lmerTest

[skanskan]

It would be great if broom also showed the output of lmerTest alongside the lme4 output.
This package simply adds a new column with the p-value.

[bbolker]

I'm in favor of this in principle. It will only work easily if the original model was fitted with lmerTest, but that seems like a reasonable restriction.

[skanskan]

Anyway, What would be the difference between coef(summary(mylmerTestmodel)) and tidy(mylmerTestmodel) ?

[bbolker]

A couple of small but potentially important differences:

• column names; term as a column rather than as row names (these differences don't matter for human readers, but do for programmatic interfaces)
• coef(summary(mylmerTestmodel)) will be very close to tidy(mylmerTestmodel,effects="fixed"), but tidy() offers a wider range of options (report estimates of RE var/cov/sd/cor, conditional modes; report confidence intervals; etc.)

[skanskan]

OK, thanks @bbolker

[bbolker]

(I'm a little baffled that I can't find this comment when I view it on GitHub: #126) The available `effects` values (**not** well documented) are `fixed`, `ran_pars` (for variances/covariances or std devs/correlations of random effects) or `ran_modes` (for conditional modes/BLUPs). Documentation leaves a lot to be desired. It's possible that some of this stuff has diverged in my fork.

…

On Sun, Aug 20, 2017 at 1:11 PM, niroshar ***@***.***> wrote: I am trying to display tidy output for random effects using lmer function in lme4. Why I am getting the following error. tidy(model_ri,effects="random") Error in tidy.merMod(model_ri, effects = "random") : unknown effect type random — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#126 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAE0Rn0qEp0XDcuNF2-wAZBvtBA1unIwks5saGjAgaJpZM4IZnev> .

[niroshar]

@bbolker I removed the comment since I was able to get it solved using "effects = ran_pars" few minutes after.
Thank you very much.

[bbolker]

In future, probably worth posting your solution rather than removing the comment (others will be similarly confused in the future ...)

…

On 17-08-20 03:29 PM, niroshar wrote: @bbolker <https://github.com/bbolker> I removed the comment since I was able to get it solved using "effects = ran_pars" few minutes after. Thank you very much. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#126 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAE0Rt-MAWvwFDXWdlx0r3Ds1HuTR2m2ks5saIkIgaJpZM4IZnev>.

[niroshar]

That's true. Yeah, I will do so. On Sun, Aug 20, 2017 at 3:42 PM, Ben Bolker <notifications@github.com> wrote:

…

In future, probably worth posting your solution rather than removing the comment (others will be similarly confused in the future ...) On 17-08-20 03:29 PM, niroshar wrote: > @bbolker <https://github.com/bbolker> I removed the comment since I was > able to get it solved using "effects = ran_pars" few minutes after. > Thank you very much. > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <#126 (comment)>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/AAE0Rt- MAWvwFDXWdlx0r3Ds1HuTR2m2ks5saIkIgaJpZM4IZnev>. > — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#126 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/APN_bbB0xcs5YgiVnb7akUyu4cbXUusoks5saJo7gaJpZM4IZnev> .

-- Best regards, Nirosha.

[bbolker]

I think this works now in broom.mixed. Can someone check? (e.g. see examples of ?tidy.merMod in broom.mixed) If not, probably best to open as a broom.mixed issue (and close this in any case)

[alexpghayes]

Just ran the examples in tidy.merMod:

library(broom)
library(broom.mixed)
#> 
#> Attaching package: 'broom.mixed'
#> The following object is masked from 'package:broom':
#> 
#>     tidyMCMC
library(lme4)
#> Warning: package 'lme4' was built under R version 3.4.4
#> Loading required package: Matrix
#> Warning: package 'Matrix' was built under R version 3.4.4


