Nothing
tests/testthat/test-02-mirtTwo.R
In mirt: Multidimensional Item Response Theory
expect_class <- function(x, class) expect_true(inherits(x, class))
test_that('poly', {
modp1 <- mirt(Science, 1, verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
expect_equal(extract.mirt(modp1, 'df'), 239)
cfs <- as.numeric(do.call(c, coef(modp1)))
expect_equal(cfs, c(1.041, 4.864, 2.64, -1.466, 1.226, 2.924, 0.901, -2.266, 2.296, 5.238, 2.216, -1.965, 1.095, 3.348, 0.992, -1.688, 0, 1),
tolerance = 1e-2)
C2 <- M2(modp1, type = 'C2')
expect_equal(C2$M2, 19.17929, tolerance=1e-4)
expect_equal(C2$p, 6.84337e-05, tolerance=1e-4)
modp2 <- mirt(Science, 1, 'sequential', verbose=FALSE)
expect_class(modp2, 'SingleGroupClass')
expect_equal(extract.mirt(modp2, 'df'), 239)
cfs <- as.numeric(do.call(c, coef(modp2)))
expect_equal(cfs, c(0.997414,4.831609,2.750703,-1.365844,1.067633,2.815793,1.104466,-1.87485,2.132679,5.021663,2.223683,-1.784284,0.9942698,3.287169,1.12849,-1.308458,0,1),
tolerance = 1e-2)
expect_equal(logLik(modp2), -1609.768, tolerance = 1e-4)
modLouis <- mirt(Science, 1, SE=T, SE.type='Louis', verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modLouis, printSE=TRUE)))
expect_equal(cfs, c(1.04236, 0.18838, 4.86544, 0.49088, 2.64044, 0.22267, -1.46621, 0.15868, 1.22569, 0.18189, 2.924, 0.23928, 0.90115, 0.14289, -2.26661, 0.20308, 2.29058, 0.48269, 5.22988, 0.72817, 2.21201, 0.3564, -1.96222, 0.32209, 1.09557, 0.18336, 3.34845, 0.27659, 0.9919, 0.14053, -1.68846, 0.16864, 0, NA, 1, NA),
tolerance = 1e-3)
expect_equal(modLouis@OptimInfo$condnum, 98.26492, tolerance = 1e-2)
modsandwich <- mirt(Science, 1, SE=T, SE.type='sandwich.Louis', verbose=FALSE)
expect_class(modsandwich, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modsandwich, printSE=TRUE)))
expect_equal(cfs, c(1.04236, 0.23843, 4.86544, 0.46794, 2.64044, 0.24659, -1.46621, 0.17162, 1.22569, 0.1922, 2.924, 0.24655, 0.90115, 0.14592, -2.26661, 0.19899, 2.29058, 0.52026, 5.22988, 0.80736, 2.21201, 0.373, -1.96222, 0.33683, 1.09557, 0.22701, 3.34845, 0.29203, 0.9919, 0.14491, -1.68846, 0.18015, 0, NA, 1, NA),
tolerance = 1e-3)
expect_equal(extract.mirt(modsandwich, 'condnum'), 141.5391, tolerance = 1e-2)
modsandwich <- mirt(Science, 1, SE=T, SE.type='sandwich', verbose=FALSE)
expect_class(modsandwich, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modsandwich, printSE=TRUE)))
expect_equal(cfs, c(1.04236, 0.23843, 4.86544, 0.46794, 2.64044, 0.24659, -1.46621, 0.17162, 1.22569, 0.1922, 2.924, 0.24655, 0.90115, 0.14592, -2.26661, 0.19899, 2.29058, 0.52026, 5.22988, 0.80736, 2.21201, 0.373, -1.96222, 0.33683, 1.09557, 0.22701, 3.34845, 0.29203, 0.9919, 0.14491, -1.68846, 0.18015, 0, NA, 1, NA),
tolerance = 1e-3)
expect_equal(extract.mirt(modsandwich, 'condnum'), 141.5346, tolerance = 1e-2)
modOakes <- mirt(Science, 1, SE=T, SE.type='Oakes', verbose=FALSE)
expect_equal(modOakes@OptimInfo$condnum, 97.8644, tolerance = 1e-4)
modp1 <- mirt(Science, 1, verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
