Last updated on 2025-02-17 07:50:37 CET.
| Flavor | Version | Tinstall | Tcheck | Ttotal | Status | Flags |
|---|---|---|---|---|---|---|
| r-devel-linux-x86_64-debian-clang | 3.1-3 | 12.70 | 216.80 | 229.50 | OK | |
| r-devel-linux-x86_64-debian-gcc | 3.1-3 | 9.14 | 143.83 | 152.97 | OK | |
| r-devel-linux-x86_64-fedora-clang | 3.1-3 | 350.10 | ERROR | |||
| r-devel-linux-x86_64-fedora-gcc | 3.1-3 | 334.06 | ERROR | |||
| r-devel-macos-arm64 | 3.1-3 | 92.00 | OK | |||
| r-devel-macos-x86_64 | 3.1-3 | 203.00 | OK | |||
| r-devel-windows-x86_64 | 3.1-3 | 15.00 | 203.00 | 218.00 | OK | |
| r-patched-linux-x86_64 | 3.1-3 | 13.24 | 201.97 | 215.21 | OK | |
| r-release-linux-x86_64 | 3.1-3 | 12.16 | 200.76 | 212.92 | OK | |
| r-release-macos-arm64 | 3.1-3 | 96.00 | OK | |||
| r-release-macos-x86_64 | 3.1-3 | 139.00 | OK | |||
| r-release-windows-x86_64 | 3.1-3 | 14.00 | 205.00 | 219.00 | OK | |
| r-oldrel-macos-arm64 | 3.1-3 | 108.00 | OK | |||
| r-oldrel-macos-x86_64 | 3.1-3 | 183.00 | NOTE | |||
| r-oldrel-windows-x86_64 | 3.1-3 | 19.00 | 261.00 | 280.00 | OK |
Version: 3.1-3
Check: tests
Result: ERROR
Running ‘test_a_utils.R’ [6s/18s]
Running ‘test_anova.R’ [8s/22s]
Running ‘test_compare_sas.R’ [6s/15s]
Running ‘test_contest1D.R’ [9s/24s]
Running ‘test_contestMD.R’ [9s/19s]
Running ‘test_contrast_utils.R’ [7s/17s]
Running ‘test_drop1.R’ [9s/22s]
Running ‘test_legacy.R’ [8s/20s]
Running ‘test_lmer.R’ [10s/24s]
Running ‘test_lmerTest_paper.R’ [36s/111s]
Running ‘test_ls_means.R’ [7s/22s]
Running ‘test_ranova_step.R’ [9s/25s]
Running ‘test_summary.R’ [10s/28s]
Running ‘test_zerovar.R’ [6s/17s]
Running ‘zlmerTest_zeroDenom.R’ [6s/12s]
Running the tests in ‘tests/test_anova.R’ failed.
Complete output:
> # test_anova.R
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> # WRE says "using if(requireNamespace("pkgname")) is preferred, if possible."
> # even in tests:
> assertError <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertError(expr, ...) else invisible()
> assertWarning <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertWarning(expr, ...) else invisible()
>
> # Kenward-Roger only available with pbkrtest and only then validated in R >= 3.3.3
> # (faulty results for R < 3.3.3 may be due to unstated dependencies in pbkrtest)
> has_pbkrtest <- requireNamespace("pbkrtest", quietly = TRUE) && getRversion() >= "3.3.3"
>
> data("sleepstudy", package="lme4")
> TOL <- 1e-4
>
> ####################################
> ## Basic anova tests
> ####################################
>
> m <- lmer(Reaction ~ Days + (Days | Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
4: Model failed to converge with 1 negative eigenvalue: -2.6e+01
>
> ####### ddf argument:
> (an1 <- anova(m)) # Also testing print method.
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2 <- anova(m, ddf="Satterthwaite"))
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2b <- anova(m, ddf="Satterthwaite", type=3))
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2c <- anova(m, ddf="Satterthwaite", type=2))
Type II Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> (an3 <- anova(m, ddf="Sat")) ## Abbreviated argument
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an3, tolerance=TOL)
+ ))
> if(has_pbkrtest) {
+ (anova(m, ddf="Kenward-Roger"))
+ (anova(m, ddf="Kenward-Roger", type=3))
+ }
Type III Analysis of Variance Table with Kenward-Roger's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 17 3.284 0.08766 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an1 <- anova(m, ddf="lme4"))
Analysis of Variance Table
npar Sum Sq Mean Sq F value
Days 1 1948.5 1948.5 3.284
> (an2 <- anova(m, ddf="lme4", type=3)) # 'type' is ignored with ddf="lme4"
Analysis of Variance Table
npar Sum Sq Mean Sq F value
Days 1 1948.5 1948.5 3.284
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> res <- assertError(anova(m, ddf="KR")) ## Error on incorrect arg.
> stopifnot(
+ grepl("'arg' should be one of ", unlist(res[[1]])$message)
+ )
>
> ## lme4 method:
> an1 <- anova(m, ddf="lme4")
> an2 <- anova(as(m, "lmerMod"))
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
>
> ###### type argument:
> (an1 <- anova(m, type="1")) # valid type arg.
Type I Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2 <- anova(m, type="I")) # same
Type I Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> (an3 <- anova(m, type=1)) # Not strictly valid, but accepted
Type I Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an3, tolerance=TOL)
+ ))
>
> (an1 <- anova(m, type="2")) # valid type arg.
Type II Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2 <- anova(m, type="II")) # same
Type II Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> (an3 <- anova(m, type=3)) # Not strictly valid, but accepted
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an3, check.attributes=FALSE, tolerance=TOL)
+ ))
>
> (an1 <- anova(m, type="3")) # valid type arg.
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2 <- anova(m, type="III")) # same
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> (an3 <- anova(m, type=3)) # Not strictly valid, but accepted
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an3, tolerance=TOL)
+ ))
> assertError(anova(m, type=0)) # Not valid arg.
> assertError(anova(m, type="i")) # Not valid arg.
>
> ####### Model comparison:
> fm <- lm(Reaction ~ Days, sleepstudy)
> (an <- anova(m, fm))
refitting model(s) with ML (instead of REML)
Data: sleepstudy
Models:
fm: Reaction ~ Days
m: Reaction ~ Days + (Days | Subject)
npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
fm 3 1906.3 1915.9 -950.15 1900.3
m 6 1763.9 1783.1 -875.97 1751.9 148.35 3 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(
+ nrow(an) == 2L,
+ rownames(an)[2] == "m"
+ )
>
> m2 <- lmer(Reaction ~ Days + I(Days^2) + (Days | Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
4: Model failed to converge with 1 negative eigenvalue: -2.6e+01
> (an <- anova(m, m2, refit=FALSE))
Data: sleepstudy
Models:
m: Reaction ~ Days + (Days | Subject)
m2: Reaction ~ Days + I(Days^2) + (Days | Subject)
npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
m 6 Inf Inf -Inf Inf
m2 7 Inf Inf -Inf Inf NaN 1 NaN
> stopifnot(
+ nrow(an) == 2L,
+ rownames(an)[1] == "m"
+ )
>
>
> ####################################
> ## Example with factor fixef:
> ####################################
>
> ## 'temp' is continuous, 'temperature' an ordered factor with 6 levels
> data("cake", package="lme4")
> m <- lmer(angle ~ recipe * temp + (1|recipe:replicate), cake)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 0.719345 (tol = 0.002, component 1)
> (an <- anova(m))
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
recipe 4.04 2.02 2 248.07 0.0947 0.9097
temp 1966.71 1966.71 1 219.41 92.2233 <2e-16 ***
recipe:temp 1.74 0.87 2 219.41 0.0408 0.9600
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an_lme4 <- anova(m, ddf="lme4"))
Analysis of Variance Table
npar Sum Sq Mean Sq F value
recipe 2 11.77 5.88 0.2759
temp 1 1966.71 1966.71 92.2233
recipe:temp 2 1.74 0.87 0.0408
>
> if(has_pbkrtest) {
+ (an_KR <- anova(m, ddf="Kenward-Roger"))
+ # res <- all.equal(an[, c("Sum Sq", "Mean Sq", "F value")],
+ # an_lme4[, c("Sum Sq", "Mean Sq", "F value")])
+ # stopifnot(isTRUE(res))
+ res <- all.equal(an[, c("Sum Sq", "Mean Sq", "F value")],
+ an_KR[, c("Sum Sq", "Mean Sq", "F value")], tolerance=TOL)
+ stopifnot(isTRUE(res))
+ }
> stopifnot(all.equal(c(2, 1, 2), an$NumDF, tol=1e-6),
+ all.equal(c(254.0157612, 222, 222), an$DenDF, tol=TOL))
Error: c(254.0157612, 222, 222) and an$DenDF are not equal:
Mean relative difference: 0.01595027
Execution halted
Running the tests in ‘tests/test_compare_sas.R’ failed.
