Ben, thanks again.
John
From: Ben Bolker
Cc: r-h...@stat.math.ethz.ch; S Ellison ; peter dalgaard
Sent: Tue, December 21, 2010 9:26:29 AM
Subject: Re: [R] logistic regression or not?
On 10-12-21 12:20 PM, array chip wrote:
> Thank you Ben, Steve and Pe
package) is another way of
handling analysis of proportions.
>
>
> *From:* Ben Bolker
> *To:* r-h...@stat.math.ethz.ch
> *Sent:* Tue, December 21, 2010 5:08:34 AM
> *Subject:* Re: [R] logistic regression or not?
>
> array chip ya
:
glm(log(percentage/(1-percentage))~treatment,data=test)
Thanks
John
From: Ben Bolker
To: r-h...@stat.math.ethz.ch
Sent: Tue, December 21, 2010 5:08:34 AM
Subject: Re: [R] logistic regression or not?
array chip yahoo.com> writes:
[snip]
> I can th
>...and before you believe in overdispersion, make sure you have a
credible explanation for it. All too often, what you really have
>is a model that doesn't fit your data properly.
Well put.
A possible fortune?
S Ellison
***
Th
On Dec 21, 2010, at 14:22 , S Ellison wrote:
> A possible caveat here.
>
> Traditionally, logistic regression was performed on the
> logit-transformed proportions, with the standard errors based on the
> residuals for the resulting linear fit. This accommodates overdispersion
> naturally, but wi
A possible caveat here.
Traditionally, logistic regression was performed on the
logit-transformed proportions, with the standard errors based on the
residuals for the resulting linear fit. This accommodates overdispersion
naturally, but without telling you that you have any.
glm with a binomial f
array chip yahoo.com> writes:
[snip]
> I can think of analyzing this data using glm() with the attached dataset:
>
> test<-read.table('test.txt',sep='\t')
> fit<-glm(cbind(positive,total-positive)~treatment,test,family=binomial)
> summary(fit)
> anova(fit, test='Chisq')
> First, is this still
Hi, I have a dataset where the response for each person on one of the 2
treatments was a proportion (percentage of certain number of markers being
positive), I also have the number of positive & negative markers available for
each person. what is the best way to analyze this kind of data?
I can
8 matches
Mail list logo
