PROX Estimation Method for the Rasch Model

This function estimates the Rasch model using the PROX algorithm (cited in Wright & Stone, 1999).

Usage

rasch.prox(dat, dat.resp=1 - is.na(dat), freq=rep(1,nrow(dat)),
    conv=0.001, maxiter=30, progress=FALSE)

Arguments

dat: An \(N \times I\) data frame of dichotomous response data. NAs are not allowed and must be indicated by zero entries in the response indicator matrix dat.resp.
dat.resp: An \(N \times I\) indicator data frame of nonmissing item responses.
freq: A vector of frequencies (or weights) of all rows in data frame dat.
conv: Convergence criterion for item parameters
maxiter: Maximum number of iterations
progress: Display progress?

Value

A list with following entries

b: Estimated item difficulties
theta: Estimated person abilities
iter: Number of iterations
sigma.i: Item standard deviations
sigma.n: Person standard deviations

References

Wright, B., & Stone, W. (1999). Measurement Essentials. Wilmington: Wide Range.

Examples

#############################################################################
# EXAMPLE 1: PROX data.read
#############################################################################

data(data.read)
mod <- sirt::rasch.prox( data.read )
mod$b       # item difficulties