Complete output:
> # test_compare_sas.R
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> # WRE says "using if(requireNamespace("pkgname")) is preferred, if possible."
> # even in tests:
> assertError <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertError(expr, ...) else invisible()
> assertWarning <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertWarning(expr, ...) else invisible()
>
> #####################################################################
>
>
> # Use contrasts to get particular estimates for the summary table:
> l <- list(Frequency="contr.SAS", Income="contr.SAS")
> m.carrots <- lmer(Preference ~ sens2*Frequency*Income
+ +(1+sens2|Consumer), data=carrots, contrasts=l)
fixed-effect model matrix is rank deficient so dropping 12 columns / coefficients
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 2 negative eigenvalues
4: Model failed to converge with 2 negative eigenvalues: -3.7e+01 -1.6e+02
> an.m <- anova(m.carrots)
Missing cells for: Frequency3:Income1, Frequency5:Income1, Frequency4:Income2, Frequency3:Income4, Frequency4:Income4, Frequency5:Income4, sens2:Frequency3:Income1, sens2:Frequency5:Income1, sens2:Frequency4:Income2, sens2:Frequency3:Income4, sens2:Frequency4:Income4, sens2:Frequency5:Income4.
Interpret type III hypotheses with care.
>
> TOL <- 1e-4
> TOL2 <- 1e-5
> # with 4 decimals should agree with SAS output
> # numbers before decimals should agree with SAS output
> stopifnot(
+ all.equal(an.m[,"Pr(>F)"],
+ c(2e-5, 0.15512, 0.06939, 0.08223, 0.52459, 0.03119, 0.48344),
+ tolerance = TOL),
+ all.equal(round(an.m$DenDF), c(83, 83, 83, 83, 83, 83, 83))
+ )
Error: an.m[, "Pr(>F)"] and c(2e-05, 0.15512, 0.06939, 0.08223, 0.52459, 0.03119, 0.48344) are not equal:
Mean relative difference: 0.7703328
Execution halted
Running the tests in ‘tests/test_legacy.R’ failed.
Complete output:
> # test_legacy.R
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
> TOL <- 1e-4
> #####################################################################
>
> # Read in data set
> load(system.file("testdata", "legacy_fits.RData", package="lmerTest"))
> # Generated with the following code using lmerTest version 2.0-37.9002
> #
> # library("lmerTest")
> # packageVersion("lmerTest")
> # fm1 <- lmer(Reaction ~ Days + (Days|Subject), sleepstudy)
> # (an1 <- anova(fm1))
> # (sfm1 <- summary(fm1))
> #
> # fm2 <- lmer(Informed.liking ~ Product + Information + Gender +
> # (1|Product:Consumer) , data=ham)
> # (an2 <- anova(fm2))
> # (sfm2 <- summary(fm2))
> #
> # save(fm1, an1, sfm1, fm2, an2, sfm2,
> # file="~/GitHub/lmerTestR/package/inst/testdata/legacy_fits.RData")
>
>
> #######################################
> ### Check that arguments for merModLmerTest and lmerModLmerTest methods match up:
>
> stopifnot(
+ isTRUE(all.equal(formals(lmerTest:::anova.merModLmerTest),
+ formals(lmerTest:::anova.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::summary.merModLmerTest),
+ formals(lmerTest:::summary.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::drop1.merModLmerTest),
+ formals(lmerTest:::drop1.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::step.merModLmerTest),
+ formals(lmerTest:::step.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::ls_means.merModLmerTest),
+ formals(lmerTest:::ls_means.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::difflsmeans.merModLmerTest),
+ formals(lmerTest:::difflsmeans.lmerModLmerTest))))
>
>
> #######################################
> ## Tests for fm1:
>
> an1new <- anova(fm1)
> sfm1new <- summary(fm1)
>
> stopifnot(
+ isTRUE(all.equal(an1new, an1, check.attributes=FALSE, tol=TOL)),
+ isTRUE(all.equal(coef(sfm1new), coef(sfm1), tol=TOL))
+ )
>
> contest(fm1, c(0, 1))
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
1 30031.01 30031.01 1 17 45.853 3.26379e-06
> contest(fm1, c(0, 1), joint=FALSE)
Estimate Std. Error df t value lower upper Pr(>|t|)
1 10.46729 1.545789 17 6.771485 7.205956 13.72862 3.26379e-06
> drop1(fm1)
Single term deletions using Satterthwaite's method:
Model:
Reaction ~ Days + (Days | Subject)
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 30031 30031 1 17 45.853 3.264e-06 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> ranova(fm1)
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -872 1756
Days in (Days | Subject) 4 -Inf Inf Inf 2 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> step(fm1)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -872 1756
Days in (Days | Subject) 0 4 -Inf Inf Inf 2 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Backward reduced fixed-effect table:
Degrees of freedom method: Satterthwaite
Eliminated Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 0 30031 30031 1 17 45.853 3.264e-06 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ Days + (Days | Subject)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
>
> fm1new <- lmer(Reaction ~ Days + (Days|Subject), sleepstudy,
+ control=lmerControl(optimizer="bobyqa"))
> stopifnot(
+ isTRUE(all.equal(drop1(fm1), drop1(fm1new), tol=TOL)),
+ isTRUE(all.equal(ranova(fm1), ranova(fm1new), tol=TOL)),
+ isTRUE(all.equal(contest(fm1, c(0, 1)), contest(fm1new, c(0, 1)), tol=TOL)),
+ isTRUE(all.equal(contest(fm1, c(0, 1), joint=FALSE),
+ contest(fm1new, c(0, 1), joint=FALSE), tol=TOL))
+ )
Error: isTRUE(all.equal(ranova(fm1), ranova(fm1new), tol = TOL)) is not TRUE
In addition: Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
Execution halted
Running the tests in ‘tests/test_ls_means.R’ failed.
