lrt.glm {datana} | R Documentation |
Computes a likelihood ratio test between a reduced model and a full model. Both models must be already fitted using and R function.
Description
Computes a likelihood ratio test between a reduced model (modr) and a full model (modr). Both models must be previously fitted by maximum likelihood using an R function such as nlme() and such, that are part of the generalized lineal models.
Usage
lrt.glm(modr, modf)
Arguments
modr |
is the object containing a previously fitted reduced model, using a glm-type of function, having less parameters than modf. |
modf |
is the object containing a previously fitted full model, using a glm-type of function, having more parameters than modr. |
Details
Double-check the order of the reduced and full model, before of using the model
Value
This function returns an object having the following elements: "loglik.Modr" maximized log-likelihood of modr; "loglik.Modf" maximized log-likelihood of modf; "dif.loglik" difference in log-likelihood between both models, and "dif.df" difference in degrees of freedong of both models, and "p-value" is the p-value for the LRT.
Author(s)
Christian Salas-Eljatib.
References
Pinheiro JC, and Bates DM. 2000. Mixed-effects models in S and Splus. Springer-Verlag, New York, NY. 528 p.
Examples
#not yet implemented