R help - Jan 2006 - lme4_0.995-2/Matrix_0.995-4 upgrade introduces error messages (change management)

I'll address two issues. The first is today's error message and the
other is change management for contributed packages on CRAN.

TODAY'S ERROR MESSAGE

I switched from the 0.995-1 versions of lme4 and Matrix to those referenced in
the subject line this afternoon. Prior to using these packages on anything else,
I applied them to code that 'worked' (provided numerical results with no
error messages) under the previous set of packages. Since I can't provide
the data, I realize this post may be of limited usefulness. Rightly or wrongly,
I've regressed my R installation back to the 0.995-1 versions of
lme4/Matrix... so I don't think that I continue to have a problem.

R version 2.2.1, 2005-12-20, i386-pc-mingw32 

attached base packages:
[1] "methods"   "stats"     "graphics" 
"grDevices" "utils"     "datasets"
[7] "base"     

other attached packages:
     lme4   lattice    Matrix 
"0.995-2" "0.12-11" "0.995-4"
> options(show.signif.stars=FALSE)
>
m1a<-lmer(cbind(prevented,control.count)~repellant+hour+(1|volunteer)+(1|date),
+ family=binomial(link='probit'), method='Laplace')
Error in dev.resids(y, mu, weights) : argument wt must be a numeric vector of
length 1 or length 219
> # probit doesn't converge
>
m1b<-lmer(cbind(prevented,control.count)~repellant+hour+(1|volunteer)+(1|date),
+ family=binomial, method='Laplace')
Error in dev.resids(y, mu, weights) : argument wt must be a numeric vector of
length 1 or length 219
> # logit is overdispersed
>
m1<-lmer(cbind(prevented,control.count)~repellant+hour+(1|volunteer)+(1|date),
+ family=quasibinomial, method='Laplace')
Error in glm.fit(X, Y, weights = weights, offset = offset, family = family,  : 
	NAs in V(mu)
> m2<-lmer(cbind(prevented,control.count)~hour+(1|volunteer)+(1|date),
+ family=quasibinomial, method='Laplace')
Error in glm.fit(X, Y, weights = weights, offset = offset, family = family,  : 
	NAs in V(mu)

CHANGE MANAGEMENT

Does CRAN keep old versions of contributed packages someplace? If not, the
strategy I've implemented today is to maintain my own repository of
contributed packages that I use. Stuff happens and change management is good.

Chuck

Charles E. White, Senior Biostatistician, MS
Walter Reed Army Institute of Research
503 Robert Grant Ave., Room 1w102
Silver Spring, MD 20910-1557
301 319-9781
Personal/Professional Site:??
http://users.starpower.net/cwhite571/professional/

White, Charles E WRAIR-Wash DC wrote:
> I'll address two issues. The first is today's error message and the
other is change management for contributed packages on CRAN.
> 
> TODAY'S ERROR MESSAGE
> 
> I switched from the 0.995-1 versions of lme4 and Matrix to those referenced
in the subject line this afternoon. Prior to using these packages on anything
else, I applied them to code that 'worked' (provided numerical results
with no error messages) under the previous set of packages. Since I can't
provide the data, I realize this post may be of limited usefulness. Rightly or
wrongly, I've regressed my R installation back to the 0.995-1 versions of
lme4/Matrix... so I don't think that I continue to have a problem.
> 
> R version 2.2.1, 2005-12-20, i386-pc-mingw32 
> 
> attached base packages:
> [1] "methods"   "stats"     "graphics" 
"grDevices" "utils"     "datasets"
> [7] "base"     
> 
> other attached packages:
>      lme4   lattice    Matrix 
> "0.995-2" "0.12-11" "0.995-4"
> 
> 
>>options(show.signif.stars=FALSE)
>>m1a<-lmer(cbind(prevented,control.count)~repellant+hour+(1|volunteer)+(1|date),
> 
> + family=binomial(link='probit'), method='Laplace')
> Error in dev.resids(y, mu, weights) : argument wt must be a numeric vector
of length 1 or length 219
> 
>># probit doesn't converge
>>m1b<-lmer(cbind(prevented,control.count)~repellant+hour+(1|volunteer)+(1|date),
> 
> + family=binomial, method='Laplace')
> Error in dev.resids(y, mu, weights) : argument wt must be a numeric vector
of length 1 or length 219
> 
>># logit is overdispersed
>>m1<-lmer(cbind(prevented,control.count)~repellant+hour+(1|volunteer)+(1|date),
> 
> + family=quasibinomial, method='Laplace')
> Error in glm.fit(X, Y, weights = weights, offset = offset, family = family,
:
> 	NAs in V(mu)
> 
>>m2<-lmer(cbind(prevented,control.count)~hour+(1|volunteer)+(1|date),
> 
> + family=quasibinomial, method='Laplace')
> Error in glm.fit(X, Y, weights = weights, offset = offset, family = family,
:
> 	NAs in V(mu)
> 
> CHANGE MANAGEMENT
> 
> Does CRAN keep old versions of contributed packages someplace? If not, the
strategy I've implemented today is to maintain my own repository of
contributed packages that I use. Stuff happens and change management is good.