Complete output:
> # test_lsmeans.R
>
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> TOL <- 1e-4
> # Kenward-Roger only available with pbkrtest and only then validated in R >= 3.3.3
> # (faulty results for R < 3.3.3 may be due to unstated dependencies in pbkrtest)
> has_pbkrtest <- requireNamespace("pbkrtest", quietly = TRUE) && getRversion() >= "3.3.3"
>
> ########### Basic model structures:
>
> # Factor * covariate:
> data("cake", package="lme4")
> model <- lmer(angle ~ recipe * temp + (1|recipe:replicate), cake)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 0.719345 (tol = 0.002, component 1)
> (lsm <- ls_means(model))
Least Squares Means table:
Estimate Std. Error df t value lower upper Pr(>|t|)
recipeA 33.1222 1.6493 48.6 20.082 29.8071 36.4374 < 2.2e-16 ***
recipeB 31.6444 1.6493 48.6 19.186 28.3293 34.9596 < 2.2e-16 ***
recipeC 31.6000 1.6493 48.6 19.159 28.2849 34.9151 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Confidence level: 95%
Degrees of freedom method: Satterthwaite
> stopifnot(
+ nrow(lsm) == 3L,
+ ncol(lsm) == 7L,
+ # Balanced, so LS-means equal raw means:
+ isTRUE(all.equal(c(with(cake, tapply(angle, recipe, mean))), lsm[, "Estimate"],
+ check.attributes=FALSE, tolerance=TOL))
+ )
>
> # Pairwise differences of LS-means:
> plsm <- ls_means(model, pairwise = TRUE)
> plsm2 <- difflsmeans(model)
> C <- as.matrix(lmerTest:::get_pairs(rownames(lsm)))
> stopifnot(
+ isTRUE(all.equal(plsm, plsm2, tolerance=TOL)),
+ isTRUE(all.equal(plsm[, "Estimate"], c(lsm[, "Estimate"] %*% C),
+ check.attributes=FALSE, tolerance=TOL))
+ )
>
> # Contrasts vectors:
> show_tests(lsm)
$recipe
(Intercept) recipeB recipeC temp recipeB:temp recipeC:temp
recipeA 1 0 0 200 0 0
recipeB 1 1 0 200 200 0
recipeC 1 0 1 200 0 200
> show_tests(plsm)
$recipe
(Intercept) recipeB recipeC temp recipeB:temp recipeC:temp
recipeA - recipeB 0 -1 0 0 -200 0
recipeA - recipeC 0 0 -1 0 0 -200
recipeB - recipeC 0 1 -1 0 200 -200
>
> # Factor * Ordered:
> model <- lmer(angle ~ recipe * temperature + (1|recipe:replicate), cake)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 0.828782 (tol = 0.002, component 1)
> (lsm2 <- ls_means(model))
Least Squares Means table:
Estimate Std. Error df t value lower upper
recipeA 33.1222 1.6367 49.7 20.237 29.8342 36.4102
recipeB 31.6444 1.6367 49.7 19.334 28.3564 34.9325
recipeC 31.6000 1.6367 49.7 19.307 28.3120 34.8880
temperature175 27.9778 1.1322 96.9 24.711 25.7307 30.2249
temperature185 29.9556 1.1322 96.9 26.458 27.7085 32.2027
temperature195 31.4222 1.1322 96.9 27.753 29.1751 33.6693
temperature205 32.1778 1.1322 96.9 28.421 29.9307 34.4249
temperature215 35.8444 1.1322 96.9 31.659 33.5973 38.0915
temperature225 35.3556 1.1322 96.9 31.228 33.1085 37.6027
recipeA:temperature175 29.1333 1.9610 96.9 14.856 25.2412 33.0254
recipeB:temperature175 26.8667 1.9610 96.9 13.700 22.9746 30.7588
recipeC:temperature175 27.9333 1.9610 96.9 14.244 24.0412 31.8254
recipeA:temperature185 31.5333 1.9610 96.9 16.080 27.6412 35.4254
recipeB:temperature185 29.4000 1.9610 96.9 14.992 25.5079 33.2921
recipeC:temperature185 28.9333 1.9610 96.9 14.754 25.0412 32.8254
recipeA:temperature195 30.8000 1.9610 96.9 15.706 26.9079 34.6921
recipeB:temperature195 31.7333 1.9610 96.9 16.182 27.8412 35.6254
recipeC:temperature195 31.7333 1.9610 96.9 16.182 27.8412 35.6254
recipeA:temperature205 33.5333 1.9610 96.9 17.100 29.6412 37.4254
recipeB:temperature205 32.1333 1.9610 96.9 16.386 28.2412 36.0254
recipeC:temperature205 30.8667 1.9610 96.9 15.740 26.9746 34.7588
recipeA:temperature215 38.6667 1.9610 96.9 19.718 34.7746 42.5588
recipeB:temperature215 34.4667 1.9610 96.9 17.576 30.5746 38.3588
recipeC:temperature215 34.4000 1.9610 96.9 17.542 30.5079 38.2921
recipeA:temperature225 35.0667 1.9610 96.9 17.882 31.1746 38.9588
recipeB:temperature225 35.2667 1.9610 96.9 17.984 31.3746 39.1588
recipeC:temperature225 35.7333 1.9610 96.9 18.222 31.8412 39.6254
Pr(>|t|)
recipeA < 2.2e-16 ***
recipeB < 2.2e-16 ***
recipeC < 2.2e-16 ***
temperature175 < 2.2e-16 ***
temperature185 < 2.2e-16 ***
temperature195 < 2.2e-16 ***
temperature205 < 2.2e-16 ***
temperature215 < 2.2e-16 ***
temperature225 < 2.2e-16 ***
recipeA:temperature175 < 2.2e-16 ***
recipeB:temperature175 < 2.2e-16 ***
recipeC:temperature175 < 2.2e-16 ***
recipeA:temperature185 < 2.2e-16 ***
recipeB:temperature185 < 2.2e-16 ***
recipeC:temperature185 < 2.2e-16 ***
recipeA:temperature195 < 2.2e-16 ***
recipeB:temperature195 < 2.2e-16 ***
recipeC:temperature195 < 2.2e-16 ***
recipeA:temperature205 < 2.2e-16 ***
recipeB:temperature205 < 2.2e-16 ***
recipeC:temperature205 < 2.2e-16 ***
recipeA:temperature215 < 2.2e-16 ***
recipeB:temperature215 < 2.2e-16 ***
recipeC:temperature215 < 2.2e-16 ***
recipeA:temperature225 < 2.2e-16 ***
recipeB:temperature225 < 2.2e-16 ***
recipeC:temperature225 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Confidence level: 95%
Degrees of freedom method: Satterthwaite
> stopifnot(
+ nrow(lsm2) == 3 + 6 + 3*6,
+ ncol(lsm) == 7L,
+ # Balanced, so LS-means equal raw means:
+ isTRUE(all.equal(lsm[1:3, ], lsm2[1:3, ],
+ check.attributes=FALSE, tolerance=TOL))
+ )
Error: isTRUE(all.equal(lsm[1:3, ], lsm2[1:3, ], check.attributes = FALSE, .... is not TRUE
Execution halted
Running the tests in ‘tests/test_ranova_step.R’ failed.
Complete output:
> # test_ranova.R
>
> # Test functionality _before_ attaching lmerTest
> stopifnot(!"lmerTest" %in% .packages()) # ensure that lmerTest is NOT attached
> data("sleepstudy", package="lme4")
> f <- function(form, data) lmerTest::lmer(form, data=data)
> form <- "Reaction ~ Days + (Days|Subject)"
> fm <- f(form, data=sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
4: Model failed to converge with 1 negative eigenvalue: -2.6e+01
> lmerTest::ranova(fm)
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 4 -Inf Inf NaN 2 NaN
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> lmerTest::rand(fm)
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 4 -Inf Inf NaN 2 NaN
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> lmerTest::step(fm)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 0 4 -Inf Inf NaN 2 NaN
Backward reduced fixed-effect table:
Degrees of freedom method: Satterthwaite
Eliminated Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ (Days | Subject)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
3: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
4: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
5: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
6: Model failed to converge with 1 negative eigenvalue: -9.2e+00
>
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> # WRE says "using if(requireNamespace("pkgname")) is preferred, if possible."