lmm1 <- lmer(Reaction ~ Days + (Days | Subject), sleepstudy)
tidy(lmm1)
#>     effect    group                 term     estimate std.error statistic
#> 1    fixed    fixed          (Intercept) 251.40510485  6.824556 36.838311
#> 2    fixed    fixed                 Days  10.46728596  1.545789  6.771485
#> 3 ran_pars  Subject       sd_(Intercept)  24.74044759        NA        NA
#> 4 ran_pars  Subject              sd_Days   5.92213327        NA        NA
#> 5 ran_pars  Subject cor_(Intercept).Days   0.06555133        NA        NA
#> 6 ran_pars Residual       sd_Observation  25.59181589        NA        NA
tidy(lmm1, effects = "fixed")
#>   effect        term  estimate std.error statistic
#> 1  fixed (Intercept) 251.40510  6.824556 36.838311
#> 2  fixed        Days  10.46729  1.545789  6.771485
tidy(lmm1, effects = "fixed", conf.int=TRUE)
#>   effect        term  estimate std.error statistic   conf.low conf.high
#> 1  fixed (Intercept) 251.40510  6.824556 36.838311 238.029221 264.78099
#> 2  fixed        Days  10.46729  1.545789  6.771485   7.437595  13.49698
tidy(lmm1, effects = "fixed", conf.int=TRUE, conf.method="profile")
#> Computing profile confidence intervals ...
#>   effect        term  estimate std.error statistic   conf.low conf.high
#> 1  fixed (Intercept) 251.40510  6.824556 36.838311 237.680695 265.12951
#> 2  fixed        Days  10.46729  1.545789  6.771485   7.358653  13.57592
tidy(lmm1, effects = "ran_modes", conf.int=TRUE)
#> Warning: package 'bindrcpp' was built under R version 3.4.4
#>       effect   group        term level    estimate std.error     conf.low
#> 1  ran_modes Subject (Intercept)   308   2.2585654 12.070837 -21.39983947
#> 2  ran_modes Subject (Intercept)   309 -40.3985769 12.070837 -64.05698171
#> 3  ran_modes Subject (Intercept)   310 -38.9602458 12.070837 -62.61865063
#> 4  ran_modes Subject (Intercept)   330  23.6904985 12.070837   0.03209363
#> 5  ran_modes Subject (Intercept)   331  22.2602027 12.070837  -1.39820213
#> 6  ran_modes Subject (Intercept)   332   9.0395259 12.070837 -14.61887895
#> 7  ran_modes Subject (Intercept)   333  16.8404311 12.070837  -6.81797371
#> 8  ran_modes Subject (Intercept)   334  -7.2325792 12.070837 -30.89098405
#> 9  ran_modes Subject (Intercept)   335  -0.3336958 12.070837 -23.99210066
#> 10 ran_modes Subject (Intercept)   337  34.8903508 12.070837  11.23194592
#> 11 ran_modes Subject (Intercept)   349 -25.2101104 12.070837 -48.86851523
#> 12 ran_modes Subject (Intercept)   350 -13.0699567 12.070837 -36.72836156
#> 13 ran_modes Subject (Intercept)   351   4.5778352 12.070837 -19.08056968
#> 14 ran_modes Subject (Intercept)   352  20.8635924 12.070837  -2.79481241
#> 15 ran_modes Subject (Intercept)   369   3.2754530 12.070837 -20.38295184
#> 16 ran_modes Subject (Intercept)   370 -25.6128694 12.070837 -49.27127428
#> 17 ran_modes Subject (Intercept)   371   0.8070397 12.070837 -22.85136511
#> 18 ran_modes Subject (Intercept)   372  12.3145393 12.070837 -11.34386550
#> 19 ran_modes Subject        Days   308   9.1989719  2.304837   4.68157445
#> 20 ran_modes Subject        Days   309  -8.6197032  2.304837 -13.13710067
#> 21 ran_modes Subject        Days   310  -5.4488799  2.304837  -9.96627739
#> 22 ran_modes Subject        Days   330  -4.8143313  2.304837  -9.33172878
#> 23 ran_modes Subject        Days   331  -3.0698946  2.304837  -7.58729207
#> 24 ran_modes Subject        Days   332  -0.2721707  2.304837  -4.78956813
#> 25 ran_modes Subject        Days   333  -0.2236244  2.304837  -4.74102191
#> 26 ran_modes Subject        Days   334   1.0745761  2.304837  -3.44282140
#> 27 ran_modes Subject        Days   335 -10.7521591  2.304837 -15.26955659
#> 28 ran_modes Subject        Days   337   8.6282840  2.304837   4.11088649
#> 29 ran_modes Subject        Days   349   1.1734142  2.304837  -3.34398323
#> 30 ran_modes Subject        Days   350   6.6142050  2.304837   2.09680752
#> 31 ran_modes Subject        Days   351  -3.0152572  2.304837  -7.53265466
#> 32 ran_modes Subject        Days   352   3.5360133  2.304837  -0.98138416
#> 33 ran_modes Subject        Days   369   0.8722166  2.304837  -3.64518087
#> 34 ran_modes Subject        Days   370   4.8224646  2.304837   0.30506714
#> 35 ran_modes Subject        Days   371  -0.9881551  2.304837  -5.50555257
#> 36 ran_modes Subject        Days   372   1.2840297  2.304837  -3.23336773
#>      conf.high
#> 1   25.9169702
#> 2  -16.7401720
#> 3  -15.3018409
#> 4   47.3489033
#> 5   45.9186076
#> 6   32.6979308
#> 7   40.4988360
#> 8   16.4258257
#> 9   23.3247090
#> 10  58.5487556
#> 11  -1.5517055
#> 12  10.5884481
#> 13  28.2362400
#> 14  44.5219973
#> 15  26.9338579
#> 16  -1.9544646
#> 17  24.4654446
#> 18  35.9729442
#> 19  13.7163694
#> 20  -4.1023057
#> 21  -0.9314824
#> 22  -0.2969338
#> 23   1.4475029
#> 24   4.2452268
#> 25   4.2937730
#> 26   5.5919736
#> 27  -6.2347616
#> 28  13.1456814
#> 29   5.6908117
#> 30  11.1316025
#> 31   1.5021403
#> 32   8.0534108
#> 33   5.3896141
#> 34   9.3398621
#> 35   3.5292424
#> 36   5.8014272
head(augment(lmm1, sleepstudy))
#>   Reaction Days Subject  .fitted     .resid   .fixed      .mu .offset
#> 1 249.5600    0     308 253.6637  -4.103670 251.4051 253.6637       0
#> 2 258.7047    1     308 273.3299 -14.625228 261.8724 273.3299       0
#> 3 250.8006    2     308 292.9962 -42.195586 272.3397 292.9962       0
#> 4 321.4398    3     308 312.6624   8.777356 282.8070 312.6624       0
#> 5 356.8519    4     308 332.3287  24.523198 293.2742 332.3287       0
#> 6 414.6901    5     308 351.9950  62.695140 303.7415 351.9950       0
#>   .sqrtXwt .sqrtrwt .weights     .wtres
#> 1        1        1        1  -4.103670
#> 2        1        1        1 -14.625228
#> 3        1        1        1 -42.195586
#> 4        1        1        1   8.777356
#> 5        1        1        1  24.523198
#> 6        1        1        1  62.695140
glance(lmm1)
#>      sigma    logLik      AIC      BIC deviance df.residual
#> 1 25.59182 -871.8141 1755.628 1774.786 1743.628         174

