[R-sig-ME] Warning message with lmer function

[ONKELINX, Thierry]

Use poly() instead of calculating the polynomials by hand. 

And note that your current random effect requires 10 parameters. And you
have only 11 data point within each group. Therefore I would simplify it
to the model below.

lmer(Temp ~ poly(dp1, 3) + (poly(dp1, 1)|group) data = set3)

HTH,

Thierry
------------------------------------------------------------------------
----
ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature
and Forest
Cel biometrie, methodologie en kwaliteitszorg / Section biometrics,
methodology and quality assurance
Gaverstraat 4
9500 Geraardsbergen
Belgium
tel. + 32 54/436 185
Thierry.Onkelinx at inbo.be
www.inbo.be

To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to
say what the experiment died of.
~ Sir Ronald Aylmer Fisher

The plural of anecdote is not data.
~ Roger Brinner

The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of
data.
~ John Tukey

-----Oorspronkelijk bericht-----
Van: r-sig-mixed-models-bounces at r-project.org
[mailto:r-sig-mixed-models-bounces at r-project.org] Namens Ken Beath
Verzonden: woensdag 16 september 2009 12:35
Aan: FMH
CC: r-sig-mixed-models at r-project.org
Onderwerp: Re: [R-sig-ME] Warning message with lmer function

The random effects variances are close to zero, which will cause lots of
problems.

This is possibly a numerical problem, as the values of the fixed effect
estimates are small. Maybe you could try scaling the covariate before
creating the polynomial terms.

Ken

On 16/09/2009, at 8:17 PM, FMH wrote:

> Dear All,
>
> I have a set of data which consist of 1575 groups with 11  
> temperature values in each group. This temperature is recorded  
> across 11 different depths of the sea. The R script are shown below:
>
> ############################################################
> #Temp : Temperature
> #dp1, dp2, dp3 : covariate with respect to linear, quadratic and  
> cubic terms
> #group : 1575 groups in which there are 11 observations in each group
> #sub3 : Data set 1
> #set3 : Data set 2
>
> dp1 <- rep(rev(seq(1,51, by = 5)),1575)
> dp2 <- dp1^2
> dp3 <- dp1^3
> group <- rep(1:1575, each = 11)
> set3 <- data.frame(sub3, dp1, dp2, dp3)
> (lm.lme3 <- lmer(Temp ~ dp1 + dp2 + dp3 + (dp1 + dp2 + dp3|group),  
> data = set3))
> ############################################################
>
>
> I tried to fit a linear mixed model via lmer function, with all the  
> fixed and random effects are included, but there is a 'Warning'  
> message given after the output, as shown below.
>
>
> ##############################################
> Linear mixed model fit by REML
> Formula: Temp ~ dp1 + dp2 + dp3 + (dp1 + dp2 + dp3 | group)
>    Data: set3
>    AIC   BIC logLik deviance REMLdev
>  65627 65743 -32799    65539   65597
> Random effects:
>  Groups   Name        Variance   Std.Dev.   Corr
>  group    (Intercept) 4.8336e-01 6.9524e-01
>           dp1         5.2199e-04 2.2847e-02  0.000
>           dp2         3.0528e-07 5.5252e-04  0.000  0.000
>           dp3         7.8888e-11 8.8819e-06 -0.966  0.000  0.000
>  Residual             1.9939e+00 1.4120e+00
> Number of obs: 17325, groups: group, 1575
> Fixed effects:
>               Estimate Std. Error t value
> (Intercept)  1.363e+01  4.043e-02   337.1
> dp1         -2.930e-01  6.328e-03   -46.3
> dp2          3.924e-03  2.870e-04    13.7
> dp3         -1.900e-05  3.628e-06    -5.2
> Correlation of Fixed Effects:
>     (Intr) dp1    dp2
> dp1 -0.739
> dp2  0.616 -0.959
> dp3 -0.563  0.904 -0.982
>
> Warning message:
> In mer_finalize(ans) : false convergence (8)
> #############################################
>
>
> Does the output given is valid? Could someone please advice on this  
> message.
>
> Thank you
> Fir
>
>
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>

_______________________________________________
R-sig-mixed-models at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

Druk dit bericht a.u.b. niet onnodig af.
Please do not print this message unnecessarily.

Dit bericht en eventuele bijlagen geven enkel de visie van de schrijver weer 
en binden het INBO onder geen enkel beding, zolang dit bericht niet bevestigd is
door een geldig ondertekend document. The views expressed in  this message 
and any annex are purely those of the writer and may not be regarded as stating 
an official position of INBO, as long as the message is not confirmed by a duly 
signed document.