[R-sig-ME] (no subject)
Hank Stevens
HStevens at MUOhio.edu
Mon Jun 9 17:43:00 CEST 2008
Hi Arnaud,
What you are doing seems right to me, but I was wondering where the
large number of observations come from. Are you recording location of
individual animals at different times (GPS tracking?)? Do you have the
same animal at different scales (within a year)? Might scale and
animal be considered separately, or is ANIMAL:SCALEFACT:YEARFACT
simply a way to ID animals in a unique fashion? what is ANNEEFACT?
You could try AIC comparisons of the two models.
Hank
On Jun 9, 2008, at 11:01 AM, ARNAUD_MOSNIER at UQAR.QC.CA wrote:
> Dear lmer users,
>
> It's the first time I use lmer for analysis and I want to be sure I
> correctly wrote my command.
> I need to test link between habitat characteristics used by animal
> we followed (1) and random locations (0). Information was collected
> during several year, several season, on several geographical scales
> and for different individuals.
> I make analysis by season so I do not consider this factor here.
>
> I have not the same individuals for each scale considered, and not
> the same scales for each year considered.
> So I have individuals nested in year and scale.
>
>
>
> First question :
>
>
> Do ANIMAL:SCALEFACT:YEARFACT is the correct way to consider this in
> lmer ?
>
>
> Generalized linear mixed model fit using Laplace
> Formula: REAL_ALEAT ~ (1 | ANIMAL:SCALEFACT:YEARFACT) + ELEVMEAN +
> ELEVSTD
> Data: rawdata
> Subset: SEASON %in% c("PR")
> Family: binomial(logit link)
> AIC BIC logLik deviance
> 293.1 325.3 -142.5 285.1
> Random effects:
> Groups Name Variance Std.Dev.
> ANIMAL:SCALEFACT:YEARFACT (Intercept) 0.0022750 0.047697
> number of obs: 23433, groups: ANIMAL:SCALEFACT:YEARFACT, 20
>
> Estimated scale (compare to 1 ) 0.9111264
>
> Fixed effects:
> Estimate Std. Error z value Pr(>|z|)
> (Intercept) -5.484582 1.249321 -4.390 1.13e-05 ***
> ELEVMEAN -0.004170 0.002086 -1.999 0.0456 *
> ELEVSTD 0.090707 0.010896 8.325 < 2e-16 ***
>
>
>
>
>
> Second and last question :
>
>
> What is the difference between ...
>
> REAL_ALEAT ~ (1 | ANIMAL:SCALEFACT:YEARFACT) + ELEVMEAN + ELEVSTD
>
> and
>
> REAL_ALEAT ~ (1 | ANIMAL) + (1 | SCALEFACT) + (1 | YEARFACT) +
> ELEVMEAN + ELEVSTD
>
>
>
>
>
> Generalized linear mixed model fit using Laplace
> Formula: REAL_ALEAT ~ (1 | ANIMAL) + (1 | SCALEFACT) + (1 |
> ANNEEFACT) + ELEVMEAN + ELEVSTD
> Data: rawdata
> Subset: SEASON %in% c("PR")
> Family: binomial(logit link)
> AIC BIC logLik deviance
> 297.0 345.4 -142.5 285.0
> Random effects:
> Groups Name Variance Std.Dev.
> ANIMAL (Intercept) 0.00159273 0.039909
> SCALEFACT (Intercept) 0.00159324 0.039915
> ANNEEFACT (Intercept) 0.00034151 0.018480
> number of obs: 23433, groups: ANIMAL, 14; SCALEFACT, 14; ANNEEFACT, 3
>
> Estimated scale (compare to 1 ) 0.9112284
>
> Fixed effects:
> Estimate Std. Error z value Pr(>|z|)
> (Intercept) -5.484582 1.251321 -4.383 1.17e-05 ***
> ELEVMEAN -0.004167 0.002088 -1.996 0.0459 *
> ELEVSTD 0.090707 0.010907 8.316 < 2e-16 ***
> ---
> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>
> Correlation of Fixed Effects:
> (Intr) ELEVME
> ELEVMEAN -0.970
> ELEVSTD 0.050 -0.213
>
>
>
> Thanks in advance,
>
> Arnaud
>
> [[alternative HTML version deleted]]
>
> <ATT00001.txt>
Dr. Hank Stevens, Associate Professor
338 Pearson Hall
Botany Department
Miami University
Oxford, OH 45056
Office: (513) 529-4206
Lab: (513) 529-4262
FAX: (513) 529-4243
http://www.cas.muohio.edu/~stevenmh/
http://www.cas.muohio.edu/ecology
http://www.muohio.edu/botany/
"If the stars should appear one night in a thousand years, how would men
believe and adore." -Ralph Waldo Emerson, writer and philosopher
(1803-1882)
More information about the R-sig-mixed-models
mailing list