[R] lmer: Estimated variance-covariance is singular, false convergence

From: Martin Klipp <a9913395_at_unet.univie.ac.at>
Date: Wed, 13 Feb 2008 23:05:43 +0100


Dear R Community!

We analyse the impact of climbing activity on cliff vegetation. During our fieldwork, we recorded 90 Transects in 3 climbing sites. The aim is to see, if the plant cover (response: Cover) is influenced only by crevice availability (predictor: Cracs), or, additional, by the distance to the climbing route (predictor: Distance). Six plots are nested within one Transect (ID.Transekt), the Transects are nested within the climbing site (Site).

I tried to fit two linear mixed models with lmer(), and to compare them by anova():

 > cracks.lmer <- lmer (Cover ~ Cracks + Distance + (Cracks + Distance |  > + Site / ID.Transekt) , method = ("ML"))

Warning message:
In .local(x, ..., value) :

   Estimated variance-covariance for factors ‘ID.Transekt:Site’, ‘Site’ is singular

 > cracks.lmer1 <- lmer (Cover ~ Cracks + (Cracks | Site / ID.Transekt) ,  > + data = Lmm.spalten.corr, method = ("ML"))

Warning messages:
1: In .local(x, ..., value) :

   Estimated variance-covariance for factor ‘Site’ is singular

2: In .local(x, ..., value) :

   nlminb returned message false convergence (8)

For this one, i changed the maxIter and msMaxIter up to 50000000, but the Warning messages remained the same.
I have many "0" values for "Cover" and "Cracks" and tried to exclude some of them in different ways (as all transects without vegetation cover), but this did not help. The current version deals with all transects wich have at least two plots with crevices. The correlation Intercept/Cracks ist -1.000 for both models. I attachet a symmary of the used data.frame, a randomly selected subset and the summarys of cracks.lmer and cracks.lmer1 in a *.txt file (modified with kate). Some xyplots are attached as *.png. (The error message on the plots means "Error using packet 5, NA/NaN/Inf in externem Funktionsaufruf", excluding these transects does not change the lmer message.)

My question is: Is it possible and does it make sense to fit these date with lmer? I do get outputs, but they dont seem really reliable to me.

I hope that this is not a stupid and unnecessary question, but i did't really find a answer in the current mailing-lists. Thanks a lot!

Martin Klipp

R version 2.6.2 (2008-02-08)
i486-pc-linux-gnu
Kubuntu 6.06.1, precompieled packages from the CRAN-mirror. lme4_0.99875-9
Matrix_0.999375-4
lattice_0.17-2
nlme_3.1-86

>summary (Lmm.spalten.corr)

  ID.Transekt     Distance       Cracks           Cover       
 11     :  6   Min.   :1.0   Min.   :  0.00   Min.   : 0.000  
 12     :  6   1st Qu.:2.0   1st Qu.:  6.00   1st Qu.: 0.000  
 14     :  6   Median :3.5   Median : 25.00   Median : 0.100  
 16     :  6   Mean   :3.5   Mean   : 31.77   Mean   : 1.133  
 17     :  6   3rd Qu.:5.0   3rd Qu.: 45.00   3rd Qu.: 0.700  
 18     :  6   Max.   :6.0   Max.   :244.00   Max.   :33.650  
 (Other):450                                                  
          Site    
 Brueggele  :162  

 Kirchler :168
 Plaetzwiese:156

#Subset of 20 randomly selectet Transects from Lmm.spalten.corr:

