[R] correlation structure in lme without random effect

Benjamin Tyner btyner at stat.purdue.edu
Mon Oct 30 18:59:41 CET 2006


Thanks!

Douglas Bates wrote:
> On 10/30/06, Benjamin Tyner <btyner at stat.purdue.edu> wrote:
>> I was hoping to fit along the lines of
>>
>> g<-gl(20,5)
>> y<-runif(100)
>> fit<-lme(fixed=y~g,correlation=corAR1(0,~1|g))
>>
>> But I get the error "Incompatible formulas for groups in "random" and
>> "correlation""
>
> Use the gls function in the nlme package to fit a model with
> correlation but no random effects specification.



More information about the R-help mailing list