Hello,
Thanks for getting back to me. I was of the impression that once the res. var.
is larger than the df then the data was overdispersed and as such the model was
not a best fit. Is this true?
Here is an example of the output from R:
Call:
glm(formula = y ~ log(conc), family = binomial)
Deviance Residuals:
1 2 3 4 5 6
0.54568 1.08474 0.04561 -2.00959 0.05772 1.33891
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -5.52815 0.85916 -6.434 1.24e-10 ***
log(conc) 0.40457 0.05938 6.813 9.56e-12 ***
---
Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 78.811 on 5 degrees of freedom
Residual deviance: 7.311 on 4 degrees of freedom
AIC: 30.45
Number of Fisher Scoring iterations: 4
>
> xv<-seq(min(log(conc)-1),max(log(conc)+1),0.01)
> lines(xv,predict(model,list(conc=exp(xv)),type="response"))
>
> dose.p(model,p=c(0.10,0.25,0.5,0.75,0.90))
Dose SE
p = 0.10: 8.233179 0.9810446
p = 0.25: 10.948665 0.6580127
p = 0.50: 13.664152 0.4703530
p = 0.75: 16.379638 0.5720159
p = 0.90: 19.095125 0.8665399
> exp(13.664152)
[1] 859539.4
> exp(13.664152+(1.96*0.4703530))
[1] 2160918
> exp(13.664152-(1.96*0.04703530))
[1] 783842
BW
Adaku
________________________________________
From: peter dalgaard [[email protected]]
Sent: 09 July 2012 20:03
To: Lawrence, Adaku
Cc: [email protected]
Subject: Re: [R] Correcting for overdispersion
On Jul 9, 2012, at 20:23 , Lawrence, Adaku wrote:
> Hello,
>
> I am trying to determine LD50 and LD95 using dose.p in MASS however some of
> the Residual variance is larger than the degrees of freedom. Please can
> anyone help with any advice as to how i can correct for this?
Er, in what sense is that a problem? Your code is not reproducible, at least
some output to look at might help.
-pd
>
> Here is the model as inputted into R
>
>
>
> y<-cbind(dead,n-dead)
>
> model<-glm(y~log(conc),binomial)
> summary(model)
>
> xv<-seq(min(log(conc)-1),max(log(conc)+1),0.01)
> lines(xv,predict(model,list(conc=exp(xv)),type="response"))
>
> dose.p(model,p=c(0.10,0.25,0.5,0.75,0.90,0.95))
>
>
>
> Thanks
>
> Adaku
>
>
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> [email protected] 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.
--
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Email: [email protected] Priv: [email protected]
[[alternative HTML version deleted]]
______________________________________________
[email protected] 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.