> # even in tests:
> assertError <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertError(expr, ...) else invisible()
> assertWarning <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertWarning(expr, ...) else invisible()
>
> TOL <- 1e-4
> #####################################################################
> data("sleepstudy", package="lme4")
>
> # Test reduction of (Days | Subject) to (1 | Subject):
> fm1 <- lmer(Reaction ~ Days + (Days|Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
4: Model failed to converge with 1 negative eigenvalue: -2.6e+01
> (an <- rand(fm1)) # 2 df test
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 4 -Inf Inf NaN 2 NaN
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> (an <- ranova(fm1)) # 2 df test
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 4 -Inf Inf NaN 2 NaN
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> step(fm1)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 0 4 -Inf Inf NaN 2 NaN
Backward reduced fixed-effect table:
Degrees of freedom method: Satterthwaite
Eliminated Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ (Days | Subject)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
3: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
4: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
5: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
6: Model failed to converge with 1 negative eigenvalue: -9.2e+00
> stopifnot(
+ nrow(an) == 2L,
+ an[2L, "Df"] == 2L
+ )
>
> # This test can also be achieved with anova():
> fm2 <- lmer(Reaction ~ Days + (1|Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> (stp <- step(fm2))
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 4 -Inf Inf
(1 | Subject) 1 3 -947 1900 -Inf 1 1
Backward reduced fixed-effect table:
Eliminated Df Sum of Sq RSS AIC F value Pr(>F)
Days 0 1 162703 567954 1452.2 71.464 9.894e-15 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ Days
> get_model(stp)
Call:
lm(formula = Reaction ~ Days, data = sleepstudy)
Coefficients:
(Intercept) Days
251.41 10.47
> (ana <- anova(fm1, fm2, refit=FALSE))
Data: sleepstudy
Models:
fm2: Reaction ~ Days + (1 | Subject)
fm1: Reaction ~ Days + (Days | Subject)
npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
fm2 4 Inf Inf -Inf Inf
fm1 6 Inf Inf -Inf Inf NaN 2 NaN
>
> stopifnot(
+ all.equal(an[2L, "LRT"], ana[2L, "Chisq"], tolerance=TOL)
+ )
>
> # Illustrate complete.test argument:
> # Test removal of (Days | Subject):
> (an <- ranova(fm1, reduce.terms = FALSE)) # 3 df test
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
(Days | Subject) 3 -947 1900 -Inf 3 1
>
> # The likelihood ratio test statistic is in this case:
> fm3 <- lm(Reaction ~ Days, sleepstudy)
> LRT <- 2*c(logLik(fm1, REML=TRUE) - logLik(fm3, REML=TRUE)) # LRT
> stopifnot(
+ nrow(an) == 2L,
+ an[2L, "Df"] == 3L,
+ all.equal(an[2L, "LRT"], LRT, tolerance=TOL)
+ )
>
> ## _NULL_ model:
> fm <- lmer(Reaction ~ -1 + (1|Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 14.9723 (tol = 0.002, component 1)
> step(fm)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 2 -Inf Inf
(1 | Subject) 1 1 -1284 2571 -Inf 1 1
Backward reduced fixed-effect table:
Eliminated Df Sum of Sq RSS AIC F value Pr(>F)
Model found:
Reaction ~ 1 - 1
> ranova(fm)
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ (1 | Subject) - 1
npar logLik AIC LRT Df Pr(>Chisq)
<none> 2 -Inf Inf
(1 | Subject) 1 -1284 2571 -Inf 1 1
> lm1 <- lm(Reaction ~ 0, data=sleepstudy)
> LRT <- 2*c(logLik(fm, REML=FALSE) - logLik(lm1, REML=FALSE))
>
> ## Tests of ML-fits agree with anova():
> fm1 <- lmer(Reaction ~ Days + (1|Subject), sleepstudy, REML=FALSE)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.84934 (tol = 0.002, component 1)
> step(fm1)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 4 -Inf Inf
(1 | Subject) 1 3 -950 1906 -Inf 1 1
Backward reduced fixed-effect table:
Eliminated Df Sum of Sq RSS AIC F value Pr(>F)
Days 0 1 162703 567954 1452.2 71.464 9.894e-15 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ Days
> lm2 <- lm(Reaction ~ Days, sleepstudy)
> (an1 <- ranova(fm1))
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (1 | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 4 -Inf Inf
(1 | Subject) 3 -950 1906 -Inf 1 1
> (an2 <- anova(fm1, lm2))
Data: sleepstudy
Models:
lm2: Reaction ~ Days
fm1: Reaction ~ Days + (1 | Subject)
npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
lm2 3 1906 1916 -950 1900
fm1 4 Inf Inf -Inf Inf 0 1 1
> j <- grep("Chi Df|Df", colnames(an2))
> stopifnot(
+ all.equal(an1[2, "LRT"], an2[2, "Chisq"], tolerance=TOL),
+ all.equal(an1[2, "Df"], an2[2, j[length(j)]], tolerance=TOL),
+ all.equal(an1[1:2, "logLik"], an2[2:1, "logLik"], tolerance=TOL)
+ )
Error: an1[2, "LRT"] and an2[2, "Chisq"] are not equal:
Mean absolute difference: Inf
Execution halted
Running the tests in ‘tests/zlmerTest_zeroDenom.R’ failed.
Complete output:
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> # Read in data set
> load(system.file("testdata","potdata.RData", package="lmerTest"))
>
> # Mixed model
> lmerout <- lmer(biomass ~ CO2*nutrients + (1|chamber),data=potdata)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 0.788311 (tol = 0.002, component 1)
> summary(lmerout)
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: biomass ~ CO2 * nutrients + (1 | chamber)
Data: potdata
REML criterion at convergence: Inf
Scaled residuals:
Min 1Q Median 3Q Max
-1.4961 -0.5349 0.0000 0.5349 1.4961
Random effects:
Groups Name Variance Std.Dev.
chamber (Intercept) 0.07571 0.2752
Residual 1.71728 1.3105
Number of obs: 24, groups: chamber, 4
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 12.8500 0.9468 11.8625 13.572 1.39e-08 ***
CO2675 2.0000 1.3390 11.8625 1.494 0.161390
nutrients2 7.3500 1.3105 7.3492 5.609 0.000682 ***
nutrients3 10.4500 1.3105 7.3492 7.974 7.15e-05 ***
nutrients4 18.6500 1.3105 7.3492 14.232 1.29e-06 ***
nutrients5 25.0500 1.3105 7.3492 19.116 1.55e-07 ***
nutrients6 29.0000 1.3105 7.3492 22.130 5.39e-08 ***
CO2675:nutrients2 -0.5000 1.8533 7.3492 -0.270 0.794737
CO2675:nutrients3 1.9000 1.8533 7.3492 1.025 0.337816
CO2675:nutrients4 3.4500 1.8533 7.3492 1.862 0.102962
CO2675:nutrients5 5.9000 1.8533 7.3492 3.184 0.014448 *
CO2675:nutrients6 5.2500 1.8533 7.3492 2.833 0.024051 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) CO2675 ntrnt2 ntrnt3 ntrnt4 ntrnt5 ntrnt6 CO2675:2 CO2675:3
CO2675 -0.707
nutrients2 -0.692 0.489
nutrients3 -0.692 0.489 0.500
nutrients4 -0.692 0.489 0.500 0.500
nutrients5 -0.692 0.489 0.500 0.500 0.500
nutrients6 -0.692 0.489 0.500 0.500 0.500 0.500
CO2675:ntr2 0.489 -0.692 -0.707 -0.354 -0.354 -0.354 -0.354
CO2675:ntr3 0.489 -0.692 -0.354 -0.707 -0.354 -0.354 -0.354 0.500
CO2675:ntr4 0.489 -0.692 -0.354 -0.354 -0.707 -0.354 -0.354 0.500 0.500
CO2675:ntr5 0.489 -0.692 -0.354 -0.354 -0.354 -0.707 -0.354 0.500 0.500
CO2675:ntr6 0.489 -0.692 -0.354 -0.354 -0.354 -0.354 -0.707 0.500 0.500
CO2675:4 CO2675:5
CO2675
nutrients2
nutrients3
nutrients4
nutrients5
nutrients6
CO2675:ntr2
CO2675:ntr3
CO2675:ntr4
CO2675:ntr5 0.500
CO2675:ntr6 0.500 0.500
optimizer (nloptwrap) convergence code: -2 (NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).)