Yes, old packages are in
   CRAN/src/contrib/Archive/
You have to compile them from source yourself, though.

Uwe Ligges

> Chuck
> 
> Charles E. White, Senior Biostatistician, MS
> Walter Reed Army Institute of Research
> 503 Robert Grant Ave., Room 1w102
> Silver Spring, MD 20910-1557
> 301 319-9781
> Personal/Professional Site: 
> http://users.starpower.net/cwhite571/professional/
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html

I hope I'm not making your life unnecessarily difficult. As I will
demonstrate below my signature, my original straight application of
lme4_0.995-2/Matrix_0.995-4 is failing without providing any
optimization information. For reference, I've provided optimization
output from lme4_0.995-1/Matrix_0.995-1. Including the lmer command
control=list(PQLmaxIt=0) or control=list(PQLmaxIt=10) produces exactly
the same error as when the commands are not included.

Chuck

Charles E. White, Senior Biostatistician, MS
Walter Reed Army Institute of Research
503 Robert Grant Ave., Room 1w102
Silver Spring, MD 20910-1557
301 319-9781

Personal/Professional Site: 
http://users.starpower.net/cwhite571/professional/ 
> sessionInfo()
R version 2.2.1, 2005-12-20, i386-pc-mingw32 

attached base packages:
[1] "methods"   "stats"     "graphics" 
"grDevices" "utils"
"datasets" 
[7] "base"     

other attached packages:
     lme4   lattice    Matrix 
"0.995-2" "0.12-11" "0.995-4"
>
m1<-lmer(cbind(Treat.Landed,Control.Landed)~Repellant+Hour.After.Applica
tion+(1|Volunteer)+(1|Date),
+ family=quasibinomial, method='Laplace')
Error in glm.fit(X, Y, weights = weights, offset = offset, family family,  : 
	NAs in V(mu)
########################################################################
##
> sessionInfo()
R version 2.2.1, 2005-12-20, i386-pc-mingw32 

attached base packages:
[1] "methods"   "stats"     "graphics" 
"grDevices" "utils"
"datasets" 
[7] "base"     

