Plausible Values for multiple group MIRT models (from tam.mml.3pl)

[thkiefer]

Hi Alex,

following this stats.stackexchange issue I came accross that there is no Plausible Value drawing for multiple group MIRT models. That ist, the following code terminates with an error.

library(TAM)

## load longitudinal data
library(sirt)
data(data.long, package="sirt")
dat <- data.long

## tamaan specification
tammodel <- "
  LAVAAN MODEL:
    T1=~ 1*I1T1__I6T1
    T2=~ 1*I3T2__I8T2
    T1 ~~ T1
    T2 ~~ T2
    T1 ~~ T2
    # constraint on item difficulties
    I3T1 + I3T2 | -5*t1
    I4T1 + I4T2 | -2*t1
    I5T1 + I5T2 | 3*t1
    I6T1 + I6T2 | 4*t1
    "

Y <- data.frame("female" = sample(0:1, nrow(data.long), replace = TRUE),
                model.matrix(~0+school, data.frame("school" = factor(rep(1:5, each = 40)))))
cohort <- rep(rep(c(2017, 2018), each = 20), 5)
mod14 <- TAM::tamaan(tammodel, resp = data.long, Y = Y, group = cohort,
                     tam.method = "tam.mml.3pl" # for multidimensional multiple group
                     )

pv <- tam.pv(tamobj = mod14)

[alexanderrobitzsch]

I agree that this specification has not been programmed for tam.pv. I will likely include an error message in this situation.