expect_equal(extract.mirt(modp1, 'df'), 239)
cfs <- as.numeric(do.call(c, coef(modp1)))
expect_equal(cfs, c(1.041, 4.864, 2.64, -1.466, 1.226, 2.924, 0.901, -2.266, 2.296, 5.238, 2.216, -1.965, 1.095, 3.348, 0.992, -1.688, 0, 1),
tolerance = 1e-2)
vals <- mirt(Science, 1, large = 'return', verbose=FALSE)
modp1 <- mirt(Science, 1, large = vals, verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp1)))
expect_equal(cfs, c(1.041, 4.864, 2.64, -1.466, 1.226, 2.924, 0.901, -2.266, 2.296, 5.238, 2.216, -1.965, 1.095, 3.348, 0.992, -1.688, 0, 1),
tolerance = 1e-2)
modp1 <- mirt(Science, 1, SE=TRUE, SE.type = 'SEM', verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
expect_equal(extract.mirt(modp1, 'condnum'), 108.8042, tolerance = 1e-2)
cfs <- as.numeric(do.call(c, coef(modp1)))
expect_equal(cfs, c(1.041, 0.656, 1.425, 4.863, 3.849, 5.876, 2.639, 2.196, 3.083, -1.466, -1.782, -1.149, 1.226, 0.887, 1.565, 2.924, 2.45, 3.398, 0.901, 0.614, 1.188, -2.266, -2.639, -1.894, 2.3, 1.325, 3.275, 5.244, 3.804, 6.685, 2.218, 1.488, 2.949, -1.967, -2.605, -1.329, 1.094, 0.727, 1.461, 3.347, 2.801, 3.893, 0.991, 0.717, 1.266, -1.688, -2.018, -1.357, 0, NA, NA, 1, NA, NA),
tolerance = 1e-2)
modp2 <- mirt(Science, 2, verbose=FALSE)
expect_class(modp2, 'SingleGroupClass')
expect_equal(modp2@Fit$df, 236)
cfs <- as.numeric(do.call(c, coef(modp2, verbose=FALSE)))
expect_equal(abs(cfs), abs(c(-1.3278,0.1081,5.1934,2.8583,-1.5996,-0.8762,1.8783,3.7248,1.1598,-2.9225,-1.4614,1.1639,4.6495,1.951,-1.7322,-1.7397,0,4.0053,1.2008,-2.0548,0,0,1,0,1)),
tolerance = 1e-2)
modp3 <- mirt(Science, 1, constrain = list(c(1,5)), parprior = list(c(4,'norm',0,1)), verbose=FALSE)
expect_class(modp3, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp3, verbose = FALSE)))
expect_true(mirt:::closeEnough(cfs - c(1.134,4.964,2.711,-1.473,1.134,2.865,0.882,-2.205,2.211,5.136,2.171,-1.91,1.129,3.383,1.006,-1.7,0,1),
-1e-2, 1e-2))
newmodel <- mirt.model('F = 1-4
CONSTRAIN = (1-2,a1)
PRIOR = (1, d1, norm, 4, 1)')
modp3 <- mirt(Science, newmodel, verbose=FALSE)
expect_class(modp3, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp3, verbose = FALSE)))
expect_true(mirt:::closeEnough(cfs - c(1.132,4.795,2.677,-1.507,1.132,2.856,0.874,-2.211,2.216,5.127,2.16,-1.927,1.128,3.374,0.999,-1.707,0,1),
-1e-2, 1e-2))
modOakes <- mirt(Science, newmodel, SE=T, SE.type='Oakes', verbose=FALSE)
expect_equal(modOakes@OptimInfo$condnum, 106.0775, tolerance = 1e-4)
modp4 <- mirt(Science, 1, itemtype = c(rep('graded',3), 'nominal'), verbose=FALSE)
expect_class(modp4, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp4, verbose = FALSE)))
expect_equal(cfs, c(1.0408, 4.862, 2.6387, -1.4664, 1.2063, 2.9083, 0.8958, -2.254, 2.3376, 5.2972, 2.2404, -1.9886, 0.7986, 0, 1.0782, 1.7756, 3, 0, 2.1964, 2.9637, 1.6742, 0, 1),
tolerance = 1e-2)
modp5 <- mirt(Science, 1, itemtype = c(rep('graded',3), 'gpcm'), SE = TRUE, SE.type = 'SEM', verbose=FALSE)
expect_class(modp5, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp5, verbose = FALSE)))