other attached packages:
     lme4   lattice    Matrix 
"0.995-1" "0.12-11" "0.995-1"
>
m1<-lmer(cbind(Treat.Landed,Control.Landed)~Repellant+Hour.After.Applica
tion+(1|Volunteer)+(1|Date),
+ family=quasibinomial, method='Laplace')
  EM iterations
  0 1643.816 ( 273.179: -0.0214) ( 90.3079: -0.0282)
  1 1640.765 ( 366.319: -0.0102) ( 248.902:-0.00169)
  2 1640.187 ( 437.442:-0.00561) ( 278.171:-0.000393)
  3 1639.953 ( 489.655:-0.00341) ( 285.979:-0.000160)
  4 1639.846 ( 527.983:-0.00221) ( 289.280:-9.04e-005)
  5 1639.793 ( 556.263:-0.00150) ( 291.184:-5.89e-005)
  6 1639.767 ( 577.229:-0.00105) ( 292.439:-4.07e-005)
  7 1639.753 ( 592.835:-0.000748) ( 293.311:-2.89e-005)
  8 1639.745 ( 604.485:-0.000541) ( 293.934:-2.09e-005)
  9 1639.741 ( 613.202:-0.000395) ( 294.386:-1.53e-005)
 10 1639.739 ( 619.736:-0.000291) ( 294.717:-1.12e-005)
 11 1639.738 ( 624.639:-0.000216) ( 294.961:-8.32e-006)
 12 1639.737 ( 628.322:-0.000161) ( 295.142:-6.19e-006)
 13 1639.736 ( 631.090:-0.000120) ( 295.277:-4.62e-006)
 14 1639.736 ( 633.172:-8.97e-005) ( 295.378:-3.46e-006)
 15 1639.736 ( 634.739:-6.72e-005) ( 295.453:-2.59e-006)
  0      1639.74: 0.00157545 0.00338463
  1      1639.74: 0.00156384 0.00338453
  2      1639.74: 0.00156308 0.00338224
  3      1639.74: 0.00156394 0.00338218
  4      1639.74: 0.00156367 0.00338136
  5      1639.74: 0.00156374 0.00338222
  6      1639.74: 0.00156367 0.00338219
  7      1639.74: 0.00156371 0.00338211
  8      1639.74: 0.00156366 0.00338206
  9      1639.74: 0.00156370 0.00338199
 10      1639.74: 0.00156370 0.00338199
  EM iterations
  0 1601.856 ( 639.508: 0.00108) ( 295.684: 0.00140)
  1 1601.814 ( 620.816:0.000875) ( 267.871:0.000262)
  2 1601.802 ( 606.495:0.000663) ( 263.248:6.65e-005)
  0      1601.80: 0.00164882 0.00379870
  1      1601.79: 0.00181161 0.00380177
  2      1601.79: 0.00176152 0.00395670
  3      1601.79: 0.00174046 0.00388111
  4      1601.79: 0.00176162 0.00387505
  5      1601.79: 0.00175271 0.00385493
  6      1601.79: 0.00176027 0.00385121
  7      1601.79: 0.00175618 0.00385019
  8      1601.79: 0.00175420 0.00384200
  9      1601.79: 0.00175615 0.00384174
 10      1601.79: 0.00175577 0.00383981
 11      1601.79: 0.00175577 0.00383981
  EM iterations
  0 1608.593 ( 569.550:-0.000323) ( 260.429:-0.000384)
  1 1608.591 ( 574.148:-0.000245) ( 267.114:-6.56e-005)
  2 1608.590 ( 577.686:-0.000179) ( 268.289:-1.70e-005)
  0      1608.59: 0.00173104 0.00372732
  1      1608.59: 0.00168995 0.00372648
  2      1608.59: 0.00170234 0.00368730
  3      1608.59: 0.00170328 0.00372838
  4      1608.59: 0.00170194 0.00372450
  5      1608.59: 0.00170465 0.00372141
  6      1608.59: 0.00170173 0.00371852
  7      1608.59: 0.00170315 0.00371466
  8      1608.59: 0.00170267 0.00371666
  9      1608.59: 0.00170246 0.00371667
 10      1608.59: 0.00170255 0.00371648
  EM iterations
  0 1608.661 ( 587.354:-6.52e-006) ( 269.072:-3.50e-006)
  1 1608.661 ( 587.452:-4.87e-006) ( 269.135:-7.21e-007)
  2 1608.661 ( 587.525:-3.58e-006) ( 269.148:-2.53e-007)
  0      1608.66: 0.00170206 0.00371543
  1      1608.66: 0.00170148 0.00371542
