marginal_psychometrics_MC: Marginal psychometrics for multiple criteria identification...

In mcbeem/giftedCalcs: Psychometrics for Gifted Program Identification

Marginal psychometrics for multiple criteria identification systems

Description

This function calculates marginal psychometrics (currently only sensitivity) for one- and two-stage multiple criteria identification systems. A plot method is available for visualizing the distributions of scores on each assessment for the identified students. This function uses Monte Carlo simulation to approximate the integrals involved in computing the metrics; as such the results will vary somewhat from run to run. Users should determine the necessary sample size (n) needed to achieve the required level of precision. The default value of n=50000 will likely suffice for most situations. Users may also wish to set a random number seed for reproducibility.

Usage

marginal_psychometrics_MC(policy, corr, rely, n=50000, nomination=NA,
  ignore_nomination=FALSE, labels=NA)

Arguments

policy	a matrix describing the identification policy. assessments are in columns, pathways in are in rows. values are percentile cutoffs. multiple requirements within a row are joined by "and" combination rules, whereas the "or" rule joins across rows
corr	a correlation matrix
rely	a vector of reliability coefficients
n	scalar, the number of samples to draw. defaults to 50,000
nomination	vector defining which columns of the policy matrix, row / column of the correlation matrix, and element of the reliability vector is the nomination. Defaults to NA, which is interpreted as no nomination stage
ignore_nomination	boolean. Should the nomination be ignored? This allows for convenient comparison of single- and two-stage versions of a multiple criteria policy without needing to respecify the other inputs. defaults to FALSE
labels	an optional vector of labels for the assessments; defaults to NA

Value

a list with the following elements: $identified: the proportion of students that are identified $gifted: the proportion of students that are gifted $sensitivity: the sensitivity $scores: a data frame of scores for identified students

Examples

policy <- matrix(c(
  .9, .9, .9, 0,
  .9, 0, .9, .9,
  .9, 0, 0, .95
), ncol = 4, byrow = TRUE)

corr <- matrix(c(
  1, .5, .4, .3,
  .5, 1, .7, .6,
  .4, .7, 1, .5,
  .3, .6, .5, 1
), byrow = TRUE, nrow = 4)

rely <- c(.8, .9, .8, .85)

result <- marginal_psychometrics_MC(
  n = 50000, policy = policy, corr = corr,
  rely = rely, nomination = 1, labels = c("nom", "IQ", "ach", "creativity"),
  ignore_nomination = FALSE
)

result

mcbeem/giftedCalcs documentation built on May 3, 2022, 3:34 a.m.