[R-sig-ME] Jointly testing several parameters of factor variable

Greg Snow Greg.Snow at imail.org
Mon Feb 2 19:27:30 CET 2009


For number (1) below you can either set up a set of contrasts such that when some of the contrasts are 0, the equality constraints of interest hold, then look at those specific contrasts (see functions 'contrasts' and 'C' for ways to specify the contrasts).  Or you can create variables recoded such that using one set of the variables is the full model and the other set is your constrained model, then just do a full/reduced model comparison between the 2 groups of variables.

-- 
Gregory (Greg) L. Snow Ph.D.
Statistical Data Center
Intermountain Healthcare
greg.snow at imail.org
801.408.8111


> -----Original Message-----
> From: r-sig-mixed-models-bounces at r-project.org [mailto:r-sig-mixed-
> models-bounces at r-project.org] On Behalf Of Stijn Ruiter
> Sent: Saturday, January 31, 2009 4:31 AM
> To: r-sig-mixed-models at r-project.org
> Subject: [R-sig-ME] Jointly testing several parameters of factor
> variable
> 
> Hi all,
> I am estimating a lmer model and now I would like to test whether some
> coefficients for specific levels of factor variable are equal.
> I am thinking of either:
> (1) applying equality constraints on several parameters for specific
> levels of a factor variable (and use that to do LRT against model
> without equality constraints), but I have no clue on how to that; or
> (2) do a Wald test in which several parameter estimates of a factor
> variable are jointly tested. Although I noticed "wald.test" from the
> "aod" package, but I don't quite follow how to apply that to my lmer
> model.
> 
> Any suggestions on how to jointly test several parameters of a lmer
> model (preferably applied to factor variables)?
> 
> Stijn
> 
> --
> Best regards,
> 
> Stijn Ruiter
> Department of Sociology / ICS
> Radboud University Nijmegen
> P.O. Box 9104
> 6500 HE Nijmegen
> Netherlands
> 
> Phone: + 31 24 361 2272
> Fax:   + 31 24 361 2399
> 
> Visiting address:
> Thomas van Aquinostraat 4.01.71
> Nijmegen
> 
> website: http://oase.uci.ru.nl/~sruiter
> 
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models




More information about the R-sig-mixed-models mailing list