  2      1608.66: 0.00170148 0.00371524
  3      1608.66: 0.00170148 0.00371524
  4      1608.66: 0.00170148 0.00371524
  EM iterations
  0 1608.660 ( 587.724:-1.09e-008) ( 269.162:5.68e-008)
  1 1608.660 ( 587.724:-5.92e-009) ( 269.161:8.25e-009)
  2 1608.660 ( 587.724:-4.02e-009) ( 269.161:1.07e-009)
  0      1608.66: 0.00170148 0.00371525
  1      1608.66: 0.00170148 0.00371525
  2      1608.66: 0.00170148 0.00371525
  EM iterations
  0 1608.660 ( 587.725:2.30e-010) ( 269.161:4.40e-010)
  1 1608.660 ( 587.725:1.83e-010) ( 269.161:7.32e-011)
  2 1608.660 ( 587.725:1.36e-010) ( 269.161:1.65e-011)
  0      1608.66: 0.00170148 0.00371525
  1      1608.66: 0.00170148 0.00371525
  0      11444.3: -1.57468 -0.114374 0.0891461 0.295675 0.322676
-0.0819240 0.0613226 -0.278625 0.252676 0.297048 0.00170148 0.00371525
  1      10461.4: -1.57468 -0.114375 0.0891456 0.295675 0.322677
-0.0819245 0.0613221 -0.278625 0.252676 0.297048 0.991395 0.146916
  2      10453.7: -1.57501 -0.118004 0.0914816 0.325860 0.316566
-0.101131 0.0995624 -0.273603 0.254018 0.290755 0.987977 0.148760
  3      10452.4: -1.57627 -0.106030 0.110693 0.344082 0.324971
-0.0605686 0.106017 -0.267820 0.245485 0.293816 0.976769 0.154694
  4      10451.5: -1.57797 -0.0856623 0.117621 0.334970 0.344039
-0.0620508 0.146502 -0.274762 0.257380 0.289187 0.968650 0.161734
  5      10450.2: -1.57831 -0.0912595 0.116721 0.344484 0.342080
-0.0541054 0.139780 -0.273456 0.253567 0.291741 0.966484 0.162502
  6      10450.1: -1.58093 -0.0960249 0.120939 0.348483 0.333461
-0.0497757 0.138781 -0.271218 0.250089 0.293405 0.960659 0.169695
  7      10449.8: -1.58338 -0.0947018 0.111567 0.349242 0.340198
-0.0491439 0.142989 -0.272130 0.253299 0.291556 0.953538 0.175865
  8      10449.7: -1.58601 -0.0918766 0.121701 0.342860 0.342149
-0.0469333 0.143516 -0.272566 0.251350 0.294516 0.946432 0.181555
  9      10449.6: -1.58943 -0.0910486 0.119831 0.352018 0.337230
-0.0454451 0.140744 -0.272584 0.256178 0.290521 0.939746 0.188275
 10      10449.5: -1.59166 -0.0935204 0.116089 0.350666 0.341477
-0.0510304 0.145357 -0.270167 0.247932 0.296975 0.933589 0.191757
 11      10449.4: -1.59447 -0.0957850 0.120865 0.343099 0.343630
-0.0473610 0.143548 -0.269864 0.255472 0.290163 0.927840 0.195228
 12      10449.1: -1.59658 -0.0901759 0.115450 0.350433 0.337106
-0.0458197 0.142501 -0.272300 0.252706 0.296086 0.921275 0.197706
 13      10449.0: -1.60106 -0.0990970 0.119617 0.350897 0.341253
-0.0521281 0.143267 -0.269335 0.253346 0.294103 0.914170 0.202652
 14      10448.9: -1.60360 -0.0884343 0.118272 0.344260 0.339332
-0.0487273 0.139916 -0.268830 0.255000 0.292972 0.906302 0.204724
 15      10448.8: -1.60708 -0.0952676 0.116544 0.350083 0.341797
-0.0438318 0.142868 -0.273987 0.255999 0.298785 0.898871 0.208172
 16      10448.6: -1.61004 -0.0936384 0.119330 0.347368 0.338683
-0.0502022 0.147287 -0.265043 0.253930 0.293356 0.891785 0.209803
 17      10448.4: -1.61572 -0.0922092 0.119692 0.348542 0.342165
-0.0453877 0.138443 -0.265999 0.256307 0.294703 0.883089 0.215247
 18      10448.4: -1.61897 -0.0915042 0.119826 0.346438 0.340360
