The model is including the extra terms, but they are folded into, well,
somewhere. The -1 only removes the intercept from the Time effect: the
other factors still have to be contrasts to something. Doesn't Crawley's
book explain this? If not, I wrote a paper a few years ago that might help:
<http://www.sekj.org/PDF/anz46-free/anz46-124.pdf>
Bob
On 23 July 2013 20:53, Leif Rasmuson <rasmuson@uoregon.edu> wrote:
> Hi All,
>
> I am working on re-analyzing per a reviewers request.
>
> The goal of the project was to determine if the presence of predatory
> fishes caused female crabs to delay the release of larvae. Number of
> releases were recorded at three time periods: 1 hour before the simulated
> tide, 3 hours after the simulated tide and 6 hours after the tide.
> Predators were introduced at high tide and removed just following the 3
> hour observation. Thus we have two categorical variables with three levels
> each. Time: pre-introduction, after introduction and after removal.
> Predator: no-predator, predator of adults and predator of larvae.
>
> The number of females was not consistent between trials so my initial
> method was to use the SRH extension of the Kruskal wallace test to examine
> the percent of of larvae released.
>
> A reviewer suggested that I use a GLM with a binomial distribution and
> logit link to analyze the larval release patterns. I used Crawley's
book to
> analyze the data based on proportions. See code below. The problem I am
> running into is that the model is not including the predator control and
> interaction is excluding the predator control and the introduction time
> periods. Is this just the nature of using a binomial distribution (and/or
> our small sample size) or is there a way to force R to include all the
> factors and run all the interactions?
>
> Thanks,
> Leif
>
> N=GLM$NumFemale-GLM$**NumFemaleRelease
> y=GLM$NumFemaleRelease
> rv <-cbind(y,N)
>
> > model1=glm(y~GLM$Time*GLM$**Treatment-1,family=binomial)
> > summary(model1)
>
> Call:
> glm(formula = y ~ GLM$Time * GLM$Treatment - 1, family = binomial)
>
> Deviance Residuals:
> Min 1Q Median 3Q Max
> -1.6024 -0.7146 -0.4284 -0.2512 3.9885
>
> Coefficients:
> Estimate Std. Error z value Pr(>|z|)
> GLM$TimeAfterIntroduction -1.8289 0.2297 -7.963 1.68e-15 ***
> GLM$TimeAfterRemoval -3.6571 0.5064 -7.222 5.14e-13 ***
> GLM$TimePreIntro -5.0626 1.0029 -5.048 4.46e-07 ***
> GLM$TreatmentPlanktivore -0.0123 0.3211 -0.038 0.969
> GLM$TreatmentPredator -0.1589 0.3311 -0.480 0.631
> GLM$TimeAfterRemoval:GLM$**TreatmentPlanktivore 0.5339 0.7132 0.749 0.454
> GLM$TimePreIntro:GLM$**TreatmentPlanktivore 0.6561 1.2708 0.516 0.606
> GLM$TimeAfterRemoval:GLM$**TreatmentPredator -0.1791 0.8399 -0.213 0.831
> GLM$TimePreIntro:GLM$**TreatmentPredator 1.7496 1.1496 1.522 0.128
> ---
> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>
> (Dispersion parameter for binomial family taken to be 1)
>
> Null deviance: 1720.2 on 270 degrees of freedom
> Residual deviance: 258.3 on 261 degrees of freedom
> AIC: 378.23
>
> Number of Fisher Scoring iterations: 6
>
> T
>
> --
> Leif K. Rasmuson
>
> Doctoral Candidate
> Oregon Institute of Marine Biology
> Phone: (253)961-1763
> E-Mail: Rasmuson@uoregon.edu
>
>> <((((º>`·.¸¸.·´¯`·...¸><((((º>
>>
>
> ______________________________**________________
> R-help@r-project.org mailing list
>
https://stat.ethz.ch/mailman/**listinfo/r-help<https://stat.ethz.ch/mailman/listinfo/r-help>
> PLEASE do read the posting guide http://www.R-project.org/**
> posting-guide.html <http://www.R-project.org/posting-guide.html>
> and provide commented, minimal, self-contained, reproducible code.
>
--
Bob O'Hara
Biodiversity and Climate Research Centre
Senckenberganlage 25
D-60325 Frankfurt am Main,
Germany
Tel: +49 69 798 40226
Mobile: +49 1515 888 5440
WWW: http://www.bik-f.de/root/index.php?page_id=219
Blog: http://occamstypewriter.org/boboh/
Journal of Negative Results - EEB: www.jnr-eeb.org
[[alternative HTML version deleted]]