[R-sig-ME] Multi level response variables

[Mike Lawrence]

I'm no expert, but I think you want family='ordinal' from the MCMCglmm
package. The package page on cran has a bunch of documentation
materials (tutorial, etc):
http://cran.r-project.org/web/packages/MCMCglmm/index.html

On Tue, Jun 29, 2010 at 8:57 AM, Chris Mcowen <cm744 at st-andrews.ac.uk> wrote:
> Dear List,
>
> I am a little unsure how to structure my model and was after some advice. I am a little unsure if this question is appropriate for this list, if it is not please just delete and accept my apologise.
>
> I have 10 factors that are categorical variables and 5 levels of response variables -
>
> WOODYNONWOODY           L2_REGIONS      SEASONALITY     FLORALSYMMETRY  - Factors               IUCN - RESPONSE
> 2                                               2                                               2                                               2                               NE
> 2                                               4                                               2                                               2                               LC
> 2                                               1                                               2                                               2                               CR
> 2                                               1                                               2                                               1                               NE
> 2                                               3                                               2                                               2                               NE
> 2                                               1                                               1                                               2                               NE
> 2                                               1                                               2                                               3                               NT
> 1                                               1                                               3                                               2                               NE
> 2                                               2                                               2                                               2                               LC
> 2                                               1                                               5                                               2                               LC
>
> The response variables relate to how threatened the species is  - from not threatened to extinct (1-5)
>
> My first approach was to divide the 5 response levels into 2 - threatened ( levels 1+2) or non threatened (levels 3,4+5) and call
>
> model1 <- lmer(THREAT~1+(1|ORDER/FAMILY) + STORAGE_ORGAN + BREEDING_SYSTEM + POLLEN_DISPERSAL + LIFE_FORM + FRUIT + ENDOSPERM + HABIT + WOODY_NONWOODY + L2_REGIONS + SEASONALITY + ALTITUDE + SEED_FRUIT + FLORAL_SYMMETRY, family=binomial)
>
> Which worked well, now i want to see how the factors influence the individual response variables i.e do species with LC for instance, posses certain factors, and it is this i am unsure how to build into a lmer model.
>
> My overall goal would be to use the model as a predictive model and ask -  "if a species has factors a ,b,c for instance , can i predict what the response level (0-5) would be".
>
> Thanks, and once again i apologise if this is not the right place to ask this type of question.
>
> Chris
>
>
> Chris Mcowen
> PhD Student
>
> Room 15
> Sir Harold Mitchell Building
> University of St Andrews
> St Andrews
> Fife
> KY16 9TH
> UK
> Phone 01334 463381
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>



-- 
Mike Lawrence
Graduate Student
Department of Psychology
Dalhousie University

Looking to arrange a meeting? Check my public calendar:
http://tr.im/mikes_public_calendar

~ Certainty is folly... I think. ~