Model failed to converge with max|grad| = 0.788311 (tol = 0.002, component 1)
>
> an.sat <- anova(lmerout)
> anova(lmerout, ddf="lme4")
Analysis of Variance Table
npar Sum Sq Mean Sq F value
CO2 1 103.33 103.33 60.1718
nutrients 5 3050.44 610.09 355.2632
CO2:nutrients 5 35.47 7.09 4.1307
> TOL <- 1e-5
> stopifnot(isTRUE(all.equal(
+ an.sat[,"DenDF"], c(2, 10, 10), tolerance=TOL
+ )))
Error: isTRUE(all.equal(an.sat[, "DenDF"], c(2, 10, 10), tolerance = TOL)) is not TRUE
Execution halted
Flavor: r-devel-linux-x86_64-fedora-clang
Version: 3.1-3
Check: tests
Result: ERROR
Running ‘test_a_utils.R’
Running ‘test_anova.R’ [8s/21s]
Running ‘test_compare_sas.R’ [6s/16s]
Running ‘test_contest1D.R’ [8s/20s]
Running ‘test_contestMD.R’ [9s/24s]
Running ‘test_contrast_utils.R’ [6s/16s]
Running ‘test_drop1.R’ [8s/21s]
Running ‘test_legacy.R’ [7s/20s]
Running ‘test_lmer.R’ [9s/26s]
Running ‘test_lmerTest_paper.R’ [35s/96s]
Running ‘test_ls_means.R’ [7s/18s]
Running ‘test_ranova_step.R’ [8s/23s]
Running ‘test_summary.R’ [8s/24s]
Running ‘test_zerovar.R’ [6s/17s]
Running ‘zlmerTest_zeroDenom.R’ [6s/17s]
Running the tests in ‘tests/test_anova.R’ failed.
Complete output:
> # test_anova.R
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> # WRE says "using if(requireNamespace("pkgname")) is preferred, if possible."
> # even in tests:
> assertError <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertError(expr, ...) else invisible()
> assertWarning <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertWarning(expr, ...) else invisible()
>
> # Kenward-Roger only available with pbkrtest and only then validated in R >= 3.3.3
> # (faulty results for R < 3.3.3 may be due to unstated dependencies in pbkrtest)
> has_pbkrtest <- requireNamespace("pbkrtest", quietly = TRUE) && getRversion() >= "3.3.3"
>
> data("sleepstudy", package="lme4")
> TOL <- 1e-4
>
> ####################################
> ## Basic anova tests
> ####################################
>
> m <- lmer(Reaction ~ Days + (Days | Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
4: Model failed to converge with 1 negative eigenvalue: -2.6e+01
>
> ####### ddf argument:
> (an1 <- anova(m)) # Also testing print method.
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2 <- anova(m, ddf="Satterthwaite"))
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2b <- anova(m, ddf="Satterthwaite", type=3))
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2c <- anova(m, ddf="Satterthwaite", type=2))
Type II Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> (an3 <- anova(m, ddf="Sat")) ## Abbreviated argument
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an3, tolerance=TOL)
+ ))
> if(has_pbkrtest) {
+ (anova(m, ddf="Kenward-Roger"))
+ (anova(m, ddf="Kenward-Roger", type=3))
+ }
Type III Analysis of Variance Table with Kenward-Roger's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 17 3.284 0.08766 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an1 <- anova(m, ddf="lme4"))
Analysis of Variance Table
npar Sum Sq Mean Sq F value
Days 1 1948.5 1948.5 3.284
> (an2 <- anova(m, ddf="lme4", type=3)) # 'type' is ignored with ddf="lme4"
Analysis of Variance Table
npar Sum Sq Mean Sq F value
Days 1 1948.5 1948.5 3.284
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> res <- assertError(anova(m, ddf="KR")) ## Error on incorrect arg.
> stopifnot(
+ grepl("'arg' should be one of ", unlist(res[[1]])$message)
+ )
>
> ## lme4 method:
> an1 <- anova(m, ddf="lme4")
> an2 <- anova(as(m, "lmerMod"))
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
>
> ###### type argument:
> (an1 <- anova(m, type="1")) # valid type arg.
Type I Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2 <- anova(m, type="I")) # same
Type I Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> (an3 <- anova(m, type=1)) # Not strictly valid, but accepted
Type I Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an3, tolerance=TOL)
+ ))
>
> (an1 <- anova(m, type="2")) # valid type arg.
Type II Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2 <- anova(m, type="II")) # same
Type II Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> (an3 <- anova(m, type=3)) # Not strictly valid, but accepted
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an3, check.attributes=FALSE, tolerance=TOL)
+ ))
>
> (an1 <- anova(m, type="3")) # valid type arg.
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an2 <- anova(m, type="III")) # same
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an2, tolerance=TOL)
+ ))
> (an3 <- anova(m, type=3)) # Not strictly valid, but accepted
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(isTRUE(
+ all.equal(an1, an3, tolerance=TOL)
+ ))
> assertError(anova(m, type=0)) # Not valid arg.
> assertError(anova(m, type="i")) # Not valid arg.
>
> ####### Model comparison:
> fm <- lm(Reaction ~ Days, sleepstudy)
> (an <- anova(m, fm))
refitting model(s) with ML (instead of REML)
Data: sleepstudy
Models:
fm: Reaction ~ Days
m: Reaction ~ Days + (Days | Subject)
npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
fm 3 1906.3 1915.9 -950.15 1900.3
m 6 1763.9 1783.1 -875.97 1751.9 148.35 3 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> stopifnot(
+ nrow(an) == 2L,
+ rownames(an)[2] == "m"
+ )
>
> m2 <- lmer(Reaction ~ Days + I(Days^2) + (Days | Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
4: Model failed to converge with 1 negative eigenvalue: -2.6e+01
> (an <- anova(m, m2, refit=FALSE))
Data: sleepstudy
Models:
m: Reaction ~ Days + (Days | Subject)
m2: Reaction ~ Days + I(Days^2) + (Days | Subject)
npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
m 6 Inf Inf -Inf Inf
m2 7 Inf Inf -Inf Inf NaN 1 NaN
> stopifnot(
+ nrow(an) == 2L,
+ rownames(an)[1] == "m"
+ )
>
>
> ####################################
> ## Example with factor fixef:
> ####################################
>
> ## 'temp' is continuous, 'temperature' an ordered factor with 6 levels
> data("cake", package="lme4")
> m <- lmer(angle ~ recipe * temp + (1|recipe:replicate), cake)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 0.719345 (tol = 0.002, component 1)
> (an <- anova(m))
Type III Analysis of Variance Table with Satterthwaite's method
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
recipe 4.04 2.02 2 248.07 0.0947 0.9097
temp 1966.71 1966.71 1 219.41 92.2233 <2e-16 ***
recipe:temp 1.74 0.87 2 219.41 0.0408 0.9600
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> (an_lme4 <- anova(m, ddf="lme4"))
Analysis of Variance Table
npar Sum Sq Mean Sq F value
recipe 2 11.77 5.88 0.2759
temp 1 1966.71 1966.71 92.2233
recipe:temp 2 1.74 0.87 0.0408
>
> if(has_pbkrtest) {
+ (an_KR <- anova(m, ddf="Kenward-Roger"))
+ # res <- all.equal(an[, c("Sum Sq", "Mean Sq", "F value")],
+ # an_lme4[, c("Sum Sq", "Mean Sq", "F value")])
+ # stopifnot(isTRUE(res))
+ res <- all.equal(an[, c("Sum Sq", "Mean Sq", "F value")],
+ an_KR[, c("Sum Sq", "Mean Sq", "F value")], tolerance=TOL)
+ stopifnot(isTRUE(res))
+ }
> stopifnot(all.equal(c(2, 1, 2), an$NumDF, tol=1e-6),
+ all.equal(c(254.0157612, 222, 222), an$DenDF, tol=TOL))
Error: c(254.0157612, 222, 222) and an$DenDF are not equal:
Mean relative difference: 0.01595027
Execution halted
Running the tests in ‘tests/test_compare_sas.R’ failed.
Complete output:
> # test_compare_sas.R
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> # WRE says "using if(requireNamespace("pkgname")) is preferred, if possible."
> # even in tests:
> assertError <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertError(expr, ...) else invisible()
> assertWarning <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertWarning(expr, ...) else invisible()
>
> #####################################################################
>
>
> # Use contrasts to get particular estimates for the summary table:
> l <- list(Frequency="contr.SAS", Income="contr.SAS")
> m.carrots <- lmer(Preference ~ sens2*Frequency*Income
+ +(1+sens2|Consumer), data=carrots, contrasts=l)
fixed-effect model matrix is rank deficient so dropping 12 columns / coefficients
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 2 negative eigenvalues
4: Model failed to converge with 2 negative eigenvalues: -3.7e+01 -1.6e+02
> an.m <- anova(m.carrots)
Missing cells for: Frequency3:Income1, Frequency5:Income1, Frequency4:Income2, Frequency3:Income4, Frequency4:Income4, Frequency5:Income4, sens2:Frequency3:Income1, sens2:Frequency5:Income1, sens2:Frequency4:Income2, sens2:Frequency3:Income4, sens2:Frequency4:Income4, sens2:Frequency5:Income4.
