Stochastically compute random effects for MixedClass objects with Metropolis-Hastings
samplers and averaging over the draws to obtain expected a posteriori predictions.
Returns a list of the estimated effects.
Arguments
- x
an estimated model object from the
mixedmirtfunction- ndraws
total number of draws to perform. Default is 1000
- thin
amount of thinning to apply. Default is to use every 10th draw
- return.draws
logical; return a list containing the thinned draws of the posterior?
References
Chalmers, R., P. (2012). mirt: A Multidimensional Item Response Theory Package for the R Environment. Journal of Statistical Software, 48(6), 1-29.
Chalmers, R. P. (2015). Extended Mixed-Effects Item Response Models with the MH-RM Algorithm. Journal of Educational Measurement, 52, 200-222. doi:10.1111/jedm.12072 doi:10.18637/jss.v048.i06
Author
Phil Chalmers rphilip.chalmers@gmail.com
Examples
# \donttest{
# make an arbitrary groups
covdat <- data.frame(group = rep(paste0('group', 1:49), each=nrow(Science)/49))
# partial credit model
mod <- mixedmirt(Science, covdat, model=1, random = ~ 1|group)
summary(mod)
#>
#> Call:
#> mixedmirt(data = Science, covdata = covdat, model = 1, random = ~1 |
#> group)
#>
#>
#> --------------
#> RANDOM EFFECT COVARIANCE(S):
#> Correlations on upper diagonal
#>
#> $Theta
#> F1
#> F1 0.95
#>
#> $group
#> COV_group
#> COV_group 0.0185
#>
effects <- randef(mod, ndraws = 2000, thin = 20)
head(effects$Theta)
#> F1
#> [1,] 0.4642744
#> [2,] 0.1621242
#> [3,] -0.6574783
#> [4,] -0.6442423
#> [5,] 0.1012442
#> [6,] 0.9235582
head(effects$group)
#> group
#> group1 0.022019000
#> group2 -0.006446242
#> group3 -0.010649155
#> group4 -0.024398305
#> group5 -0.009860348
#> group6 0.006403203
# }