R/BIFIE_NMI_inference_parameters.R

In BIFIEsurvey: Tools for Survey Statistics in Educational Assessment

## File Name: BIFIE_NMI_inference_parameters.R
## File Version: 0.18


###########################################################
# NMI inference, helper function
BIFIE_NMI_inference_parameters <- function( parsM, parsrepM, fayfac,
    RR, Nimp, Nimp_NMI, comp_cov=FALSE )
{

    res0 <- bifie_comp_vcov_within( parsM, parsrepM, fayfac, RR, Nimp )
    u_diag <- res0$u_diag

    NV <- length(u_diag) / Nimp
    u_diag <- array( u_diag, dim=c( NV, Nimp_NMI[2], Nimp_NMI[1] ) )
    u_diag <- aperm( u_diag, c(3,2,1) )
    u <- array( 0, dim=c( Nimp_NMI[1], Nimp_NMI[2], NV, NV ) )

    for (ii in seq( 1, Nimp_NMI[1] ) ){
        for (jj in seq( 1, Nimp_NMI[2] ) ){
            if (NV>1){
                diag(u[ii,jj,,]) <- u_diag[ii,jj,]
            }
            if (NV==1){
                u[ii,jj,1,1] <- u_diag[ii,jj,1]
            }
        }
    }
    qhat <- array( parsM, dim=c(NV, Nimp_NMI[2], Nimp_NMI[1] ) )
    qhat <- aperm( qhat, c(3,2,1) )
    # inference using miceadds package
    res1 <- miceadds::NMIcombine( qhat=qhat, u=u, comp_cov=comp_cov, is_list=FALSE )
    res1$df <- ifelse( res1$df > 1000, Inf, res1$df )

    # output management
    res1$pars <- res1$qbar
    res1$pars_se <- sqrt( diag( res1$Tm ) )
    res1$pars_fmi <- res1$lambda
    res1$pars_fmiB <- res1$lambda_Between
    res1$pars_fmiW <- res1$lambda_Within
    res1$pars_varBetween1 <- diag(res1$Bm)
    res1$pars_varBetween2 <- diag(res1$Wm)
    res1$pars_varWithin <- diag(res1$ubar)
    return(res1)
}
#################################################################