glmm1 <- glmer(cbind(incidence, size - incidence) ~ period + (1 | herd),
               data = cbpp, family = binomial)
tidy(glmm1)
#>     effect group           term   estimate std.error statistic
#> 1    fixed fixed    (Intercept) -1.3983429 0.2312140 -6.047828
#> 2    fixed fixed        period2 -0.9919250 0.3031506 -3.272053
#> 3    fixed fixed        period3 -1.1282162 0.3228301 -3.494767
#> 4    fixed fixed        period4 -1.5797454 0.4220492 -3.743037
#> 5 ran_pars  herd sd_(Intercept)  0.6420699        NA        NA
#>        p.value
#> 1 1.468113e-09
#> 2 1.067695e-03
#> 3 4.744754e-04
#> 4 1.818098e-04
#> 5           NA
tidy(glmm1, effects = "fixed")
#>   effect        term  estimate std.error statistic      p.value
#> 1  fixed (Intercept) -1.398343 0.2312140 -6.047828 1.468113e-09
#> 2  fixed     period2 -0.991925 0.3031506 -3.272053 1.067695e-03
#> 3  fixed     period3 -1.128216 0.3228301 -3.494767 4.744754e-04
#> 4  fixed     period4 -1.579745 0.4220492 -3.743037 1.818098e-04
head(augment(glmm1, cbpp))
#>   herd incidence size period    .fitted     .resid    .fixed        .mu
#> 1    1         2   14      1 -0.8087134 -1.4377078 -1.398343 0.30816472
#> 2    1         3   12      2 -1.8006384  0.9884720 -2.390268 0.14177337
#> 3    1         4    9      3 -1.9369296  2.3566883 -2.526559 0.12598556
#> 4    1         0    5      4 -2.3884588 -0.9370227 -2.978088 0.08405701
#> 5    2         3   22      1 -1.6974362 -0.2431732 -1.398343 0.15480041
#> 6    2         1   18      2 -2.6893612 -0.1428294 -2.390268 0.06360406
#>   .offset  .sqrtXwt  .sqrtrwt .weights     .wtres       .eta
#> 1       0 1.7276542  8.103473       14 -1.3395656 -0.8087134
#> 2       0 1.2083394  9.930984       12  1.0747970 -1.8006384
#> 3       0 0.9954992  9.040690        9  2.8790881 -1.9369296
#> 4       0 0.6204492  8.058678        5 -0.6773884 -2.3884588
#> 5       0 1.6965905 12.967183       22 -0.2390730 -1.6974362
#> 6       0 1.0354006 17.384575       18 -0.1399198 -2.6893612
glance(glmm1)
#>   sigma    logLik      AIC      BIC deviance df.residual
#> 1     1 -92.02657 194.0531 204.1799 73.47428          51