expect_equal(cfs, c(1.057,0.659,1.454,4.876,3.908,5.844,2.65,2.206,3.093,-1.472,-1.799,-1.146,1.219,0.865,1.573,2.918,2.444,3.391,0.9,0.615,1.185,-2.263,-2.662,-1.864,2.254,1.244,3.265,5.177,3.606,6.747,2.19,1.395,2.985,-1.942,-2.587,-1.298,0.771,0.441,1.1,0,NA,NA,1,NA,NA,2,NA,NA,3,NA,NA,0,NA,NA,2.16,1.537,2.782,2.973,2.276,3.671,1.767,1.128,2.407,0,NA,NA,1,NA,NA),
tolerance = 1e-2)
modp6 <- mirt(Science, 1, dentype="empiricalhist", verbose = FALSE, TOL=1e-3)
expect_class(modp6, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp6, verbose = FALSE)))
expect_equal(cfs, c(0.856,5.072,2.639,-1.35,1.095,2.951,0.968,-2.181,2.601,5.541,2.634,-1.752,0.988,3.443,1.058,-1.595,0,1),
tolerance = 1e-2)
fm0 <- fscores(modp1, method='EAP', response.pattern = c(1,2,3,4), full.scores=FALSE)
expect_equal(as.numeric(fm0[,c('F1','SE_F1')]), c(-0.3494903, 0.6004922), tolerance=1e-4)
fm1 <- fscores(modp1, verbose = FALSE, full.scores=FALSE)
expect_class(fm1, 'matrix')
expect_true(mirt:::closeEnough(fm1[1:6,'F1'] - c(-2.7488324, -1.4190140, -0.7154329, -0.4453752, -2.5438490, -1.2481072), -1e-2, 1e-2))
fm1b <- fscores(modp1, verbose = FALSE, full.scores=TRUE)
expect_equal(cor(fm1b, rowSums(Science))[1], .969, tolerance = .02)
fm2 <- fscores(modp2, rotate = 'oblimin', verbose = FALSE, full.scores=FALSE)
expect_class(fm2, 'matrix')
expect_true(mirt:::closeEnough(abs(as.numeric(fm2[1:6,c('F1','F2')])) -
c(2.5966,1.8668,0.6578,1.1597,2.4204,0.7001,2.412,0.8689,0.0258,0.2609,2.3376,1.4844),
-1e-2, 1e-2))
fm3 <- fscores(modp3, rotate = 'oblimin', full.scores = TRUE, verbose = FALSE)
expect_class(fm3, 'matrix')
fm4 <- fscores(modp4, verbose = FALSE, full.scores=FALSE)
expect_class(fm4, 'matrix')
expect_true(mirt:::closeEnough(fm4[1:6,'F1'] - c(-2.7320802, -1.4459303, -0.7910009, -0.5438761, -2.5310045, -1.1434060), -1e-2, 1e-2))
fm5 <- fscores(modp5, verbose = FALSE, full.scores=FALSE)
expect_class(fm5, 'matrix')
expect_true(mirt:::closeEnough(fm5[1:6,'F1'] - c(-2.7249561, -1.4446593, -0.7364399, -0.5627047, -2.5174376, -1.1732461), -1e-2, 1e-2))
resmat <- residuals(modp3, type = 'Q3', Theta = fm3[,'F'], verbose = FALSE)
expect_equal(as.numeric(resmat), c(1,-0.209,-0.283,0.057,-0.209,1,-0.149,-0.235,-0.283,-0.149,1,-0.33,0.057,-0.235,-0.33,1), tolerance=1e-2)
resmatLD <- residuals(modp3, type = 'LD', verbose = FALSE)
expect_equal(as.numeric(resmatLD), c(NA,23.88453,12.78836,22.16431,-0.1425131,NA,10.44856,22.30602,-0.1042807,0.09425936,NA,17.52566,0.1372851,-0.1377233,-0.1220769,NA), tolerance=1e-2)
resmatG2 <- residuals(modp3, type = 'LDG2', verbose = FALSE)
expect_equal(as.numeric(resmatG2), c(NA,21.11214,10.52294,19.91545,-0.1339869,NA,10.25686,17.55354,-0.09459427,0.0933907,NA,18.39732,0.1301342,-0.122174,-0.1250759,NA), tolerance=1e-2)
cof1 <- coef(modp1)
expect_class(cof1, 'list')
cof2 <- coef(modp2, verbose = FALSE)
expect_class(cof2, 'list')
IP1 <- itemplot(modp1, 1)
IP2 <- itemplot(modp2, 1)
expect_class(IP1, 'trellis')
expect_class(IP2, 'trellis')
