[R] proportional splitplot with glmmPQL

[Ronaldo Reis Jr.]

Hi ALL,

sorry by this repeated question, but I'm a student and I dont have
money to buy any book more for now and I dont find any answer in network.

I have a splitplot with binomial errors (proportion data).

I try to make a model with glmmPQL.

But my doubt are:

Exit any method to compare models made with glmmPQL?
How know about the variables significance and model significance?
How make a model simplification (variables simplification and levels
amalgamation)?

If it is not possible with glmmPQL, how is the best method for make this
analysis? Using lme with the data transformed in log(y/n-y), where y is the
observation and n is the total.

Thanks for ALL and sorry again.

Inte
Ronaldo

--
|   //|\\   [*****************************]
|| ( õ õ )  [Ronaldo Reis Júnior          ]
|     V     [ESALQ/USP-Entomologia, CP-09 ]
||  / l \   [13418-900 Piracicaba - SP    ]
|  /(lin)\  [Fone: 19-429-4199 r.229      ]
||/(linux)\ [chrysopa at insecta.ufv.br      ]
|/ (linux) \[ICQ#: 5692561                ]
||  ( x )   [*****************************]
||| _/ \_ Powered by Gnu/Debian Woody
-----------------------------------
Insecta - Entomologia
Departamento de Biologia Animal
Universidade Federal de Viçosa