[R] Nested variables

[DanielWC]

Hi again
I read the chapter and it says that I can nest B in A by writing A/B. But it
doesnt seem to work properly. The package I use is pscl, I am trying to
model zero inflated data. Here is the full model:

zip1<-zeroinfl(Visitation ~ Habitat/Site + Sps + TP + Resource +
Sps*Habitat/Site + Sps*Resource + Sps*TP + TP*Resource + TP*Habitat/Site |
TP + Sps + Resource + Habitat/Site + TP*Sps + Sps*Resource + TP*Resource +
TP*Habitat/Site, data = data1)

where as you can see I have tried to include Habitat/Site. But it doesnt
seem like its possible to create interaction links that way. I would like
create the interaction link between say Sps and Habitat/Site. If I write
Sps*Habitat/Site R somehow thinks that its a 3-way interaction link and
still includes habitat as a single variable. Here is the summary just so you
can see:

Count model coefficients (poisson with log link):
                   Estimate Std. Error z value Pr(>|z|)    
(Intercept)       3.8643338  0.3450539  11.199  < 2e-16 ***
Habitat          -3.2757473  0.4871726  -6.724 1.77e-11 ***
Sps              -0.3332371  0.1117680  -2.982  0.00287 ** 
TP               -0.3857513  0.0975556  -3.954 7.68e-05 ***
Resource          0.0130935  0.0009545  13.717  < 2e-16 ***
Habitat:Site      0.2487012  0.0415908   5.980 2.24e-09 ***
Habitat:Sps       0.5432845  0.1230434   4.415 1.01e-05 ***
Sps:Resource     -0.0041420  0.0002632 -15.738  < 2e-16 ***
Sps:TP            0.0438085  0.0189411   2.313  0.02073 *  
TP:Resource      -0.0009413  0.0002887  -3.260  0.00111 ** 
Habitat:TP        0.1549435  0.1205884   1.285  0.19883    
Habitat:Site:Sps -0.0305628  0.0107794  -2.835  0.00458 ** 
Habitat:Site:TP  -0.0096408  0.0098497  -0.979  0.32768    

Zero-inflation model coefficients (binomial with logit link):
                  Estimate Std. Error z value Pr(>|z|)    
(Intercept)     -1.3160850  0.8200631  -1.605   0.1085    
TP               0.2808474  0.2657544   1.057   0.2906    
Sps              0.8023648  0.1281520   6.261 3.82e-10 ***
Resource        -0.0075967  0.0035817  -2.121   0.0339 *  
Habitat         -1.3122811  1.1348922  -1.156   0.2476    
Habitat:Site     0.1390498  0.0984716   1.412   0.1579    
TP:Sps          -0.0189088  0.0406299  -0.465   0.6417    
Sps:Resource    -0.0005548  0.0006843  -0.811   0.4175    
TP:Resource      0.0007100  0.0008454   0.840   0.4010    
TP:Habitat       0.7195254  0.3666809   1.962   0.0497 *  
TP:Habitat:Site -0.0782610  0.0314229  -2.491   0.0128 *

Thanks again


Mark Difford wrote:
> 
> Daniel,
> 
>>> Yes I am trying to model such data, and i need R to know that Site is
>>> nested within Habitat. 
>>> Do I use some kind of command before running the model (like factor()
>>> and so on) or do i 
>>> write it in the model formula. If so, how?
> 
> You still are not telling the list enough, since nesting specifications
> are sometimes package/program specific. You could help yourself by reading
> some of R's basic documentation about how to specify a nested model. Then
> "it" will tell you what to do.
> 
> ?formula
> 
> Or read §11.1 of "An Introduction to R" (one of R's manuals).
> 
> Regards, Mark.
> 
> 
> DanielWC wrote:
>> 
>> Hi. 
>> Yes I am trying to model such data, and i need R to know that Site is
>> nested within Habitat. Do I use some kind of command before running the
>> model (like factor() and so on) or do i write it in the model formula. If
>> so, how?
>> 
>> Thanks
>> 
>> 
>> 
>> Douglas Bates-2 wrote:
>>> 
>>> On Fri, May 29, 2009 at 7:50 AM, DanielWC <daniel.carstensen at gmail.com>
>>> wrote:
>>> 
>>>> Hello
>>>> I am working with a biological data including variables called Habitat
>>>> and
>>>> Site, example:
>>> 
>>>> Habitat     Site
>>> 
>>>> Forest      Low
>>>> Forest      Low
>>>> Forest      High
>>>> Forest      High
>>> 
>>>> I want to tell R that the Site variable is nested within the Forest
>>>> variable
>>>> (that it is not a new variable).
>>>> Does anyone know how to do this?
>>> 
>>> If you could explain what you mean by "tell R", it would help.  Are
>>> you trying to model such data and you want to know how to express the
>>> relationship between such factors in a model formula?
>>> 
>>> ______________________________________________
>>> R-help at r-project.org 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.
>>> 
>>> 
>> 
>> 
> 
> 

-- 
View this message in context: http://www.nabble.com/Nested-variables-tp23779402p23830019.html
Sent from the R help mailing list archive at Nabble.com.