Displaying 2 results from an estimated 2 matches for "cohort2004".
2012 May 04
2
Binomial GLM, chisq.test, or?
...o site 1 and this sort of things.
I thought I might do a binomial GLM considering as response variable the
site, I tried to run the more general model to have a look to overdispersion
but I believe there is something wrong even before worrying about
overdispersion. I know (I did a chisq.test) that cohort2004 is very diversly
represented between the two sites but it is not reflected in the results of
GLM. Here there are the results of chisq.test and the GLM:
1)
/>
age_cohort<-as.table(rbind(c(142,95,46,33,14,59,18,12,7,1,0),c(258,144,54,70,20,11,6,8,2,3,1)))
> dimnames(age_cohort)<-list(si...
2012 May 07
1
Can't find the error in a Binomial GLM I am doing, please help
...:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -12.57 324.74 -0.039 0.969
cohort2000 11.47 324.75 0.035 0.972
cohort2001 13.82 324.74 0.043 0.966
cohort2002 12.97 324.74 0.040 0.968
cohort2003 13.66 324.74 0.042 0.966
*cohort2004 14.25 324.74 0.044 0.965*
cohort2006 12.21 324.74 0.038 0.970
cohort2007 11.81 324.74 0.036 0.971
cohort2008 12.41 324.74 0.038 0.970
cohort2009 12.15 324.74 0.037 0.970
cohort2010 11.97 324.74 0.037 0.971
(Dispersion pa...