[R-sig-ME] syntax for indicating fixed covariates in nlmer ?

Ben Bolker bolker at ufl.edu
Wed Mar 11 12:20:32 CET 2009


Laurent Gentzbittel wrote:
> Dear all,
> 
> I would like to use the nlmer for fitting logistic curves to disease
> index curves. I'm typically comparing different plant lines (several
> plants from each line are subjected to a pathogen) and would like to
> test if the 'population' (fixed effects) parameters of the curves are
> the same between the lines.
> I need to specify a random effect for each parameter to account for
> plant-to-plant variability and a fixed effect for some parameter to
> account for 'general' differences among lines.
> I previously used the nlme function of the nlme package and was able to
> specify both fixed and random effects for each parameter, including a
> fixed covariate 'line' for example.
> 
> I would like to switch to the lme4 package, but I was unable to find how
> to specify fixed covariates (fixed effects) in the formula.
> A search in the different mail lists and Googl'ing  was also unsuccessful.
> 

   The fixed effects formula looks just like that in nlme.  The random
effects are specified differently -- where nlme would use (for example)

fixed=response~(something),random=~1|plant+a|plant+b|plant

nlmer would use

formula = response~(something)+(1|plant)+(a|plant)+(b|plant)

  Ben Bolker


-- 
Ben Bolker
Associate professor, Biology Dep't, Univ. of Florida
bolker at ufl.edu / www.zoology.ufl.edu/bolker
GPG key: www.zoology.ufl.edu/bolker/benbolker-publickey.asc




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