IRT.repDesign function

Generation of a Replicate Design for IRT.jackknife

This function generates a Jackknife replicate design which is necessary to use the IRT.jackknife function. The function is a wrapper to BIFIE.data.jack in the BIFIEsurvey package.

IRT.repDesign(data, wgt=NULL, jktype="JK_TIMSS", jkzone=NULL, jkrep=NULL,
   jkfac=NULL, fayfac=1, wgtrep="W_FSTR", ngr=100, Nboot=200, seed=.Random.seed)

Arguments

• data: Dataset which must contain weights and item responses
• wgt: Vector with sample weights
• jktype: Type of jackknife procedure for creating the BIFIE.data object. jktype="JK_TIMSS" refers to TIMSS/PIRLS datasets. The type "JK_GROUP" creates jackknife weights based on a user defined grouping, the type "JK_RANDOM" creates random groups. The number of random groups can be defined in ngr. The argument type="RW_PISA" extracts the replicated design with balanced repeated replicate weights from PISA datasets into objects of class IRT.repDesign. Bootstrap samples can be obtained by type="BOOT".
• jkzone: Variable name for jackknife zones. If jktype="JK_TIMSS", then jkzone="JKZONE". However, this default can be overwritten.
• jkrep: Variable name containing Jackknife replicates
• jkfac: Factor for multiplying jackknife replicate weights. If jktype="JK_TIMSS", then jkfac=2.
• fayfac: Fay factor. For Jackknife, the default is 1. For a Bootstrap with $R$ samples with replacement, the Fay factor is $1/R$.
• wgtrep: Already available replicate design
• ngr: Number of groups
• Nboot: Number of bootstrap samples
• seed: Random seed

Returns

A list with following entries - wgt: Vector with weights

• wgtrep: Matrix containing the replicate design

• fayfac: Fay factor needed for Jackknife calculations

See Also

See IRT.jackknife for further examples.

See the BIFIE.data.jack function in the BIFIEsurvey package.

Examples

## Not run:

# load the BIFIEsurvey package
library(BIFIEsurvey)

#############################################################################
# EXAMPLE 1: Design with Jackknife replicate weights in TIMSS
#############################################################################

data(data.timss11.G4.AUT, package="CDM")
dat <- CDM::data.timss11.G4.AUT$data
# generate design
rdes <- CDM::IRT.repDesign( data=dat,  wgt="TOTWGT", jktype="JK_TIMSS",
             jkzone="JKCZONE", jkrep="JKCREP" )
str(rdes)

#############################################################################
# EXAMPLE 2: Bootstrap resampling
#############################################################################

data(sim.qmatrix, package="CDM")
q.matrix <- CDM::sim.qmatrix

# simulate data according to the DINA model
dat <- CDM::sim.din(N=2000, q.matrix=q.matrix )$dat

# bootstrap with 300 random samples
rdes <- CDM::IRT.repDesign( data=dat, jktype="BOOT", Nboot=300 )
## End(Not run)