The next part gives me some warnings that I though were errors at first but otherwise appears fine:

startvec <- c(Asym = 200, xmid = 725, scal = 350)
nm1 <- nlmer(circumference ~ SSlogis(age, Asym, xmid, scal) ~ Asym|Tree,
             Orange, start = startvec)
tidy(nm1)
#> Warning in vcov.merMod(object, use.hessian = use.hessian): variance-covariance matrix computed from finite-difference Hessian is
#> not positive definite or contains NA values: falling back to var-cov estimated from RX
#> Warning in vcov.merMod(object, correlation = correlation, sigm = sig): variance-covariance matrix computed from finite-difference Hessian is
#> not positive definite or contains NA values: falling back to var-cov estimated from RX
#>     effect    group           term   estimate std.error statistic
#> 1    fixed    fixed           Asym 192.052792  15.58385  12.32383
#> 2    fixed    fixed           xmid 727.904473  34.43847  21.13637
#> 3    fixed    fixed           scal 348.072134  26.30812  13.23060
#> 4 ran_pars     Tree        sd_Asym  31.646324        NA        NA
#> 5 ran_pars Residual sd_Observation   7.843009        NA        NA
tidy(nm1, effects = "fixed")
#> Warning in vcov.merMod(object, use.hessian = use.hessian): variance-covariance matrix computed from finite-difference Hessian is
#> not positive definite or contains NA values: falling back to var-cov estimated from RX

#> Warning in vcov.merMod(object, use.hessian = use.hessian): variance-covariance matrix computed from finite-difference Hessian is
#> not positive definite or contains NA values: falling back to var-cov estimated from RX
#>   effect term estimate std.error statistic
#> 1  fixed Asym 192.0528  15.58385  12.32383
#> 2  fixed xmid 727.9045  34.43847  21.13637
#> 3  fixed scal 348.0721  26.30812  13.23060
head(augment(nm1, Orange))
#> Warning in indices[which(stats::complete.cases(original))] =
#> seq_len(nrow(x)): number of items to replace is not a multiple of
#> replacement length
#>   Tree  age circumference   .fitted     .resid       .mu .offset  .sqrtXwt
#> 1    1  118            30  24.01051   5.989486  24.01051       0 0.1477654
#> 2    1  484            58  53.89014   4.109864  53.89014       0 0.3316505
#> 3    1  664            87  73.80812  13.191884  73.80812       0 0.4542296
#> 4    1 1004           115 111.87829   3.121713 111.87829       0 0.6885209
#> 5    1 1231           120 131.50149 -11.501489 131.50149       0 0.8092860
#> 6    1 1372           142 140.42155   1.578450 140.42155       0 0.8641818
#>   .sqrtrwt .weights     .wtres     .gam
#> 1        1        1   5.989486 162.4908
#> 2        1        1   4.109864 162.4908
#> 3        1        1  13.191884 162.4908
#> 4        1        1   3.121713 162.4908
#> 5        1        1 -11.501489 162.4908
#> 6        1        1   1.578450 162.4908
glance(nm1)
#>      sigma    logLik      AIC      BIC deviance df.residual
#> 1 7.843009 -131.5719 273.1438 280.9205 263.1438          30

Closing as all seems to be in order.

[github-actions]

This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.