Interpret type III hypotheses with care.
>
> TOL <- 1e-4
> TOL2 <- 1e-5
> # with 4 decimals should agree with SAS output
> # numbers before decimals should agree with SAS output
> stopifnot(
+ all.equal(an.m[,"Pr(>F)"],
+ c(2e-5, 0.15512, 0.06939, 0.08223, 0.52459, 0.03119, 0.48344),
+ tolerance = TOL),
+ all.equal(round(an.m$DenDF), c(83, 83, 83, 83, 83, 83, 83))
+ )
Error: an.m[, "Pr(>F)"] and c(2e-05, 0.15512, 0.06939, 0.08223, 0.52459, 0.03119, 0.48344) are not equal:
Mean relative difference: 0.7703328
Execution halted
Running the tests in ‘tests/test_legacy.R’ failed.
Complete output:
> # test_legacy.R
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
> TOL <- 1e-4
> #####################################################################
>
> # Read in data set
> load(system.file("testdata", "legacy_fits.RData", package="lmerTest"))
> # Generated with the following code using lmerTest version 2.0-37.9002
> #
> # library("lmerTest")
> # packageVersion("lmerTest")
> # fm1 <- lmer(Reaction ~ Days + (Days|Subject), sleepstudy)
> # (an1 <- anova(fm1))
> # (sfm1 <- summary(fm1))
> #
> # fm2 <- lmer(Informed.liking ~ Product + Information + Gender +
> # (1|Product:Consumer) , data=ham)
> # (an2 <- anova(fm2))
> # (sfm2 <- summary(fm2))
> #
> # save(fm1, an1, sfm1, fm2, an2, sfm2,
> # file="~/GitHub/lmerTestR/package/inst/testdata/legacy_fits.RData")
>
>
> #######################################
> ### Check that arguments for merModLmerTest and lmerModLmerTest methods match up:
>
> stopifnot(
+ isTRUE(all.equal(formals(lmerTest:::anova.merModLmerTest),
+ formals(lmerTest:::anova.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::summary.merModLmerTest),
+ formals(lmerTest:::summary.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::drop1.merModLmerTest),
+ formals(lmerTest:::drop1.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::step.merModLmerTest),
+ formals(lmerTest:::step.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::ls_means.merModLmerTest),
+ formals(lmerTest:::ls_means.lmerModLmerTest))),
+ isTRUE(all.equal(formals(lmerTest:::difflsmeans.merModLmerTest),
+ formals(lmerTest:::difflsmeans.lmerModLmerTest))))
>
>
> #######################################
> ## Tests for fm1:
>
> an1new <- anova(fm1)
> sfm1new <- summary(fm1)
>
> stopifnot(
+ isTRUE(all.equal(an1new, an1, check.attributes=FALSE, tol=TOL)),
+ isTRUE(all.equal(coef(sfm1new), coef(sfm1), tol=TOL))
+ )
>
> contest(fm1, c(0, 1))
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
1 30031.01 30031.01 1 17 45.853 3.26379e-06
> contest(fm1, c(0, 1), joint=FALSE)
Estimate Std. Error df t value lower upper Pr(>|t|)
1 10.46729 1.545789 17 6.771485 7.205956 13.72862 3.26379e-06
> drop1(fm1)
Single term deletions using Satterthwaite's method:
Model:
Reaction ~ Days + (Days | Subject)
Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 30031 30031 1 17 45.853 3.264e-06 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> ranova(fm1)
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -872 1756
Days in (Days | Subject) 4 -Inf Inf Inf 2 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> step(fm1)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -872 1756
Days in (Days | Subject) 0 4 -Inf Inf Inf 2 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Backward reduced fixed-effect table:
Degrees of freedom method: Satterthwaite
Eliminated Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 0 30031 30031 1 17 45.853 3.264e-06 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ Days + (Days | Subject)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
>
> fm1new <- lmer(Reaction ~ Days + (Days|Subject), sleepstudy,
+ control=lmerControl(optimizer="bobyqa"))
> stopifnot(
+ isTRUE(all.equal(drop1(fm1), drop1(fm1new), tol=TOL)),
+ isTRUE(all.equal(ranova(fm1), ranova(fm1new), tol=TOL)),
+ isTRUE(all.equal(contest(fm1, c(0, 1)), contest(fm1new, c(0, 1)), tol=TOL)),
+ isTRUE(all.equal(contest(fm1, c(0, 1), joint=FALSE),
+ contest(fm1new, c(0, 1), joint=FALSE), tol=TOL))
+ )
Error: isTRUE(all.equal(ranova(fm1), ranova(fm1new), tol = TOL)) is not TRUE
In addition: Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
Execution halted
Running the tests in ‘tests/test_ls_means.R’ failed.
Complete output:
> # test_lsmeans.R
>
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> TOL <- 1e-4
> # Kenward-Roger only available with pbkrtest and only then validated in R >= 3.3.3
> # (faulty results for R < 3.3.3 may be due to unstated dependencies in pbkrtest)
> has_pbkrtest <- requireNamespace("pbkrtest", quietly = TRUE) && getRversion() >= "3.3.3"
>
> ########### Basic model structures:
>
> # Factor * covariate:
> data("cake", package="lme4")
> model <- lmer(angle ~ recipe * temp + (1|recipe:replicate), cake)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 0.719345 (tol = 0.002, component 1)
> (lsm <- ls_means(model))
Least Squares Means table:
Estimate Std. Error df t value lower upper Pr(>|t|)
recipeA 33.1222 1.6493 48.6 20.082 29.8071 36.4374 < 2.2e-16 ***
recipeB 31.6444 1.6493 48.6 19.186 28.3293 34.9596 < 2.2e-16 ***
recipeC 31.6000 1.6493 48.6 19.159 28.2849 34.9151 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Confidence level: 95%
Degrees of freedom method: Satterthwaite
> stopifnot(
+ nrow(lsm) == 3L,
+ ncol(lsm) == 7L,
+ # Balanced, so LS-means equal raw means:
+ isTRUE(all.equal(c(with(cake, tapply(angle, recipe, mean))), lsm[, "Estimate"],
+ check.attributes=FALSE, tolerance=TOL))
+ )
>
> # Pairwise differences of LS-means:
> plsm <- ls_means(model, pairwise = TRUE)
> plsm2 <- difflsmeans(model)
> C <- as.matrix(lmerTest:::get_pairs(rownames(lsm)))
> stopifnot(
+ isTRUE(all.equal(plsm, plsm2, tolerance=TOL)),
+ isTRUE(all.equal(plsm[, "Estimate"], c(lsm[, "Estimate"] %*% C),
+ check.attributes=FALSE, tolerance=TOL))
+ )
>
> # Contrasts vectors:
> show_tests(lsm)
$recipe
(Intercept) recipeB recipeC temp recipeB:temp recipeC:temp
recipeA 1 0 0 200 0 0
recipeB 1 1 0 200 200 0
recipeC 1 0 1 200 0 200
> show_tests(plsm)
$recipe
(Intercept) recipeB recipeC temp recipeB:temp recipeC:temp
recipeA - recipeB 0 -1 0 0 -200 0
recipeA - recipeC 0 0 -1 0 0 -200
recipeB - recipeC 0 1 -1 0 200 -200
>
> # Factor * Ordered:
> model <- lmer(angle ~ recipe * temperature + (1|recipe:replicate), cake)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 0.828782 (tol = 0.002, component 1)
> (lsm2 <- ls_means(model))
Least Squares Means table:
Estimate Std. Error df t value lower upper
recipeA 33.1222 1.6367 49.7 20.237 29.8342 36.4102
recipeB 31.6444 1.6367 49.7 19.334 28.3564 34.9325
recipeC 31.6000 1.6367 49.7 19.307 28.3120 34.8880