-0.0538914 0.143168 -0.273599 0.260039 0.300747 0.876663 0.215492
 19      10448.1: -1.62170 -0.0959069 0.114654 0.350710 0.339778
-0.0497085 0.142583 -0.264006 0.254848 0.298294 0.869750 0.215881
 20      10448.0: -1.62425 -0.0925439 0.121244 0.342900 0.337405
-0.0463290 0.142340 -0.266442 0.261394 0.295969 0.861491 0.216977
 21      10447.8: -1.63033 -0.0931746 0.119288 0.346826 0.344990
-0.0511664 0.144844 -0.264393 0.258442 0.302008 0.853196 0.217943
 22      10447.7: -1.63145 -0.0916493 0.118219 0.352653 0.337064
-0.0455886 0.138990 -0.264261 0.263362 0.297954 0.845762 0.217530
 23      10447.4: -1.63584 -0.0963003 0.117473 0.344293 0.334583
-0.0471378 0.145940 -0.261730 0.260632 0.302341 0.838119 0.219307
 24      10447.3: -1.63779 -0.0939463 0.112834 0.349643 0.342489
-0.0464460 0.141193 -0.262403 0.264623 0.301302 0.828109 0.218743
 25      10447.1: -1.64064 -0.0902682 0.123597 0.349856 0.343063
-0.0531242 0.139255 -0.260173 0.263578 0.303512 0.820044 0.219590
 26      10446.8: -1.64322 -0.0935995 0.115354 0.350843 0.338844
-0.0474707 0.141749 -0.260037 0.264904 0.305199 0.809290 0.219342
 27      10446.6: -1.64619 -0.0945572 0.119069 0.342510 0.338353
-0.0456821 0.147251 -0.257505 0.267071 0.304463 0.798134 0.219737
 28      10446.3: -1.64999 -0.0947331 0.118412 0.349323 0.341641
-0.0476797 0.140415 -0.256079 0.268139 0.307440 0.786893 0.221035
 29      10446.2: -1.65084 -0.0886278 0.118513 0.347099 0.337841
-0.0542178 0.146136 -0.256303 0.268253 0.309310 0.775711 0.219820
 30      10445.8: -1.65368 -0.0922684 0.119903 0.347534 0.339643
-0.0467803 0.141972 -0.253972 0.271075 0.309212 0.763495 0.220808
 31      10445.7: -1.65432 -0.0934434 0.114649 0.351249 0.341609
-0.0511498 0.147058 -0.254179 0.267805 0.313519 0.751625 0.219946
 32      10445.3: -1.65655 -0.0926586 0.119584 0.348001 0.339997
-0.0492817 0.143213 -0.252695 0.272859 0.310855 0.738776 0.221235
 33      10445.1: -1.65724 -0.0930273 0.115031 0.351075 0.341181
-0.0465371 0.141509 -0.252119 0.268540 0.315546 0.725421 0.220247
 34      10444.7: -1.65927 -0.0927543 0.118641 0.348636 0.340453
-0.0497414 0.144647 -0.251144 0.273531 0.312746 0.711868 0.221569
 35      10444.4: -1.66014 -0.0927254 0.117734 0.349532 0.340062
-0.0444686 0.139415 -0.250910 0.272214 0.315254 0.697846 0.220846
 36      10441.9: -1.69136 -0.104869 0.114555 0.361175 0.357164
-0.0599348 0.140898 -0.232009 0.280566 0.334682 0.511023 0.248880
 37      10436.0: -1.70375 -0.0828958 0.116911 0.349359 0.350304
-0.0512119 0.156769 -0.216600 0.296166 0.352977 0.320550 0.255465
 38      10434.6: -1.70384 -0.0880833 0.121505 0.352662 0.345165
-0.0429543 0.146445 -0.218435 0.306303 0.345228 0.316458 0.254844
 39      10434.2: -1.70422 -0.0934200 0.122745 0.347648 0.341108
-0.0496339 0.146789 -0.218789 0.304552 0.349584 0.299201 0.252209
 40      10433.7: -1.70444 -0.0925859 0.115720 0.350862 0.340392
-0.0469562 0.141048 -0.216638 0.310348 0.347099 0.282092 0.249917
 41      10433.3: -1.70446 -0.0931099 0.118574 0.346919 0.339547
-0.0516847 0.143506 -0.213742 0.309185 0.352231 0.263546 0.246707
 42      10432.6: -1.70471 -0.0941059 0.113513 0.351775 0.338769
