sirt (version 1.5-0)

md.pattern.sirt: Response Pattern in a Binary Matrix

Description

Computes different statistics of the response pattern in a binary matrix.

Usage

md.pattern.sirt(dat)

Arguments

dat

A binary data matrix

Value

• A list with following entries:
• dat
  Original dataset
• dat.resp1
  Indices for responses of 1's
• dat.resp0
  Indices for responses of 0's
• resp_patt
  Vector of response patterns
• unique_resp_patt
  Unique response patterns
• unique_resp_patt_freq
  Frequencies of unique response patterns
• unique_resp_patt_firstobs
  First observation in original dataset dat of a unique response pattern
• freq1
  Frequencies of 1's
• freq0
  Frequencies of 0's
• dat.ordered
  Dataset according to response patterns

See Also

See also the md.pattern function in the mice package.

Examples

#############################################################################
# SIMULATED EXAMPLE 1
#############################################################################
set.seed(7654)
N <- 21         # number of rows
I <- 4          # number of columns
dat <- matrix( 1*(runif(N*I) > .3 ) , N, I )
res <- md.pattern.sirt(dat)
# plot of response patterns
res$dat.ordered
image( z=t(res$dat.ordered) , y =1:N , x=1:I , xlab="Items"  , ylab="Persons")
# 0's are yellow and 1's are red

#############################################################################
# EXAMPLE 2: Item response patterns for dataset data.read
#############################################################################

data(data.read)
dat <- data.read  ; N <- nrow(dat) ; I <- ncol(dat)
# order items according to p values
dat <- dat[ , order(colMeans(dat , na.rm=TRUE )) ]
# analyzing response pattern
res <- md.pattern.sirt(dat)
res$dat.ordered
image( z=t(res$dat.ordered) , y =1:N , x=1:I , xlab="Items"  , ylab="Persons")