see: Bootstrap standard error of equating (SEE)

In twolodzko/equi: equi: Equipercentile Equating

Description

Compute bootstrap Standard Error of Equating (SEE).

Usage

see(data, fun, clusters, reps = 100)

## S3 method for class 'see'
plot(x, type = "l", col = 1, lty = 1, ylab = "yx",
  xlab = "x", alpha = 0.5, ...)

Arguments

data	a data frame.
fun	a function to pass to bootstrap. The function must return a concordance table with source test scores in first column named "x", and equated scores in the second column named "yx" as in equi or lin functions.
clusters	vector or list of vectors of cluster groupping variables.
reps	number of bootstrap repetitions (100 by default).
x	object to plot or print.
type,col,lty,xlab,ylab	arguments of the plot function.
alpha	plot lines opacity (range [0, 1]).
...	potentially further arguments passed from other methods.

Value

Returns Bias, SE and RMSE values for each score point and averaged SEE values coputed with mean or average weighted on score points probabilities.

References

Efron, B. & Tibshirani, R.J. (1993). An Introduction to the Bootstrap. London: Chapman & Hall/CRC.

Field, C.A. & Welsh, A.H. (2007). Bootstrapping clustered data. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 69(3), 369-390.

Kolen, M.J. & Brennan, R.J. (2004). Test Equating, Scaling, and Linking: Methods and Practices. New York: Springer-Verlag.

Rena, S., Lai, H., Tong, W., Aminzadeh, M., Hou, X. & Lai, S. (2010). Nonparametric bootstrapping for hierarchical data. Journal of Applied Statistics, 37(9), 1487-1498.

von Davier, A.A., Holland, P.W. & Thayer, D.T. (2004). The Kernel Method of Test Equating. New York: Springer-Verlag.

Wang, C. (2011). An investigation of bootstrap methods for estimating the standard error of equating under the common-item nonequivalent groups design. doctoral PhD diss., University of Iowa. http://ir.uiowa.edu/etd/1188

See Also

equi, smoothtab

Examples

data(Tests)
x <- Tests[Tests$Sample == "P", "x"]
y <- Tests[Tests$Sample == "P", "y"]

data <- data.frame(x=x, y=y)

# a function to be passed to bootstrap

myfun1 <- function(data) {
  eq <- equi(smoothtab(data$x, data$y, presmoothing=TRUE))
  return(eq$conc)
}

# note: the number of iterations is small
#       only to run faster as an example

see(data, myfun1, reps=25)

## Bootstrap for NEAT-CE design

data(Tests)

myfun2 <- function(Tests) {
  p <- Tests[Tests$Sample == "P", 1:2]
  q <- Tests[Tests$Sample == "Q", 2:3]
 eq <- equi(smoothtab(p), smoothtab(q))
  return(eq$conc)
}

see(Tests, myfun2, reps=25)

twolodzko/equi documentation built on Nov. 11, 2020, 4:04 p.m.