-0.0495017 0.139931 -0.208146 0.319561 0.356341 0.225173 0.237935
 43      10432.4: -1.70797 -0.0937163 0.121158 0.343195 0.343718
-0.0530556 0.140490 -0.193316 0.328469 0.374068 0.205060 0.213949
 44      10432.2: -1.71930 -0.0944377 0.117711 0.350098 0.336539
-0.0507862 0.141410 -0.175205 0.348005 0.387963 0.197098 0.190250
 45      10432.1: -1.72151 -0.0949748 0.116203 0.344533 0.342755
-0.0494401 0.137980 -0.173972 0.351927 0.392206 0.197160 0.187444
 46      10432.1: -1.72269 -0.0948574 0.117125 0.346960 0.339561
-0.0515517 0.139812 -0.171871 0.353255 0.393072 0.196681 0.187603
 47      10432.1: -1.72279 -0.0949262 0.117043 0.347170 0.339564
-0.0509965 0.139315 -0.171544 0.353063 0.393329 0.196471 0.187595
 48      10432.1: -1.72306 -0.0947929 0.117129 0.347058 0.339491
-0.0511342 0.139607 -0.171074 0.353431 0.393525 0.195959 0.187574
 49      10432.1: -1.72379 -0.0949714 0.116977 0.347264 0.339587
-0.0507304 0.139303 -0.170283 0.354277 0.394605 0.195653 0.187622
 50      10432.1: -1.73745 -0.0949212 0.117176 0.346877 0.338784
-0.0514425 0.139079 -0.154894 0.371388 0.410965 0.196248 0.187693
 51      10432.0: -1.74347 -0.0950768 0.116443 0.347248 0.340567
-0.0494151 0.141337 -0.145484 0.376406 0.418066 0.196017 0.187831
 52      10432.0: -1.74954 -0.0940385 0.118101 0.347600 0.340126
-0.0478509 0.143164 -0.139359 0.385729 0.424524 0.196150 0.187119
 53      10432.0: -1.75110 -0.0936333 0.117473 0.347457 0.340236
-0.0495394 0.140504 -0.138277 0.385992 0.426828 0.195005 0.188405
 54      10432.0: -1.75362 -0.0954679 0.116741 0.346808 0.339022
-0.0500229 0.140646 -0.136244 0.388144 0.427714 0.194715 0.189040
 55      10432.0: -1.75554 -0.0949921 0.117023 0.347058 0.339582
-0.0496439 0.141035 -0.134474 0.389739 0.430382 0.194467 0.186809
 56      10432.0: -1.75591 -0.0944204 0.117272 0.347378 0.339771
-0.0498677 0.140480 -0.133556 0.391030 0.430849 0.194460 0.186355
 57      10432.0: -1.75622 -0.0945095 0.117389 0.347319 0.339824
-0.0499044 0.140630 -0.132562 0.391734 0.432105 0.194408 0.187217
 58      10432.0: -1.75686 -0.0947003 0.117059 0.347200 0.339619
-0.0495625 0.140879 -0.131663 0.392896 0.432997 0.194223 0.187654
 59      10432.0: -1.75778 -0.0947472 0.117253 0.347206 0.339628
-0.0495707 0.140869 -0.130626 0.393636 0.433921 0.194453 0.186924
 60      10432.0: -1.75877 -0.0944993 0.117209 0.347352 0.339856
-0.0498309 0.140826 -0.129775 0.394661 0.434882 0.194274 0.186688
 61      10432.0: -1.75948 -0.0944810 0.117426 0.347291 0.339689
-0.0496809 0.140634 -0.128936 0.395592 0.435697 0.194265 0.187723
 62      10432.0: -1.75997 -0.0947607 0.117194 0.347261 0.339698
-0.0495161 0.140802 -0.127874 0.396307 0.436611 0.194448 0.186709
 63      10432.0: -1.75997 -0.0946737 0.117224 0.347231 0.339640
-0.0496096 0.140894 -0.127978 0.396412 0.436636 0.194408 0.186713
 64      10432.0: -1.76005 -0.0946805 0.117222 0.347294 0.339677
-0.0495418 0.140844 -0.127949 0.396448 0.436763 0.194315 0.186808
 65      10432.0: -1.76015 -0.0946524 0.117252 0.347272 0.339682
-0.0495778 0.140898 -0.127841 0.396564 0.436825 0.194276 0.186895