fit <- suppressMessages(itemfit(modp2, c('S_X2', 'Zh')))
expect_equal(fit$Zh, c(1.287093, 4.581860, 3.934030, 3.989205), tolerance=1e-4)
expect_equal(fit$S_X2, c(3.941804,10.58528,6.913239,10.11068), tolerance=1e-4)
fs <- fscores(modp1, method = 'WLE', verbose=FALSE, full.scores=FALSE)
expect_equal(as.numeric(fs[1:3, 5:6]), c(-5.7024116, -2.1162737, -1.1386969, 1.5797286,
0.6321478, 0.6544024), tolerance = 1e-4)
##rating scale test
if(FALSE){
rm(list=ls())
set.seed(1234)
a <- matrix(rep(1, 10))
d <- matrix(c(1,0.5,-.5,-1), 10, 4, byrow = TRUE)
cc <- seq(-1, 1, length.out=10)
data <- simdata(a, d + cc, 2000, itemtype = rep('graded',10))
sv <- mirt(data, 1, itemtype = 'grsm', pars = 'values', verbose=FALSE)
sv[,'value'] <- c(as.vector(t(cbind(a,d,cc))),0,1)
save(data, sv, file = 'tests/tests/testdata/rst.rds')
}
load('testdata/rst.rds')
grsm <- mirt(data, 1, itemtype = 'grsm', pars = sv, calcNull= FALSE, verbose=FALSE, TOL=1e-5)
rsm <- mirt(data, 1, itemtype = 'rsm', calcNull= FALSE, verbose=FALSE, TOL = 1e-3)
expect_class(grsm, 'SingleGroupClass')
expect_class(rsm, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(grsm, verbose = FALSE)))
expect_equal(cfs, c(0.958,1,0.491,-0.543,-1.034,-1,0.986,1,0.491,-0.543,-1.034,-0.765,0.993,1,0.491,-0.543,-1.034,-0.543,1.026,1,0.491,-0.543,-1.034,-0.272,0.994,1,0.491,-0.543,-1.034,-0.104,0.986,1,0.491,-0.543,-1.034,0.18,0.956,1,0.491,-0.543,-1.034,0.405,1.039,1,0.491,-0.543,-1.034,0.579,0.963,1,0.491,-0.543,-1.034,0.879,0.946,1,0.491,-0.543,-1.034,1.137,0,1),
tolerance = 1e-2)
cfs <- as.numeric(do.call(c, coef(rsm, verbose = FALSE)))
expect_equal(cfs, c(1,1.3931,-0.479,1.2474,-0.7059,0,1,1.3931,-0.479,1.2474,-0.7059,0.0779,1,1.3931,-0.479,1.2474,-0.7059,0.164,1,1.3931,-0.479,1.2474,-0.7059,0.2547,1,1.3931,-0.479,1.2474,-0.7059,0.321,1,1.3931,-0.479,1.2474,-0.7059,0.4254,1,1.3931,-0.479,1.2474,-0.7059,0.504,1,1.3931,-0.479,1.2474,-0.7059,0.5722,1,1.3931,-0.479,1.2474,-0.7059,0.6694,1,1.3931,-0.479,1.2474,-0.7059,0.7629,0,0.134),
tolerance = 1e-2)
expect_equal(extract.mirt(rsm, 'df'), 9765610)
expect_equal(extract.mirt(grsm, 'df'), 9765601)
graded <- mirt(data, 1, verbose = FALSE)
gM2 <- M2(graded, calcNull=TRUE)
expect_equal(gM2$M2, 16.68317, tolerance = 1e-4)
expect_equal(gM2$df, 5)
expect_equal(gM2$CFI, .9723554, tolerance = 1e-4)
expect_equal(gM2$SRMSR, 0.01984457, tolerance = 1e-4)
#item and test info
Theta <- matrix(seq(-4,4,.01))
x <- extract.item(modp1, 1)
iinfo <- iteminfo(x, Theta)
expect_class(iinfo, 'numeric')
iinfo <- iteminfo(x, Theta, total.info=FALSE)
expect_class(iinfo, 'matrix')
tinfo <- testinfo(modp1, Theta)
expect_class(tinfo, 'numeric')
ER <- fscores(modp2, returnER = TRUE)
expect_equal(as.numeric(ER), c(0.4882546, 0.5099054), tolerance=1e-4)
suppressWarnings(ER2 <- fscores(modp2, returnER = TRUE, mean = c(-1, 1), cov = matrix(c(1.5,1,1,2), 2)))
expect_equal(as.numeric(ER2), c(0.5433246, 0.5929488), tolerance=1e-4)
})
Try the mirt package in your browser
Any scripts or data that you put into this service are public.