Grouped Data: Cover ~ Cracks + Distance | Site/ID.Transekt

    ID.Transekt Distance Cracks Cover        Site
1            11        1     25  0.14 Plaetzwiese
2            11        2     30  7.10 Plaetzwiese
3            11        3     53  0.80 Plaetzwiese
4            11        4     25  0.70 Plaetzwiese
5            11        5     25  0.90 Plaetzwiese
6            11        6     28  0.00 Plaetzwiese
19           14        1     10  0.00 Plaetzwiese
20           14        2     10  0.00 Plaetzwiese
21           14        3      0  0.00 Plaetzwiese
22           14        4     10  0.00 Plaetzwiese
23           14        5      0  0.00 Plaetzwiese
24           14        6      0  0.00 Plaetzwiese
31           16        1     25  0.00 Plaetzwiese
32           16        2     20  0.00 Plaetzwiese
33           16        3     38  8.20 Plaetzwiese
34           16        4     45 16.00 Plaetzwiese
35           16        5     60  6.50 Plaetzwiese
36           16        6      0  0.00 Plaetzwiese
49           19        1     45  0.18 Plaetzwiese
50           19        2     17  0.10 Plaetzwiese
51           19        3      7  0.00 Plaetzwiese
52           19        4      8  0.00 Plaetzwiese
53           19        5      0  0.00 Plaetzwiese
54           19        6      5  0.00 Plaetzwiese
67           22        1     44  0.00 Plaetzwiese
68           22        2     25  0.00 Plaetzwiese
69           22        3      0  0.00 Plaetzwiese
70           22        4      0  0.00 Plaetzwiese
71           22        5     14  0.26 Plaetzwiese
72           22        6      0  0.00 Plaetzwiese
85           25        1     65  0.35 Plaetzwiese
86           25        2     43  0.81 Plaetzwiese
87           25        3     31  2.00 Plaetzwiese
88           25        4      0  0.00 Plaetzwiese
89           25        5      0  0.00 Plaetzwiese
90           25        6      6  0.00 Plaetzwiese
109          29        1      0  0.00 Plaetzwiese
110          29        2      0  0.00 Plaetzwiese
111          29        3    105  0.85 Plaetzwiese
112          29        4     20  0.80 Plaetzwiese
113          29        5      5  0.10 Plaetzwiese
114          29        6      0  0.00 Plaetzwiese
199          44        1     72  0.85    Kirchler
200          44        2     20  0.05    Kirchler
201          44        3      3  0.00    Kirchler
202          44        4      0  0.00    Kirchler
203          44        5      9  0.00    Kirchler
204          44        6      0  0.00    Kirchler
205          45        1     35  1.75    Kirchler
206          45        2     35  1.20    Kirchler
207          45        3     19  0.30    Kirchler
208          45        4     48  0.25    Kirchler
209          45        5     30  0.00    Kirchler
210          45        6     17  0.00    Kirchler
223          48        1     48  0.12    Kirchler
224          48        2      5  0.00    Kirchler
225          48        3     10  0.00    Kirchler
226          48        4     43  2.25    Kirchler
227          48        5     39  2.90    Kirchler
228          48        6     46  0.75    Kirchler
235          50        1     24  0.25    Kirchler
236          50        2     93  0.25    Kirchler
237          50        3      0  0.00    Kirchler
238          50        4      7  0.00    Kirchler
239          50        5      0  0.00    Kirchler
240          50        6      7  0.00    Kirchler
295          60        1     55  3.10    Kirchler
296          60        2     27  1.50    Kirchler
297          60        3     50  3.60    Kirchler
298          60        4     75  5.50    Kirchler
299          60        5     65  1.25    Kirchler
300          60        6      0  0.00    Kirchler
307          62        1     75  2.00    Kirchler
308          62        2     20  0.20    Kirchler
309          62        3     10  0.30    Kirchler
310          62        4      0  0.00    Kirchler
311          62        5     10  0.00    Kirchler
312          62        6     50  2.00    Kirchler
319          64        1    109  5.20    Kirchler
320          64        2     48  2.72    Kirchler
321          64        3     67  8.00    Kirchler
322          64        4     23  0.00    Kirchler
323          64        5      0  0.00    Kirchler
324          64        6      0  0.00    Kirchler
331          66        1     76  0.01    Kirchler
332          66        2     83  1.30    Kirchler
333          66        3     24  0.30    Kirchler
334          66        4    100 13.80    Kirchler
335          66        5     25  0.00    Kirchler
336          66        6      0  0.00    Kirchler
349          69        1    103  6.05    Kirchler
350          69        2     63  1.35    Kirchler
351          69        3     20  0.00    Kirchler
352          69        4     15  0.20    Kirchler
353          69        5      0  0.00    Kirchler
354          69        6     15  0.00    Kirchler
367          72        1    244  0.15   Brueggele
368          72        2    180  1.55   Brueggele
369          72        3    165 11.95   Brueggele
370          72        4    137  5.90   Brueggele
371          72        5    145  7.20   Brueggele
372          72        6     93  4.30   Brueggele
397          77        1      6  0.02   Brueggele
398          77        2      7  0.00   Brueggele
399          77        3     25  0.00   Brueggele
400          77        4     60  0.20   Brueggele
401          77        5     59  0.70   Brueggele
402          77        6     42  0.05   Brueggele
433          83        1     55  1.10   Brueggele
434          83        2     45  1.00   Brueggele
435          83        3      0  0.00   Brueggele
436          83        4     25  0.00   Brueggele
437          83        5     15  0.10   Brueggele
438          83        6      5  0.00   Brueggele

