


> library (MASS)
> cond1 <- c(20, 50, 3, 9, 15, 20) 	
> cond2 <- c(18, 32, 44, 180, 25, 17) 	
> cond3 <- c(4, 11, 12, 44, 15, 6) 	
> testData <- data.frame(res = c(cond1, cond2, cond3),
+        conditions = factor(rep(1:3, c(6,6,6)))) 
> fm <- glm( res ~ conditions, family = negative.binomial(theta = 1),
testData)
> fm0 <- update(fm, .~1)
> anova(fm0, fm, test="Chisq", dispersion = 1)
Analysis of Deviance Table

Model 1: res ~ 1
Model 2: res ~ conditions
  Resid. Df Resid. Dev Df Deviance P(>|Chi|)
1        17    16.8979                      
2        15    11.6307  2   5.2672    0.0668

On the other hand:

> 1 - pchisq(5.2672, 2)
[1] 0.07181945
> 

The fix:

Replace line 16 of anova.glm:

        return(anova.glmlist(c(list(object), dotargs), test = test))

by

	return(anova.glmlist(c(list(object), dotargs), 
		dispersion = dispersion, test = test))


--please do not edit the information below--

Version:
 platform = i386-pc-mingw32
 arch = i386
 os = mingw32
 system = i386, mingw32
 status = 
 major = 1
 minor = 5.1
 year = 2002
 month = 06
 day = 17
 language = R

Windows 2000 Professional (build 2195) Service Pack 1.0

Search Path:
 .GlobalEnv, package:MASS, package:ctest, Autoloads, package:base

Bill Venables, 
CMIS, CSIRO Marine Laboratories, 
PO Box 120, Cleveland, Qld. 4163
AUSTRALIA
Phone:  +61 7 3826 7251  
Fax:    +61 7 3826 7304
Mobile: +61 419 634 642
<mailto: Bill.Venables@csiro.au>
http://www.cmis.csiro.au/bill.venables/


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel 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-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._