mirt documentation built on April 3, 2025, 11:59 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
expect_class <- function(x, class) expect_true(inherits(x, class))
test_that('poly', {
modp1 <- mirt(Science, 1, verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
expect_equal(extract.mirt(modp1, 'df'), 239)
cfs <- as.numeric(do.call(c, coef(modp1)))
expect_equal(cfs, c(1.041, 4.864, 2.64, -1.466, 1.226, 2.924, 0.901, -2.266, 2.296, 5.238, 2.216, -1.965, 1.095, 3.348, 0.992, -1.688, 0, 1),
tolerance = 1e-2)
C2 <- M2(modp1, type = 'C2')
expect_equal(C2$M2, 19.17929, tolerance=1e-4)
expect_equal(C2$p, 6.84337e-05, tolerance=1e-4)
modp2 <- mirt(Science, 1, 'sequential', verbose=FALSE)
expect_class(modp2, 'SingleGroupClass')
expect_equal(extract.mirt(modp2, 'df'), 239)
cfs <- as.numeric(do.call(c, coef(modp2)))
expect_equal(cfs, c(0.997414,4.831609,2.750703,-1.365844,1.067633,2.815793,1.104466,-1.87485,2.132679,5.021663,2.223683,-1.784284,0.9942698,3.287169,1.12849,-1.308458,0,1),
tolerance = 1e-2)
expect_equal(logLik(modp2), -1609.768, tolerance = 1e-4)
modLouis <- mirt(Science, 1, SE=T, SE.type='Louis', verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modLouis, printSE=TRUE)))
expect_equal(cfs, c(1.04236, 0.18838, 4.86544, 0.49088, 2.64044, 0.22267, -1.46621, 0.15868, 1.22569, 0.18189, 2.924, 0.23928, 0.90115, 0.14289, -2.26661, 0.20308, 2.29058, 0.48269, 5.22988, 0.72817, 2.21201, 0.3564, -1.96222, 0.32209, 1.09557, 0.18336, 3.34845, 0.27659, 0.9919, 0.14053, -1.68846, 0.16864, 0, NA, 1, NA),
tolerance = 1e-3)
expect_equal(modLouis@OptimInfo$condnum, 98.26492, tolerance = 1e-2)
modsandwich <- mirt(Science, 1, SE=T, SE.type='sandwich.Louis', verbose=FALSE)
expect_class(modsandwich, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modsandwich, printSE=TRUE)))
expect_equal(cfs, c(1.04236, 0.23843, 4.86544, 0.46794, 2.64044, 0.24659, -1.46621, 0.17162, 1.22569, 0.1922, 2.924, 0.24655, 0.90115, 0.14592, -2.26661, 0.19899, 2.29058, 0.52026, 5.22988, 0.80736, 2.21201, 0.373, -1.96222, 0.33683, 1.09557, 0.22701, 3.34845, 0.29203, 0.9919, 0.14491, -1.68846, 0.18015, 0, NA, 1, NA),
tolerance = 1e-3)
expect_equal(extract.mirt(modsandwich, 'condnum'), 141.5391, tolerance = 1e-2)
modsandwich <- mirt(Science, 1, SE=T, SE.type='sandwich', verbose=FALSE)
expect_class(modsandwich, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modsandwich, printSE=TRUE)))
expect_equal(cfs, c(1.04236, 0.23843, 4.86544, 0.46794, 2.64044, 0.24659, -1.46621, 0.17162, 1.22569, 0.1922, 2.924, 0.24655, 0.90115, 0.14592, -2.26661, 0.19899, 2.29058, 0.52026, 5.22988, 0.80736, 2.21201, 0.373, -1.96222, 0.33683, 1.09557, 0.22701, 3.34845, 0.29203, 0.9919, 0.14491, -1.68846, 0.18015, 0, NA, 1, NA),
tolerance = 1e-3)
expect_equal(extract.mirt(modsandwich, 'condnum'), 141.5346, tolerance = 1e-2)
modOakes <- mirt(Science, 1, SE=T, SE.type='Oakes', verbose=FALSE)
expect_equal(modOakes@OptimInfo$condnum, 97.8644, tolerance = 1e-4)
modp1 <- mirt(Science, 1, verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
