lordplot: Create plots in the style of Lord (1953)

In begavett/lordplot: Plot that Relates Test Scores to the Traits Underlying the Test

Description

This package replicates plots produced by Lord in his 1953 article, "The relation of test score to the trait underlying the test." This function takes item response theory parameters (currently 2- and 3-parameter logistic) as the primary inputs and uses them to create plots relating a theoretical normal distribution (representing the latent trait and plotted on the x-axis) to the "true score" distribution (representing a large number of possible observed test results and plotted on the y-axis).

Usage

lordplot(a = c(0.2, 0.4, 0.6), b = c(-2, -1, 0), c = 0, D = 1.7,
  xlim = c(-4, 4), xby = 0.5, main = NULL, xcol = "lightgrey",
  ycol = "lightblue", xlabcol = xcol, ylabcol = ycol)

Arguments

a	numeric; item discrimination parameter(s), a vector of length n, where n is the number of test items.
b	numeric; item difficulty/location parameter(s), a vector of length n, where n is the number of test items.
c	numeric; item guessing parameter(s), either a single value or a vector of length n, where n is the number of test items.
D	numeric; constant used to generate results that are roughly equal to the probit scale.
xlim	numeric vector of length 2, giving the range of the x coordinates.
xby	numeric; gives the size of histogram breaks along the x-axis.
main	character; plot title.
xcol	the color of the x-axis histogram. The x-axis label also defaults to this color if not specified.
ycol	the color of the y-axis histogram. The y-axis label also defaults to this color if not specified.
xlabcol	the color of the x-axis label (defaults to xcol).
ylabcol	the color of the y-axis label (defaults to ycol).

Author(s)

Brandon E. Gavett, [e-mail]

Richard Jones, [e-mail]

References

Lord, F. M. (1953). The relation of test score to the trait underlying the test. Educational and Psychological Measurement, 4, 517-549.

Examples

## An example using data and results from the ltm package.
library(ltm)
data(LSAT)
LSAT2PL <- coef(ltm(LSAT ~ z1))
lordplot(a = LSAT2PL[, 2], b = LSAT2PL[, 1])

begavett/lordplot documentation built on Nov. 4, 2019, 7:13 a.m.