R help - Nov 2012 - Package "glmulti": Include a variable in ALL models

Dear all,

I have a question about the glmulti package. I want to include some variables in
all models. To that end I applied the wrapper function as shown in the examples
(http://www.inside-r.org/packages/cran/glmulti/docs/glmulti).

To include the variable "Geslacht" in all models:
> glm.redefined = function(formula, data, always="", ...)
{glm(as.formula(paste(deparse(formula), always)), data=data, ...)}
> output<-glmulti(ZBK3~ 
Onafhankelijk*Zelfconcept*Hyperactief*Populariteit, data=IME,
fitfunc=glm.redefined, level=2, crit=aic, method="g", always=
"+Geslacht")
However, in the output, the variable "Geslacht" is not included...
:
> print(output)
glmulti.analysis
Method: g / Fitting: glm.redefined / IC used: aic
Level: 2 / Marginality: FALSE
>From 100 models:
Best IC: -1212.52589228289
Best model:
[1] "ZBK3 ~ 1 + Onafhankelijk + Hyperactief + Populariteit +
Populariteit:Onafhankelijk + "
[2] "    Populariteit:Zelfconcept + Populariteit:Hyperactief"
[1] "ZBK3 ~ 1 + Onafhankelijk + Hyperactief + Populariteit +
Populariteit:Onafhankelijk"
[1] "ZBK3 ~ 1 + Onafhankelijk + Hyperactief + Populariteit +
Populariteit:Onafhankelijk + "
[2] "    Populariteit:Zelfconcept"
[1] "ZBK3 ~ 1 + Onafhankelijk + Hyperactief + Populariteit +
Populariteit:Onafhankelijk + "
[2] "    Populariteit:Hyperactief"
Evidence weight: 0.0594426803705217
Worst IC: -1203.94552868636
16 models within 2 IC units.
67 models to reach 95% of evidence weight.
Convergence after 700 generations.
Time elapsed: 4.5986722668012 minutes.


Hopefully you can tell me what is wrong here.

Thank you very much in advance,

Kind regards,

Machteld

Machteld Vandecandelaere
University of Leuven - Centre for Educational Effectiveness and Evaluation
Dekenstraat 2, box 3773 - room VHI 2.43
B-3000 Leuven
Tel: + 0032 16 325 780
E-mail:
Machteld.Vandecandelaere@ppw.kuleuven.be<mailto:Machteld.Vandecandelaere@ppw.kuleuven.be>
URL: http://ppw.kuleuven.be/o_en_o/COE



	[[alternative HTML version deleted]]

Solved! Thanx to Vincent Calcagno:

/Dear Machteld,
In the coming 2.0 version of glmulti, there will be a more proper way to
force variables in. In the current version, it has to be done ad-hoc, and
you did it well.
The 'formulas' only show terms that are shuffled by glmulti; the
constant
terms are always there but they do not show up.
To make sure of this, get the fitted model objects themselves
(output at objects[[1]] for instance) and print their summary: they will
include the forced variable in their formula.

Best wishes,
Vincent/

Also possible with the coef(object) call.



-----
Machteld Vandecandelaere
University of Leuven - Centre for Educational Effectiveness and Evaluation
Dekenstraat 2, box 3773 ? room VHI 2.43
B-3000 Leuven
Tel: + 0032 16 325 780
E-mail: Machteld.Vandecandelaere at ppw.kuleuven.be 
URL: http://ppw.kuleuven.be/o_en_o/COE


--
View this message in context:
http://r.789695.n4.nabble.com/Package-glmulti-Include-a-variable-in-ALL-models-tp4648849p4648890.html
Sent from the R help mailing list archive at Nabble.com.