expect_equal(extract.mirt(modp1, 'df'), 239)
cfs <- as.numeric(do.call(c, coef(modp1)))
expect_equal(cfs, c(1.041, 4.864, 2.64, -1.466, 1.226, 2.924, 0.901, -2.266, 2.296, 5.238, 2.216, -1.965, 1.095, 3.348, 0.992, -1.688, 0, 1),
tolerance = 1e-2)
vals <- mirt(Science, 1, large = 'return', verbose=FALSE)
modp1 <- mirt(Science, 1, large = vals, verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp1)))
expect_equal(cfs, c(1.041, 4.864, 2.64, -1.466, 1.226, 2.924, 0.901, -2.266, 2.296, 5.238, 2.216, -1.965, 1.095, 3.348, 0.992, -1.688, 0, 1),
tolerance = 1e-2)
modp1 <- mirt(Science, 1, SE=TRUE, SE.type = 'SEM', verbose=FALSE)
expect_class(modp1, 'SingleGroupClass')
expect_equal(extract.mirt(modp1, 'condnum'), 108.8042, tolerance = 1e-2)
cfs <- as.numeric(do.call(c, coef(modp1)))
expect_equal(cfs, c(1.041, 0.656, 1.425, 4.863, 3.849, 5.876, 2.639, 2.196, 3.083, -1.466, -1.782, -1.149, 1.226, 0.887, 1.565, 2.924, 2.45, 3.398, 0.901, 0.614, 1.188, -2.266, -2.639, -1.894, 2.3, 1.325, 3.275, 5.244, 3.804, 6.685, 2.218, 1.488, 2.949, -1.967, -2.605, -1.329, 1.094, 0.727, 1.461, 3.347, 2.801, 3.893, 0.991, 0.717, 1.266, -1.688, -2.018, -1.357, 0, NA, NA, 1, NA, NA),
tolerance = 1e-2)
modp2 <- mirt(Science, 2, verbose=FALSE)
expect_class(modp2, 'SingleGroupClass')
expect_equal(modp2@Fit$df, 236)
cfs <- as.numeric(do.call(c, coef(modp2, verbose=FALSE)))
expect_equal(abs(cfs), abs(c(-1.3278,0.1081,5.1934,2.8583,-1.5996,-0.8762,1.8783,3.7248,1.1598,-2.9225,-1.4614,1.1639,4.6495,1.951,-1.7322,-1.7397,0,4.0053,1.2008,-2.0548,0,0,1,0,1)),
tolerance = 1e-2)
modp3 <- mirt(Science, 1, constrain = list(c(1,5)), parprior = list(c(4,'norm',0,1)), verbose=FALSE)
expect_class(modp3, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp3, verbose = FALSE)))
expect_true(mirt:::closeEnough(cfs - c(1.134,4.964,2.711,-1.473,1.134,2.865,0.882,-2.205,2.211,5.136,2.171,-1.91,1.129,3.383,1.006,-1.7,0,1),
-1e-2, 1e-2))
newmodel <- mirt.model('F = 1-4
CONSTRAIN = (1-2,a1)
PRIOR = (1, d1, norm, 4, 1)')
modp3 <- mirt(Science, newmodel, verbose=FALSE)
expect_class(modp3, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp3, verbose = FALSE)))
expect_true(mirt:::closeEnough(cfs - c(1.132,4.795,2.677,-1.507,1.132,2.856,0.874,-2.211,2.216,5.127,2.16,-1.927,1.128,3.374,0.999,-1.707,0,1),
-1e-2, 1e-2))
modOakes <- mirt(Science, newmodel, SE=T, SE.type='Oakes', verbose=FALSE)
expect_equal(modOakes@OptimInfo$condnum, 106.0775, tolerance = 1e-4)
modp4 <- mirt(Science, 1, itemtype = c(rep('graded',3), 'nominal'), verbose=FALSE)
expect_class(modp4, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp4, verbose = FALSE)))
expect_equal(cfs, c(1.0408, 4.862, 2.6387, -1.4664, 1.2063, 2.9083, 0.8958, -2.254, 2.3376, 5.2972, 2.2404, -1.9886, 0.7986, 0, 1.0782, 1.7756, 3, 0, 2.1964, 2.9637, 1.6742, 0, 1),
tolerance = 1e-2)
modp5 <- mirt(Science, 1, itemtype = c(rep('graded',3), 'gpcm'), SE = TRUE, SE.type = 'SEM', verbose=FALSE)
expect_class(modp5, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp5, verbose = FALSE)))
