declaring the variables

An embedded and charset-unspecified text was scrubbed...
Name: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-mixed-models/attachments/20101116/d29b97e7/attachment.pl>
Hi Michael,

You will need to use a different function than glm(). Something from
an additional R package such as nlme or lme4.

There's some information here:
http://glmm.wikidot.com/

With glmm() from lme4 your model specification would look something like this:

fish <- glmm(Catch ~ Season + Tide + (1|Angler),
family=gaussian(identity), data=fishcatch)

Assuming Catch is appropriately modelled as gaussian that is.

andydolman at gmail.com
I am running a GLM with catch data from a fishery. ?I understand how to run
a GLM in R. ?What I don't understand is how to declare the types of
variables that I have. ?For example, I am modeling catch with several
different types of variables. ?Here are details on my variables:

Variable ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Type
comment

Anglers ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? random factor ? ? ? ? ? ? ? ? ? I
have the names of over 100 anglers.

Season ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? fixed factors ? ? ? ? ? ? ? ? ? ? ? I
broke it up into the 4 seasons

Tide phase ? ? ? ? ? ? ? ? ? ? ? ? ?Continuous variable ? ? ? ?I have the
proportion of the tide phase during the fishing event

My R code for the GLM is:

fish <- glm(Catch ~ Angler +Season + Tide, family=gaussian(identity),
data=fishcatch)

However, how do declare the types of variables used in the model? ?For
example, how do I declare that the angler variable is a random factor?

Any help would be greatly appreciated.

Mike

? ? ? ?[[alternative HTML version deleted]]

_______________________________________________
R-sig-mixed-models at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

Hello!
Maybe I am missing something, but I think you should first check how R 
"understands" your data structure:
 > str(<data-frame-name>)
 From this you can explitily define type that you need/want, like:
 > <data-frame-name>$Anglers = as.factor(<data-frame-name>$Anglers)
 > <data-frame-name>$Season = as.factor(<data-frame-name>$Season)
 > <data-frame-name>$Tide.phase = as.numeric(<data-frame-name>$Tide.phase)
Also, you can use: as.logical(), as.integer(), even as.character() and 
some others.

Again, I apologize if I understood you problem wrongly, but this seems 
to me a solution.

Best,
PM
I am running a GLM with catch data from a fishery.  I understand how to run
a GLM in R.  What I don't understand is how to declare the types of
variables that I have.  For example, I am modeling catch with several
different types of variables.  Here are details on my variables:

Variable                               Type
comment

Anglers                                 random factor                   I
have the names of over 100 anglers.

Season                                 fixed factors                       I
broke it up into the 4 seasons

Tide phase                          Continuous variable        I have the
proportion of the tide phase during the fishing event

My R code for the GLM is:

fish<- glm(Catch ~ Angler +Season + Tide, family=gaussian(identity),
data=fishcatch)

However, how do declare the types of variables used in the model?  For
example, how do I declare that the angler variable is a random factor?

Any help would be greatly appreciated.

Mike

	[[alternative HTML version deleted]]

_______________________________________________
R-sig-mixed-models at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

Sorry, again, mail went before I concluded.
As to second part:
fish<- glm(Catch ~ Angler +Season + Tide, family=gaussian(identity),
data=fishcatch)

However, how do declare the types of variables used in the model?  For
example, how do I declare that the angler variable is a random factor?

Then, you should use something like:

fish <- lmer(Catch ~ Season + Tide + (1|Angler), data=fishcatch)

In lmer(), Gaussian is the default and linear mixed-model is fit, but 
you can also use 'binomial', 'poisson' etc, and then generalized linear 
mixed-model is fit.

Best,
PM