[R] LMER: How to specify Random Effects

[Ubuntu Diego]

Thanks for your feedback.
Actually Plant is nested since High and Low are qualitative relative 
values (High in Michigan is not the same as High in Sienna).


S Ellison wrote:
> The Plant classification is not nested; it's an effect across all
> countires and states and probably a fixed effct (assuming you want to
> measure its size or significance). But the state is nested in country.
>
> That would suggest to me 
> lmer(Reaction~Drug+Plant+(1|Country/State),...)
>
> (or Plant*Drug, if you want an interaction)
>
> However, though your states are nested in countrythey are readily
> identifiable by lmer as different for each country (you do not have
> country1: State1, State2..., Country2:State1,State2..) so the nesting by
> country probably doesn't need to be specified. I think the above would,
> _in this case_ be almost identical to 
>
> lmer(Reaction~Drug+Plant+(1|Country)+(1|State),...)
>
>
>
>
>
>   
>>>> Ubuntu Diego <ubuntu.diego at gmail.com> 01/12/2009 16:22:09 >>>
>>>>         
> I saw different specifications for Random Effects and I'm confused
> about 
> the use of "/" and  the use of "(0+...|)" .
> Let say we have a nested structure where some countries have some 
> several plants in different states and we measure the reaction to a
> drug.
>
> The list of Countries = USA, France, Italy
> The States for USA = Michigan, Florida, California
> The States for France =  Paris, Orleans
> The States for Italy =  Venezia, Sienna, Florence, Rome, Napoli ,
> Sicilia
>
> Plants were classified as High and Low
>
> is this the way to specify a possible model ?
>
> lmer(Reaction ~ Drug + (1| Country / State / Plant) , data)
>
> or should I use something like this
> A) lmer(Reaction ~ Drug + (0| Country / State / Plant) , data)
> B) lmer(Reaction ~ Drug + (1| Country ) + (0+Country | State / Plant) ,
>
> data)
> C) lmer(Reaction ~ Drug + (1| Country ) + (0+Country | State / Plant) +
>
> (0+Country + State | Plant), data)
> D) lmer(Reaction ~ Drug + (1| Country ) + (0+Country | State ) + 
> (0+Country + State | Plant), data)
>
> Thanks
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help 
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html 
> and provide commented, minimal, self-contained, reproducible code.
>
> *******************************************************************
> This email and any attachments are confidential. Any u...{{dropped:9}}