 66      10432.0: -1.76026 -0.0946463 0.117233 0.347280 0.339677
-0.0495408 0.140861 -0.127736 0.396679 0.436936 0.194229 0.186908
 67      10432.0: -1.76037 -0.0946545 0.117245 0.347265 0.339681
-0.0495587 0.140902 -0.127618 0.396777 0.437052 0.194202 0.186926
 68      10432.0: -1.76046 -0.0946495 0.117251 0.347297 0.339682
-0.0495406 0.140898 -0.127506 0.396928 0.437150 0.194192 0.186917
 69      10432.0: -1.76055 -0.0946422 0.117237 0.347264 0.339676
-0.0495445 0.140893 -0.127380 0.397023 0.437290 0.194186 0.186930
 70      10432.0: -1.76066 -0.0946661 0.117261 0.347292 0.339685
-0.0495541 0.140922 -0.127271 0.397146 0.437395 0.194167 0.186931
 71      10432.0: -1.76090 -0.0946560 0.117228 0.347271 0.339666
-0.0495355 0.140895 -0.127036 0.397390 0.437615 0.194173 0.186938
 72      10432.0: -1.76094 -0.0946313 0.117272 0.347291 0.339721
-0.0495014 0.140922 -0.126762 0.397597 0.437908 0.194168 0.187017
 73      10432.0: -1.76117 -0.0946427 0.117274 0.347272 0.339696
-0.0495377 0.140953 -0.126576 0.397825 0.438082 0.194202 0.186803
 74      10432.0: -1.76158 -0.0946616 0.117253 0.347316 0.339677
-0.0495179 0.140918 -0.126503 0.397935 0.438176 0.194094 0.186901
 75      10432.0: -1.76140 -0.0946345 0.117251 0.347299 0.339693
-0.0495395 0.140916 -0.126245 0.398162 0.438420 0.194118 0.186986
 76      10432.0: -1.76141 -0.0946363 0.117267 0.347278 0.339688
-0.0495374 0.140918 -0.126244 0.398163 0.438431 0.194115 0.186983
 77      10432.0: -1.76143 -0.0946396 0.117255 0.347286 0.339689
-0.0495327 0.140920 -0.126237 0.398173 0.438435 0.194114 0.186975
 78      10432.0: -1.76147 -0.0946394 0.117259 0.347277 0.339683
-0.0495314 0.140928 -0.126219 0.398188 0.438453 0.194112 0.186960
 79      10432.0: -1.76169 -0.0946427 0.117253 0.347281 0.339694
-0.0495116 0.140935 -0.126132 0.398263 0.438534 0.194134 0.186893
 80      10432.0: -1.76166 -0.0946452 0.117262 0.347287 0.339683
-0.0495108 0.140937 -0.125985 0.398428 0.438681 0.194128 0.186879
 81      10432.0: -1.76189 -0.0946304 0.117263 0.347296 0.339691
-0.0495325 0.140936 -0.125929 0.398483 0.438741 0.194137 0.186972
 82      10432.0: -1.76190 -0.0946413 0.117247 0.347281 0.339689
-0.0495284 0.140929 -0.125781 0.398617 0.438904 0.194099 0.186915
 83      10432.0: -1.76201 -0.0946397 0.117254 0.347278 0.339693
-0.0494968 0.140946 -0.125642 0.398758 0.439026 0.194118 0.186973
 84      10432.0: -1.76218 -0.0946350 0.117265 0.347288 0.339688
-0.0495205 0.140928 -0.125534 0.398871 0.439121 0.194086 0.186879
 85      10432.0: -1.76226 -0.0946434 0.117254 0.347285 0.339675
-0.0495004 0.140968 -0.125402 0.399021 0.439273 0.194094 0.186884
 86      10432.0: -1.76230 -0.0946341 0.117261 0.347287 0.339686
-0.0494978 0.140955 -0.125331 0.399077 0.439348 0.194119 0.186884
 87      10432.0: -1.76230 -0.0946341 0.117261 0.347287 0.339686
-0.0494978 0.140955 -0.125331 0.399077 0.439348 0.194119
0.186884
>
-----Original Message-----
From: Douglas Bates [mailto:dmbates at gmail.com] 
Sent: Friday, January 27, 2006 6:33 PM
To: White, Charles E WRAIR-Wash DC
Subject: Re: lme4_0.995-2/Matrix_0.995-4 upgrade introduces error
messages (change management)