>cracks.lmer <- (lmer (Cover ~ Cracks + Distance + (Cracks + Distance | Site / ID.Transekt) , data = Lmm.spalten.corr, method = ("ML"))
>summary (cracks.lmer)

Linear mixed-effects model fit by maximum likelihood Formula: Cover ~ Cracks + Distance + (Cracks + Distance | Site/ID.Transekt)

   Data: Lmm.spalten.corr
  AIC BIC logLik MLdeviance REMLdeviance  1918 1981 -944 1888 1904 Random effects:

 Groups           Name        Variance   Std.Dev.   Corr          
 ID.Transekt:Site (Intercept) 1.8358e-01 4.2846e-01               
                  Cracks      2.1744e-03 4.6631e-02 -1.000        
                  Distance    1.0280e-09 3.2062e-05  0.002 -0.002 
 Site             (Intercept) 1.0328e-09 3.2137e-05               
                  Cracks      1.0435e-09 3.2303e-05 0.044         
                  Distance    1.0280e-09 3.2062e-05 0.057  0.122  
 Residual                     2.0560e+00 1.4339e+00               
number of obs: 486, groups: ID.Transekt:Site, 81; Site, 3

Fixed effects:

             Estimate Std. Error t value
(Intercept) -0.585846   0.194017  -3.020
Cracks       0.040834   0.006053   6.746
Distance     0.099521   0.041494   2.398

Correlation of Fixed Effects:
         (Intr) Cracks
Cracks   -0.455       

Distance -0.821 0.080

>cracks.lmer1 <- (lmer (Cover ~ Cracks + Distance + (Cracks + Distance | Site / ID.Transekt) , data = Lmm.spalten.corr, method = ("ML"))
>summary (cracks.lmer1)

Linear mixed-effects model fit by maximum likelihood Formula: Cover ~ Cracks + (Cracks | Site/ID.Transekt)

   Data: Lmm.spalten.corr
  AIC BIC logLik MLdeviance REMLdeviance  1910 1943 -946.9 1894 1905 Random effects:

 Groups           Name        Variance   Std.Dev.   Corr   
 ID.Transekt:Site (Intercept) 1.9162e-01 4.3774e-01        
                  Cracks      2.1710e-03 4.6594e-02 -1.000 
 Site             (Intercept) 1.1199e-09 3.3465e-05        
                  Cracks      1.0434e-09 3.2302e-05 0.261  
 Residual                     2.0869e+00 1.4446e+00        
number of obs: 486, groups: ID.Transekt:Site, 81; Site, 3

Fixed effects:

             Estimate Std. Error t value
(Intercept) -0.203951   0.111849  -1.823
Cracks       0.039669   0.006037   6.571

Correlation of Fixed Effects:
       (Intr)

Cracks -0.688



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.
cracks.png cracks_distance.png
Received on Wed 13 Feb 2008 - 22:50:40 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Wed 13 Feb 2008 - 23:30:14 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive