psychmeta (version 2.6.4)

simulate_r_sample: Simulation of data with measurement error and range-restriction artifacts

Description

This function simulates a psychometric sample and produces correlation matrices, artifact information, and other descriptive statistics that have been affected by measurement error and/or range restriction. It allows the formation of composite variables within the simulation and allows selection to be performed on any or all variables, including composites. By setting the sample size to n = Inf, users can explore the effects of measurement error and/or range restriction on parameters without the influence of sampling error. To generate multiple samples and compile a database of simulated statistics, see the simulate_r_database function.

Usage

simulate_r_sample(
  n,
  rho_mat,
  rel_vec = rep(1, ncol(rho_mat)),
  mu_vec = rep(0, ncol(rho_mat)),
  sigma_vec = rep(1, ncol(rho_mat)),
  sr_vec = rep(1, ncol(rho_mat)),
  k_items_vec = rep(1, ncol(rho_mat)),
  wt_mat = NULL,
  sr_composites = NULL,
  var_names = NULL,
  composite_names = NULL,
  n_as_ni = FALSE,
  ...
)

Value

A list of study information, including correlations, reliabilities, standard deviations, means, and u ratios for true scores and for observed scores.

Arguments

n

Number of cases to simulate before performing selection. If Inf, function will simulate parameter values.

rho_mat

Matrix of true-score correlations.

rel_vec

Vector of reliabilities corresponding to the variables in rho_mat.

mu_vec

Vector of means.

sigma_vec

Vector of observed-score standard deviations.

sr_vec

Vector of selection ratios corresponding to the variables in rho_mat (set selection ratios to 1 for variables that should not be used in selection).

k_items_vec

Number of test items comprising each of the variables to be simulated (all are single-item variables by default).

wt_mat

Optional matrix of weights to use in forming a composite of the variables in rho_mat. Matrix should have as many rows (or vector elements) as there are variables in rho_mat.

sr_composites

Optional vector selection ratios for composite variables. If not NULL, sr_composites must have as many elements as there are columns in wt_mat.

var_names

Vector of variable names corresponding to the variables in rho_mat.

composite_names

Optional vector of names for composite variables.

n_as_ni

Logical argument determining whether n specifies the incumbent sample size (TRUE) or the applicant sample size (FALSE; default). This can only be TRUE when only one variable is involved in selection.

...

Further arguments.

Examples

if (FALSE) {
## Generate data for a simple sample with two variables:
simulate_r_sample(n = 1000, rho_mat = matrix(c(1, .5, .5, 1), 2, 2),
          rel_vec = c(.8, .8), sr_vec = c(1, .5), var_names = c("Y", "X"))

## Generate data for samples with five variables, of which subsets are used to form composites:
rho_mat <- matrix(.5, 5, 5)
diag(rho_mat) <- 1

## Simulate parameters by supply n = Inf
simulate_r_sample(n = Inf, rho_mat = rho_mat,
                rel_vec = rep(.8, 5), sr_vec = c(1, 1, 1, 1, .5),
                wt_mat = cbind(c(0, 0, 0, .3, 1), c(1, .3, 0, 0, 0)), sr_composites = c(.7, .5))

## Finite sample sizes allow the generation of sample data
simulate_r_sample(n = 1000, rho_mat = rho_mat,
                rel_vec = rep(.8, 5), sr_vec = c(1, 1, 1, 1, .5),
                wt_mat = cbind(c(0, 0, 0, .3, 1), c(1, .3, 0, 0, 0)), sr_composites = c(.7, .5))
}