Sorry to hear of the difficulties, Charles.

One thing to try is to turn on the verbose output so fit your models
after setting

options(verbose=TRUE)

Another thing that may be interesting to try is to go to optimization
of the Laplace approximation deviance directly without doing any PQL
iterations.

My theory has been that the PQL iterations help to stabilize the
optimization process but it appears that sometimes they do more harm
than good.

Can you let me know what the verbose output shows?  The thing to watch
for is what I call "ping-ponging" of the PQL iterations.  One set of
iterations converges to one optimum that determines weights that send
it to another optimum that determines weights that sends it back to
the original optimum.


On 1/27/06, White, Charles E WRAIR-Wash DC
<charles.edwin.white at us.army.mil> wrote:
> I'll address two issues. The first is today's error message and the
other is change management for contributed packages on
CRAN.
>
> TODAY'S ERROR MESSAGE
>
> I switched from the 0.995-1 versions of lme4 and Matrix to those
referenced in the subject line this afternoon. Prior to using these
packages on anything else, I applied them to code that 'worked'
(provided numerical results with no error messages) under the previous
set of packages. Since I can't provide the data, I realize this post may
be of limited usefulness. Rightly or wrongly, I've regressed my R
installation back to the 0.995-1 versions of lme4/Matrix... so I don't
think that I continue to have a problem.
>
> R version 2.2.1, 2005-12-20, i386-pc-mingw32
>
> attached base packages:
> [1] "methods"   "stats"     "graphics" 
"grDevices" "utils"
"datasets"
> [7] "base"
>
> other attached packages:
>      lme4   lattice    Matrix
> "0.995-2" "0.12-11" "0.995-4"
>
> > options(show.signif.stars=FALSE)
> >
m1a<-lmer(cbind(prevented,control.count)~repellant+hour+(1|volunteer)+(1
|date),
> + family=binomial(link='probit'), method='Laplace')
> Error in dev.resids(y, mu, weights) : argument wt must be a numeric
vector of length 1 or length 219
> > # probit doesn't converge
> >
m1b<-lmer(cbind(prevented,control.count)~repellant+hour+(1|volunteer)+(1
|date),
> + family=binomial, method='Laplace')
> Error in dev.resids(y, mu, weights) : argument wt must be a numeric
vector of length 1 or length 219
> > # logit is overdispersed
> >
m1<-lmer(cbind(prevented,control.count)~repellant+hour+(1|volunteer)+(1|
date),
> + family=quasibinomial, method='Laplace')
> Error in glm.fit(X, Y, weights = weights, offset = offset, family family, 
:
>         NAs in V(mu)
> >
m2<-lmer(cbind(prevented,control.count)~hour+(1|volunteer)+(1|date),
> + family=quasibinomial, method='Laplace')
> Error in glm.fit(X, Y, weights = weights, offset = offset, family family, 
:
>         NAs in V(mu)
>
> CHANGE MANAGEMENT
>
> Does CRAN keep old versions of contributed packages someplace? If not,
the strategy I've implemented today is to maintain my own repository of
contributed packages that I use. Stuff happens and change management is
good.
>
> Chuck
>
> Charles E. White, Senior Biostatistician, MS
> Walter Reed Army Institute of Research
> 503 Robert Grant Ave., Room 1w102
> Silver Spring, MD 20910-1557
> 301 319-9781
> Personal/Professional Site:
> http://users.starpower.net/cwhite571/professional/
>
>