[R] Parameter Estimates needed from lmer output

[David Cross]

If I remember correctly, coef(m1) would do it ... but it has been a while since I last used lmer, and I am working only from memory.

Cheers

David Cross
d.cross at tcu.edu
www.davidcross.us




On May 18, 2011, at 6:29 PM, Stephen Peterson wrote:

> Hello,
> I am looking for some help on how I may be able to view estimated
> values for 3 response variables with 1 fixed and 1 random effect using
> lmer.
> My data is proportional cover of three habitat variables (bare ground,
> grass cover, shrub cover) that was collected during 3 years (1976,
> 2000, 2010) on 5 study plots, each plot divided into 50 m square
> cells.
> Portion of dataset (proportions were log transformed)
> year	plot	cell_id bare_trans	grass_trans	shrub_trans
> 0	wh	whi1	-0.678240631	-0.892213913	-0.158328393
> 0	wh	whi2	-0.774640426	-0.745665597	-0.164722747
> 0	wh	whi3	-0.600670894	-0.545056465	-0.30835479
> 0	wh	whi4	-0.461018617	-0.704273962	-0.315083353
> 0	wh	whi5	-0.518221954	-0.643432282	-0.303575808
> 0	wh	whi6	-0.598043065	-0.588487184	-0.286051968
> 0	wh	whi7	-0.581336622	-0.356760604	-0.4880035
> 0	wh	whj1	-0.650114241	-0.706560469	-0.215255255
> 
> I am treating the group of response variables (bare_trans,
> grass_trans, shrub_trans) as one multivariate response.
> The year (0, 1, 2) is my fixed effect and cell_id (whi1 . . .) is my
> random effect.
> 
> My model is:
> m1 <- lmer(cbind(bare_trans,grass_trans,shrub_trans) ~ year +
> (1|cell_id),data=whdata)
> 
> Summary output is:
> Linear mixed model fit by REML
> Formula: cbind(bare_trans, grass_trans, shrub_trans) ~ year + (1 | cell_id)
>   Data: whdata
>    AIC    BIC logLik deviance REMLdev
> -97.86 -88.14  52.93   -119.1  -105.9
> Random effects:
> Groups   Name        Variance Std.Dev.
> cell_id  (Intercept) 0.000000 0.00000
> Residual             0.014523 0.12051
> Number of obs: 84, groups: cell_id, 28
> 
> Fixed effects:
>            Estimate Std. Error t value
> (Intercept) -0.53781    0.02079  -25.87
> year         0.24182    0.01610   15.02
> 
> Correlation of Fixed Effects:
>     (Intr)
> year -0.775
> 
> What is missing from this output that I need are estimated
> coefficients of the 3 response variables (bare_trans, grass_trans,
> shrub_trans) for each year (0, 1, 2), standard errors and p-values.
> 
> Any idea if lmer even generates these estimates? And if so, is there a
> way of digging them out of the R blackbox?
> If not, if anyone has suggestions on a more appropriate package to use
> that would be great.
> I essentially want to perform a MANOVA on my 3 response variables
> while accounting for fixed and random effects.
> 
> Any help would be appreciated.
> 
> Thank you,
> Stephen L. Peterson
> Utah State University
> 
> ______________________________________________
> 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.