temperature175 27.9778 1.1322 96.9 24.711 25.7307 30.2249
temperature185 29.9556 1.1322 96.9 26.458 27.7085 32.2027
temperature195 31.4222 1.1322 96.9 27.753 29.1751 33.6693
temperature205 32.1778 1.1322 96.9 28.421 29.9307 34.4249
temperature215 35.8444 1.1322 96.9 31.659 33.5973 38.0915
temperature225 35.3556 1.1322 96.9 31.228 33.1085 37.6027
recipeA:temperature175 29.1333 1.9610 96.9 14.856 25.2412 33.0254
recipeB:temperature175 26.8667 1.9610 96.9 13.700 22.9746 30.7588
recipeC:temperature175 27.9333 1.9610 96.9 14.244 24.0412 31.8254
recipeA:temperature185 31.5333 1.9610 96.9 16.080 27.6412 35.4254
recipeB:temperature185 29.4000 1.9610 96.9 14.992 25.5079 33.2921
recipeC:temperature185 28.9333 1.9610 96.9 14.754 25.0412 32.8254
recipeA:temperature195 30.8000 1.9610 96.9 15.706 26.9079 34.6921
recipeB:temperature195 31.7333 1.9610 96.9 16.182 27.8412 35.6254
recipeC:temperature195 31.7333 1.9610 96.9 16.182 27.8412 35.6254
recipeA:temperature205 33.5333 1.9610 96.9 17.100 29.6412 37.4254
recipeB:temperature205 32.1333 1.9610 96.9 16.386 28.2412 36.0254
recipeC:temperature205 30.8667 1.9610 96.9 15.740 26.9746 34.7588
recipeA:temperature215 38.6667 1.9610 96.9 19.718 34.7746 42.5588
recipeB:temperature215 34.4667 1.9610 96.9 17.576 30.5746 38.3588
recipeC:temperature215 34.4000 1.9610 96.9 17.542 30.5079 38.2921
recipeA:temperature225 35.0667 1.9610 96.9 17.882 31.1746 38.9588
recipeB:temperature225 35.2667 1.9610 96.9 17.984 31.3746 39.1588
recipeC:temperature225 35.7333 1.9610 96.9 18.222 31.8412 39.6254
Pr(>|t|)
recipeA < 2.2e-16 ***
recipeB < 2.2e-16 ***
recipeC < 2.2e-16 ***
temperature175 < 2.2e-16 ***
temperature185 < 2.2e-16 ***
temperature195 < 2.2e-16 ***
temperature205 < 2.2e-16 ***
temperature215 < 2.2e-16 ***
temperature225 < 2.2e-16 ***
recipeA:temperature175 < 2.2e-16 ***
recipeB:temperature175 < 2.2e-16 ***
recipeC:temperature175 < 2.2e-16 ***
recipeA:temperature185 < 2.2e-16 ***
recipeB:temperature185 < 2.2e-16 ***
recipeC:temperature185 < 2.2e-16 ***
recipeA:temperature195 < 2.2e-16 ***
recipeB:temperature195 < 2.2e-16 ***
recipeC:temperature195 < 2.2e-16 ***
recipeA:temperature205 < 2.2e-16 ***
recipeB:temperature205 < 2.2e-16 ***
recipeC:temperature205 < 2.2e-16 ***
recipeA:temperature215 < 2.2e-16 ***
recipeB:temperature215 < 2.2e-16 ***
recipeC:temperature215 < 2.2e-16 ***
recipeA:temperature225 < 2.2e-16 ***
recipeB:temperature225 < 2.2e-16 ***
recipeC:temperature225 < 2.2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Confidence level: 95%
Degrees of freedom method: Satterthwaite
> stopifnot(
+ nrow(lsm2) == 3 + 6 + 3*6,
+ ncol(lsm) == 7L,
+ # Balanced, so LS-means equal raw means:
+ isTRUE(all.equal(lsm[1:3, ], lsm2[1:3, ],
+ check.attributes=FALSE, tolerance=TOL))
+ )
Error: isTRUE(all.equal(lsm[1:3, ], lsm2[1:3, ], check.attributes = FALSE, .... is not TRUE
Execution halted
Running the tests in ‘tests/test_ranova_step.R’ failed.
Complete output:
> # test_ranova.R
>
> # Test functionality _before_ attaching lmerTest
> stopifnot(!"lmerTest" %in% .packages()) # ensure that lmerTest is NOT attached
> data("sleepstudy", package="lme4")
> f <- function(form, data) lmerTest::lmer(form, data=data)
> form <- "Reaction ~ Days + (Days|Subject)"
> fm <- f(form, data=sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
4: Model failed to converge with 1 negative eigenvalue: -2.6e+01
> lmerTest::ranova(fm)
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 4 -Inf Inf NaN 2 NaN
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> lmerTest::rand(fm)
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 4 -Inf Inf NaN 2 NaN
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> lmerTest::step(fm)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 0 4 -Inf Inf NaN 2 NaN
Backward reduced fixed-effect table:
Degrees of freedom method: Satterthwaite
Eliminated Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ (Days | Subject)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
3: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
4: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
5: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
6: Model failed to converge with 1 negative eigenvalue: -9.2e+00
>
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> # WRE says "using if(requireNamespace("pkgname")) is preferred, if possible."
> # even in tests:
> assertError <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertError(expr, ...) else invisible()
> assertWarning <- function(expr, ...)
+ if(requireNamespace("tools")) tools::assertWarning(expr, ...) else invisible()
>
> TOL <- 1e-4
> #####################################################################
> data("sleepstudy", package="lme4")
>
> # Test reduction of (Days | Subject) to (1 | Subject):
> fm1 <- lmer(Reaction ~ Days + (Days|Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
3: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
4: Model failed to converge with 1 negative eigenvalue: -2.6e+01
> (an <- rand(fm1)) # 2 df test
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 4 -Inf Inf NaN 2 NaN
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> (an <- ranova(fm1)) # 2 df test
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 4 -Inf Inf NaN 2 NaN
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> step(fm1)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
Days in (Days | Subject) 0 4 -Inf Inf NaN 2 NaN
Backward reduced fixed-effect table:
Degrees of freedom method: Satterthwaite
Eliminated Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
Days 1 1948.5 1948.5 1 115.69 3.284 0.07255 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ (Days | Subject)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
3: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
4: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
unable to evaluate scaled gradient
5: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge: degenerate Hessian with 1 negative eigenvalues
6: Model failed to converge with 1 negative eigenvalue: -9.2e+00
> stopifnot(
+ nrow(an) == 2L,
+ an[2L, "Df"] == 2L
+ )
>
> # This test can also be achieved with anova():
> fm2 <- lmer(Reaction ~ Days + (1|Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.97927 (tol = 0.002, component 1)
> (stp <- step(fm2))
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 4 -Inf Inf
(1 | Subject) 1 3 -947 1900 -Inf 1 1
Backward reduced fixed-effect table:
Eliminated Df Sum of Sq RSS AIC F value Pr(>F)
Days 0 1 162703 567954 1452.2 71.464 9.894e-15 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ Days
> get_model(stp)
