Description Usage Arguments Details Value Author(s) References See Also Examples
bfactor
fits a confirmatory maximum likelihood two-tier/bifactor/testlet model to
dichotomous and polytomous data under the item response theory paradigm.
The IRT models are fit using a dimensional reduction EM algorithm so that regardless
of the number of specific factors estimated the model only uses the number of
factors in the second-tier structure plus 1. For the bifactor model the maximum
number of dimensions is only 2 since the second-tier only consists of a
ubiquitous unidimensional factor. See mirt
for appropriate methods to be used
on the objects returned from the estimation.
1 2 |
data |
a |
model |
a numeric vector specifying which factor loads on which
item. For example, if for a 4 item test with two specific factors, the first
specific factor loads on the first two items and the second specific factor
on the last two, then the vector is |
model2 |
a two-tier model specification object defined by |
group |
a factor variable indicating group membership used for multiple group analyses |
quadpts |
number of quadrature nodes to use after accounting for the reduced number of dimensions.
Scheme is the same as the one used in |
invariance |
see |
... |
additional arguments to be passed to the estimation engine. See |
bfactor
follows the item factor analysis strategy explicated by
Gibbons and Hedeker (1992), Gibbons et al. (2007), and Cai (2010).
Nested models may be compared via an approximate
chi-squared difference test or by a reduction in AIC or BIC (accessible via
anova
). See mirt
for more details regarding the
IRT estimation approach used in this package.
The two-tier model has a specific block diagonal covariance structure between the primary and secondary latent traits. Namely, the secondary latent traits are assumed to be orthogonal to all traits and have a fixed variance of 1, while the primary traits can be organized to vary and covary with other primary traits in the model.
Σ_{two-tier} = ≤ft(\begin{array}{cc} G & 0 \\ 0 & diag(S) \end{array} \right)
The bifactor model is a special case of the two-tier model when G above is a 1x1 matrix, and therefore only 1 primary factor is being modeled. Evaluation of the numerical integrals for the two-tier model requires only ncol(G) + 1 dimensions for integration since the S second order (or 'specific') factors require only 1 integration grid due to the dimension reduction technique.
Note: for multiple group two-tier analyses only the second-tier means and variances should be freed since the specific factors are not treated independently due to the dimension reduction technique.
function returns an object of class SingleGroupClass
(SingleGroupClass-class) or MultipleGroupClass
(MultipleGroupClass-class).
Phil Chalmers rphilip.chalmers@gmail.com
Cai, L. (2010). A two-tier full-information item factor analysis model with applications. Psychometrika, 75, 581-612.
Chalmers, R., P. (2012). mirt: A Multidimensional Item Response Theory Package for the R Environment. Journal of Statistical Software, 48(6), 1-29. doi: 10.18637/jss.v048.i06
Gibbons, R. D., & Hedeker, D. R. (1992). Full-information Item Bi-Factor Analysis. Psychometrika, 57, 423-436.