expect_equal(cfs, c(1.057,0.659,1.454,4.876,3.908,5.844,2.65,2.206,3.093,-1.472,-1.799,-1.146,1.219,0.865,1.573,2.918,2.444,3.391,0.9,0.615,1.185,-2.263,-2.662,-1.864,2.254,1.244,3.265,5.177,3.606,6.747,2.19,1.395,2.985,-1.942,-2.587,-1.298,0.771,0.441,1.1,0,NA,NA,1,NA,NA,2,NA,NA,3,NA,NA,0,NA,NA,2.16,1.537,2.782,2.973,2.276,3.671,1.767,1.128,2.407,0,NA,NA,1,NA,NA),
tolerance = 1e-2)
modp6 <- mirt(Science, 1, dentype="empiricalhist", verbose = FALSE, TOL=1e-3)
expect_class(modp6, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(modp6, verbose = FALSE)))
expect_equal(cfs, c(0.856,5.072,2.639,-1.35,1.095,2.951,0.968,-2.181,2.601,5.541,2.634,-1.752,0.988,3.443,1.058,-1.595,0,1),
tolerance = 1e-2)
fm0 <- fscores(modp1, method='EAP', response.pattern = c(1,2,3,4), full.scores=FALSE)
expect_equal(as.numeric(fm0[,c('F1','SE_F1')]), c(-0.3494903, 0.6004922), tolerance=1e-4)
fm1 <- fscores(modp1, verbose = FALSE, full.scores=FALSE)
expect_class(fm1, 'matrix')
expect_true(mirt:::closeEnough(fm1[1:6,'F1'] - c(-2.7488324, -1.4190140, -0.7154329, -0.4453752, -2.5438490, -1.2481072), -1e-2, 1e-2))
fm1b <- fscores(modp1, verbose = FALSE, full.scores=TRUE)
expect_equal(cor(fm1b, rowSums(Science))[1], .969, tolerance = .02)
fm2 <- fscores(modp2, rotate = 'oblimin', verbose = FALSE, full.scores=FALSE)
expect_class(fm2, 'matrix')
expect_true(mirt:::closeEnough(abs(as.numeric(fm2[1:6,c('F1','F2')])) -
c(2.5966,1.8668,0.6578,1.1597,2.4204,0.7001,2.412,0.8689,0.0258,0.2609,2.3376,1.4844),
-1e-2, 1e-2))
fm3 <- fscores(modp3, rotate = 'oblimin', full.scores = TRUE, verbose = FALSE)
expect_class(fm3, 'matrix')
fm4 <- fscores(modp4, verbose = FALSE, full.scores=FALSE)
expect_class(fm4, 'matrix')
expect_true(mirt:::closeEnough(fm4[1:6,'F1'] - c(-2.7320802, -1.4459303, -0.7910009, -0.5438761, -2.5310045, -1.1434060), -1e-2, 1e-2))
fm5 <- fscores(modp5, verbose = FALSE, full.scores=FALSE)
expect_class(fm5, 'matrix')
expect_true(mirt:::closeEnough(fm5[1:6,'F1'] - c(-2.7249561, -1.4446593, -0.7364399, -0.5627047, -2.5174376, -1.1732461), -1e-2, 1e-2))
resmat <- residuals(modp3, type = 'Q3', Theta = fm3[,'F'], verbose = FALSE)
expect_equal(as.numeric(resmat), c(1,-0.209,-0.283,0.057,-0.209,1,-0.149,-0.235,-0.283,-0.149,1,-0.33,0.057,-0.235,-0.33,1), tolerance=1e-2)
resmatLD <- residuals(modp3, type = 'LD', verbose = FALSE)
expect_equal(as.numeric(resmatLD), c(NA,23.88453,12.78836,22.16431,-0.1425131,NA,10.44856,22.30602,-0.1042807,0.09425936,NA,17.52566,0.1372851,-0.1377233,-0.1220769,NA), tolerance=1e-2)
resmatG2 <- residuals(modp3, type = 'LDG2', verbose = FALSE)
expect_equal(as.numeric(resmatG2), c(NA,21.11214,10.52294,19.91545,-0.1339869,NA,10.25686,17.55354,-0.09459427,0.0933907,NA,18.39732,0.1301342,-0.122174,-0.1250759,NA), tolerance=1e-2)
cof1 <- coef(modp1)
expect_class(cof1, 'list')
cof2 <- coef(modp2, verbose = FALSE)
expect_class(cof2, 'list')
IP1 <- itemplot(modp1, 1)
IP2 <- itemplot(modp2, 1)
expect_class(IP1, 'trellis')
expect_class(IP2, 'trellis')
