Hi, I applied a linear mixed effect model in my data using the nlme package. lme2<-lme(distance~temperature*condition, random=~+1|trial, data) and then anova. I want to ask if it is posible to get the least squares means for the interaction effect and the corresponding 95%ci. And then plot this values. Thank you Panagiotis -- View this message in context: http://r.789695.n4.nabble.com/Question-about-linear-mixed-effects-model-nlme-tp3871203p3871203.html Sent from the R help mailing list archive at Nabble.com.
Question about linear mixed effects model (nlme)
3 messages · Panagiotis, Bert Gunter, Ben Bolker
An embedded and charset-unspecified text was scrubbed... Name: not available URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20111004/887e2133/attachment.pl>
Bert Gunter <gunter.berton <at> gene.com> writes:
Below. On Tue, Oct 4, 2011 at 7:34 AM, Panagiotis <pat2 <at> hi.is> wrote:
Hi, I applied a linear mixed effect model in my data using the nlme package. lme2<-lme(distance~temperature*condition, random=~+1|trial, data) and then anova. I want to ask if it is posible to get the least squares means for the interaction effect and the corresponding 95%ci. And then plot this values.
Uh-Oh. You may have unloosed "The Wrath of Khan" -- or at least of Venables. (An explanation of this cryptic remark should follow from others, so please do not ask me what it means if you do not know).
You should probably ask (a version of) this question on the r-sig-mixed-models list instead. What do you mean by "the least squares means for the interaction effect"? How is it different from the estimate of the interaction parameter? You can use the predict() function if you want to calculate predicted values for any particular combination of predictors (you probably want to specify level=0 to get the population-level effects). Getting 'good' confidence intervals for mixed-effect models is surprisingly difficult. If you are willing to ignore the uncertainty of the among-trial variance, you can use a modification of the recipe found at http://glmm.wikidot.com/faq