Gibbons, R. D., Darrell, R. B., Hedeker, D., Weiss, D. J., Segawa, E., Bhaumik, D. K., Kupfer, D. J., Frank, E., Grochocinski, V. J., & Stover, A. (2007). Full-Information item bifactor analysis of graded response data. Applied Psychological Measurement, 31, 4-19.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 | ## Not run:
###load SAT12 and compute bifactor model with 3 specific factors
data(SAT12)
data <- key2binary(SAT12,
key = c(1,4,5,2,3,1,2,1,3,1,2,4,2,1,5,3,4,4,1,4,3,3,4,1,3,5,1,3,1,5,4,5))
specific <- c(2,3,2,3,3,2,1,2,1,1,1,3,1,3,1,2,1,1,3,3,1,1,3,1,3,3,1,3,2,3,1,2)
mod1 <- bfactor(data, specific)
summary(mod1)
itemplot(mod1, 18, drop.zeros = TRUE) #drop the zero slopes to allow plotting
###Try with fixed guessing parameters added
guess <- rep(.1,32)
mod2 <- bfactor(data, specific, guess = guess)
coef(mod2)
anova(mod1, mod2)
## don't estimate specific factor for item 32
specific[32] <- NA
mod3 <- bfactor(data, specific)
anova(mod1, mod3)
# same, but decalred manually (not run)
#sv <- mod2values(mod1)
#sv$value[220] <- 0 #parameter 220 is the 32 items specific slope
#sv$est[220] <- FALSE
#mod3 <- bfactor(data, specific, pars = sv) #with excellent starting values
#########
# mixed itemtype example
#simulate data
a <- matrix(c(
1,0.5,NA,
1,0.5,NA,
1,0.5,NA,
1,0.5,NA,
1,0.5,NA,
1,0.5,NA,
1,0.5,NA,
1,NA,0.5,
1,NA,0.5,
1,NA,0.5,
1,NA,0.5,
1,NA,0.5,
1,NA,0.5,
1,NA,0.5),ncol=3,byrow=TRUE)
d <- matrix(c(
-1.0,NA,NA,
-1.5,NA,NA,
1.5,NA,NA,
0.0,NA,NA,
2.5,1.0,-1,
3.0,2.0,-0.5,
3.0,2.0,-0.5,
3.0,2.0,-0.5,
2.5,1.0,-1,
2.0,0.0,NA,
-1.0,NA,NA,
-1.5,NA,NA,
1.5,NA,NA,
0.0,NA,NA),ncol=3,byrow=TRUE)
items <- rep('2PL', 14)
items[5:10] <- 'graded'
sigma <- diag(3)
dataset <- simdata(a,d,2000,itemtype=items,sigma=sigma)
specific <- c(rep(1,7),rep(2,7))
simmod <- bfactor(dataset, specific)
coef(simmod)
#########
# testlet response model
#simulate data
set.seed(1234)
a <- matrix(0, 12, 4)
a[,1] <- rlnorm(12, .2, .3)
ind <- 1
for(i in 1:3){
a[ind:(ind+3),i+1] <- a[ind:(ind+3),1]
ind <- ind+4
}
print(a)
d <- rnorm(12, 0, .5)
sigma <- diag(c(1, .5, 1, .5))
dataset <- simdata(a,d,2000,itemtype=rep('2PL', 12),sigma=sigma)
# estimate by applying constraints and freeing the latent variances
specific <- c(rep(1,4),rep(2,4), rep(3,4))
model <- "G = 1-12
CONSTRAIN = (1, a1, a2), (2, a1, a2), (3, a1, a2), (4, a1, a2),
(5, a1, a3), (6, a1, a3), (7, a1, a3), (8, a1, a3),
(9, a1, a4), (10, a1, a4), (11, a1, a4), (12, a1, a4)
COV = S1*S1, S2*S2, S3*S3"
simmod <- bfactor(dataset, specific, model)
coef(simmod, simplify=TRUE)
#########
# Two-tier model
#simulate data
set.seed(1234)
a <- matrix(c(
0,1,0.5,NA,NA,
0,1,0.5,NA,NA,
0,1,0.5,NA,NA,
0,1,0.5,NA,NA,
0,1,0.5,NA,NA,
0,1,NA,0.5,NA,
0,1,NA,0.5,NA,
0,1,NA,0.5,NA,
1,0,NA,0.5,NA,
1,0,NA,0.5,NA,
1,0,NA,0.5,NA,
1,0,NA,NA,0.5,
1,0,NA,NA,0.5,
1,0,NA,NA,0.5,
1,0,NA,NA,0.5,
1,0,NA,NA,0.5),ncol=5,byrow=TRUE)
d <- matrix(rnorm(16))
items <- rep('2PL', 16)
sigma <- diag(5)
sigma[1,2] <- sigma[2,1] <- .4
dataset <- simdata(a,d,2000,itemtype=items,sigma=sigma)
specific <- c(rep(1,5),rep(2,6),rep(3,5))
model <- '
G1 = 1-8
G2 = 9-16
COV = G1*G2'
#quadpts dropped for faster estimation, but not as precise
simmod <- bfactor(dataset, specific, model, quadpts = 9, TOL = 1e-3)
coef(simmod, simplify=TRUE)
summary(simmod)
itemfit(simmod, QMC=TRUE)
M2(simmod, QMC=TRUE)
residuals(simmod, QMC=TRUE)
## End(Not run)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.