Call:
lm(formula = Reaction ~ Days, data = sleepstudy)
Coefficients:
(Intercept) Days
251.41 10.47
> (ana <- anova(fm1, fm2, refit=FALSE))
Data: sleepstudy
Models:
fm2: Reaction ~ Days + (1 | Subject)
fm1: Reaction ~ Days + (Days | Subject)
npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
fm2 4 Inf Inf -Inf Inf
fm1 6 Inf Inf -Inf Inf NaN 2 NaN
>
> stopifnot(
+ all.equal(an[2L, "LRT"], ana[2L, "Chisq"], tolerance=TOL)
+ )
>
> # Illustrate complete.test argument:
> # Test removal of (Days | Subject):
> (an <- ranova(fm1, reduce.terms = FALSE)) # 3 df test
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (Days | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 6 -Inf Inf
(Days | Subject) 3 -947 1900 -Inf 3 1
>
> # The likelihood ratio test statistic is in this case:
> fm3 <- lm(Reaction ~ Days, sleepstudy)
> LRT <- 2*c(logLik(fm1, REML=TRUE) - logLik(fm3, REML=TRUE)) # LRT
> stopifnot(
+ nrow(an) == 2L,
+ an[2L, "Df"] == 3L,
+ all.equal(an[2L, "LRT"], LRT, tolerance=TOL)
+ )
>
> ## _NULL_ model:
> fm <- lmer(Reaction ~ -1 + (1|Subject), sleepstudy)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 14.9723 (tol = 0.002, component 1)
> step(fm)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 2 -Inf Inf
(1 | Subject) 1 1 -1284 2571 -Inf 1 1
Backward reduced fixed-effect table:
Eliminated Df Sum of Sq RSS AIC F value Pr(>F)
Model found:
Reaction ~ 1 - 1
> ranova(fm)
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ (1 | Subject) - 1
npar logLik AIC LRT Df Pr(>Chisq)
<none> 2 -Inf Inf
(1 | Subject) 1 -1284 2571 -Inf 1 1
> lm1 <- lm(Reaction ~ 0, data=sleepstudy)
> LRT <- 2*c(logLik(fm, REML=FALSE) - logLik(lm1, REML=FALSE))
>
> ## Tests of ML-fits agree with anova():
> fm1 <- lmer(Reaction ~ Days + (1|Subject), sleepstudy, REML=FALSE)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 1.84934 (tol = 0.002, component 1)
> step(fm1)
Backward reduced random-effect table:
Eliminated npar logLik AIC LRT Df Pr(>Chisq)
<none> 4 -Inf Inf
(1 | Subject) 1 3 -950 1906 -Inf 1 1
Backward reduced fixed-effect table:
Eliminated Df Sum of Sq RSS AIC F value Pr(>F)
Days 0 1 162703 567954 1452.2 71.464 9.894e-15 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model found:
Reaction ~ Days
> lm2 <- lm(Reaction ~ Days, sleepstudy)
> (an1 <- ranova(fm1))
ANOVA-like table for random-effects: Single term deletions
Model:
Reaction ~ Days + (1 | Subject)
npar logLik AIC LRT Df Pr(>Chisq)
<none> 4 -Inf Inf
(1 | Subject) 3 -950 1906 -Inf 1 1
> (an2 <- anova(fm1, lm2))
Data: sleepstudy
Models:
lm2: Reaction ~ Days
fm1: Reaction ~ Days + (1 | Subject)
npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
lm2 3 1906 1916 -950 1900
fm1 4 Inf Inf -Inf Inf 0 1 1
> j <- grep("Chi Df|Df", colnames(an2))
> stopifnot(
+ all.equal(an1[2, "LRT"], an2[2, "Chisq"], tolerance=TOL),
+ all.equal(an1[2, "Df"], an2[2, j[length(j)]], tolerance=TOL),
+ all.equal(an1[1:2, "logLik"], an2[2:1, "logLik"], tolerance=TOL)
+ )
Error: an1[2, "LRT"] and an2[2, "Chisq"] are not equal:
Mean absolute difference: Inf
Execution halted
Running the tests in ‘tests/zlmerTest_zeroDenom.R’ failed.
Complete output:
> library(lmerTest)
Loading required package: lme4
Loading required package: Matrix
Attaching package: 'lmerTest'
The following object is masked from 'package:lme4':
lmer
The following object is masked from 'package:stats':
step
>
> # Read in data set
> load(system.file("testdata","potdata.RData", package="lmerTest"))
>
> # Mixed model
> lmerout <- lmer(biomass ~ CO2*nutrients + (1|chamber),data=potdata)
Warning messages:
1: In optwrap(optimizer, devfun, getStart(start, rho$pp), lower = rho$lower, :
convergence code -2 from nloptwrap: NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 0.788311 (tol = 0.002, component 1)
> summary(lmerout)
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: biomass ~ CO2 * nutrients + (1 | chamber)
Data: potdata
REML criterion at convergence: Inf
Scaled residuals:
Min 1Q Median 3Q Max
-1.4961 -0.5349 0.0000 0.5349 1.4961
Random effects:
Groups Name Variance Std.Dev.
chamber (Intercept) 0.07571 0.2752
Residual 1.71728 1.3105
Number of obs: 24, groups: chamber, 4
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 12.8500 0.9468 11.8625 13.572 1.39e-08 ***
CO2675 2.0000 1.3390 11.8625 1.494 0.161390
nutrients2 7.3500 1.3105 7.3492 5.609 0.000682 ***
nutrients3 10.4500 1.3105 7.3492 7.974 7.15e-05 ***
nutrients4 18.6500 1.3105 7.3492 14.232 1.29e-06 ***
nutrients5 25.0500 1.3105 7.3492 19.116 1.55e-07 ***
nutrients6 29.0000 1.3105 7.3492 22.130 5.39e-08 ***
CO2675:nutrients2 -0.5000 1.8533 7.3492 -0.270 0.794737
CO2675:nutrients3 1.9000 1.8533 7.3492 1.025 0.337816
CO2675:nutrients4 3.4500 1.8533 7.3492 1.862 0.102962
CO2675:nutrients5 5.9000 1.8533 7.3492 3.184 0.014448 *
CO2675:nutrients6 5.2500 1.8533 7.3492 2.833 0.024051 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) CO2675 ntrnt2 ntrnt3 ntrnt4 ntrnt5 ntrnt6 CO2675:2 CO2675:3
CO2675 -0.707
nutrients2 -0.692 0.489
nutrients3 -0.692 0.489 0.500
nutrients4 -0.692 0.489 0.500 0.500
nutrients5 -0.692 0.489 0.500 0.500 0.500
nutrients6 -0.692 0.489 0.500 0.500 0.500 0.500
CO2675:ntr2 0.489 -0.692 -0.707 -0.354 -0.354 -0.354 -0.354
CO2675:ntr3 0.489 -0.692 -0.354 -0.707 -0.354 -0.354 -0.354 0.500
CO2675:ntr4 0.489 -0.692 -0.354 -0.354 -0.707 -0.354 -0.354 0.500 0.500
CO2675:ntr5 0.489 -0.692 -0.354 -0.354 -0.354 -0.707 -0.354 0.500 0.500
CO2675:ntr6 0.489 -0.692 -0.354 -0.354 -0.354 -0.354 -0.707 0.500 0.500
CO2675:4 CO2675:5
CO2675
nutrients2
nutrients3
nutrients4
nutrients5
nutrients6
CO2675:ntr2
CO2675:ntr3
CO2675:ntr4
CO2675:ntr5 0.500
CO2675:ntr6 0.500 0.500
optimizer (nloptwrap) convergence code: -2 (NLOPT_INVALID_ARGS: Invalid arguments (e.g. lower bounds are bigger than upper bounds, an unknown algorithm was specified, etcetera).)
Model failed to converge with max|grad| = 0.788311 (tol = 0.002, component 1)
>
> an.sat <- anova(lmerout)
> anova(lmerout, ddf="lme4")
Analysis of Variance Table
npar Sum Sq Mean Sq F value
CO2 1 103.33 103.33 60.1718
nutrients 5 3050.44 610.09 355.2632
CO2:nutrients 5 35.47 7.09 4.1307
> TOL <- 1e-5
> stopifnot(isTRUE(all.equal(
+ an.sat[,"DenDF"], c(2, 10, 10), tolerance=TOL
+ )))
Error: isTRUE(all.equal(an.sat[, "DenDF"], c(2, 10, 10), tolerance = TOL)) is not TRUE
Execution halted
Flavor: r-devel-linux-x86_64-fedora-gcc
Version: 3.1-3
Check: package dependencies
Result: NOTE
Package suggested but not available for checking: ‘pbkrtest’
Flavor: r-oldrel-macos-x86_64
Version: 3.1-3
Check: Rd cross-references
Result: NOTE
Package unavailable to check Rd xrefs: ‘pbkrtest’
Flavor: r-oldrel-macos-x86_64