R/SingleGroup-methods.R
traditional2mirt.RdThis is a helper function for users who have previously available traditional/classical
IRT parameters and want to know the equivalent slope-intercept translation used in mirt.
Note that this function assumes that the supplied models are unidimensional by definition (i.e.,
will have only one slope/discrimination). If there is no supported slope-intercept transformation
available then the original vector of parameters will be returned by default.
traditional2mirt(x, cls, ncat)a vector of parameters to tranform
the class or itemtype of the supplied model
the number of categories implied by the IRT model
a named vector of slope-intercept parameters (if supported)
Supported class transformations for the cls input are:
Form must be: (discrimination, difficulty, lower-bound, upper-bound)
Form must be: (discrimination, difficulty 1, difficulty 2, ..., difficulty k-1)
Form must be: (discrimination, difficulty 1, difficulty 2, ..., difficulty k-1)
Form must be: (discrimination 1, discrimination 2, ..., discrimination k, difficulty 1, difficulty 2, ..., difficulty k)
# classical 3PL model
vec <- c(a=1.5, b=-1, g=.1, u=1)
slopeint <- traditional2mirt(vec, '3PL', ncat=2)
slopeint
#> a1 d g u
#> 1.5 1.5 0.1 1.0
# classical graded model (four category)
vec <- c(a=1.5, b1=-1, b2=0, b3=1.5)
slopeint <- traditional2mirt(vec, 'graded', ncat=4)
slopeint
#> a1 d1 d2 d3
#> 1.50 1.50 0.00 -2.25
# classical generalize partial credit model (four category)
vec <- c(a=1.5, b1=-1, b2=0, b3=1.5)
slopeint <- traditional2mirt(vec, 'gpcm', ncat=4)
slopeint
#> a1 ak0 ak1 ak2 ak3 d0 d1 d2 d3
#> 1.50 0.00 1.00 2.00 3.00 0.00 1.50 1.50 -0.75
# classical nominal model (4 category)
vec <- c(a1=.5, a2 = -1, a3=1, a4=-.5, d1=1, d2=-1, d3=-.5, d4=.5)
slopeint <- traditional2mirt(vec, 'nominal', ncat=4)
slopeint
#> a1 ak0 ak1 ak2 ak3 d0 d1
#> -0.3333333 0.0000000 4.5000000 -1.5000000 3.0000000 0.0000000 -2.0000000
#> d2 d3
#> -1.5000000 -0.5000000