fit <- suppressMessages(itemfit(modp2, c('S_X2', 'Zh')))
expect_equal(fit$Zh, c(1.287093, 4.581860, 3.934030, 3.989205), tolerance=1e-4)
expect_equal(fit$S_X2, c(3.941804,10.58528,6.913239,10.11068), tolerance=1e-4)
fs <- fscores(modp1, method = 'WLE', verbose=FALSE, full.scores=FALSE)
expect_equal(as.numeric(fs[1:3, 5:6]), c(-5.7024116, -2.1162737, -1.1386969, 1.5797286,
0.6321478, 0.6544024), tolerance = 1e-4)
##rating scale test
if(FALSE){
rm(list=ls())
set.seed(1234)
a <- matrix(rep(1, 10))
d <- matrix(c(1,0.5,-.5,-1), 10, 4, byrow = TRUE)
cc <- seq(-1, 1, length.out=10)
data <- simdata(a, d + cc, 2000, itemtype = rep('graded',10))
sv <- mirt(data, 1, itemtype = 'grsm', pars = 'values', verbose=FALSE)
sv[,'value'] <- c(as.vector(t(cbind(a,d,cc))),0,1)
save(data, sv, file = 'tests/tests/testdata/rst.rds')
}
load('testdata/rst.rds')
grsm <- mirt(data, 1, itemtype = 'grsm', pars = sv, calcNull= FALSE, verbose=FALSE, TOL=1e-5)
rsm <- mirt(data, 1, itemtype = 'rsm', calcNull= FALSE, verbose=FALSE, TOL = 1e-3)
expect_class(grsm, 'SingleGroupClass')
expect_class(rsm, 'SingleGroupClass')
cfs <- as.numeric(do.call(c, coef(grsm, verbose = FALSE)))
expect_equal(cfs, c(0.958,1,0.491,-0.543,-1.034,-1,0.986,1,0.491,-0.543,-1.034,-0.765,0.993,1,0.491,-0.543,-1.034,-0.543,1.026,1,0.491,-0.543,-1.034,-0.272,0.994,1,0.491,-0.543,-1.034,-0.104,0.986,1,0.491,-0.543,-1.034,0.18,0.956,1,0.491,-0.543,-1.034,0.405,1.039,1,0.491,-0.543,-1.034,0.579,0.963,1,0.491,-0.543,-1.034,0.879,0.946,1,0.491,-0.543,-1.034,1.137,0,1),
tolerance = 1e-2)
cfs <- as.numeric(do.call(c, coef(rsm, verbose = FALSE)))
expect_equal(cfs, c(1,1.3931,-0.479,1.2474,-0.7059,0,1,1.3931,-0.479,1.2474,-0.7059,0.0779,1,1.3931,-0.479,1.2474,-0.7059,0.164,1,1.3931,-0.479,1.2474,-0.7059,0.2547,1,1.3931,-0.479,1.2474,-0.7059,0.321,1,1.3931,-0.479,1.2474,-0.7059,0.4254,1,1.3931,-0.479,1.2474,-0.7059,0.504,1,1.3931,-0.479,1.2474,-0.7059,0.5722,1,1.3931,-0.479,1.2474,-0.7059,0.6694,1,1.3931,-0.479,1.2474,-0.7059,0.7629,0,0.134),
tolerance = 1e-2)
expect_equal(extract.mirt(rsm, 'df'), 9765610)
expect_equal(extract.mirt(grsm, 'df'), 9765601)
graded <- mirt(data, 1, verbose = FALSE)
gM2 <- M2(graded, calcNull=TRUE)
expect_equal(gM2$M2, 16.68317, tolerance = 1e-4)
expect_equal(gM2$df, 5)
expect_equal(gM2$CFI, .9723554, tolerance = 1e-4)
expect_equal(gM2$SRMSR, 0.01984457, tolerance = 1e-4)
#item and test info
Theta <- matrix(seq(-4,4,.01))
x <- extract.item(modp1, 1)
iinfo <- iteminfo(x, Theta)
expect_class(iinfo, 'numeric')
iinfo <- iteminfo(x, Theta, total.info=FALSE)
expect_class(iinfo, 'matrix')
tinfo <- testinfo(modp1, Theta)
expect_class(tinfo, 'numeric')
ER <- fscores(modp2, returnER = TRUE)
expect_equal(as.numeric(ER), c(0.4882546, 0.5099054), tolerance=1e-4)
suppressWarnings(ER2 <- fscores(modp2, returnER = TRUE, mean = c(-1, 1), cov = matrix(c(1.5,1,1,2), 2)))
expect_equal(as.numeric(ER2), c(0.5433246, 0.5929488), tolerance=1e-4)
})
Try the mirt package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.