Conditional Standard Error of Measurement Tools for Test Scores

Compute and compare conditional standard errors of measurement (CSEM) across score distributions using methods from classical test theory. Includes approaches for smoothing, bootstrapped CSEM, standardized CSEM, CSEM for scale scores, and assessment of properties of split-half scores. Also supports comparison with global standard errors derived from reliability coefficients and graphical visualization of CSEM curves and relative precision across observed score ranges. Some of these implemented methods are based on work by Lord (1955) , Feldt and Qualls (1996) , McNeish and Dumas (2025) .

CSEM: A Package for Computing Conditional Standard Errors of Measurement

The CSEM package provides functions to compute various methods of Conditional Standard Error of Measurement (CSEM), including:

• MF-CSEM: Mollenkopf-Feldt method
• Thorndike-CSEM: Based on half-test differences
• Binomial-CSEM: Based on the binomial model
• Other Reliability-Based Methods

Additionally, the package includes tools for: - Splitting test items into halves (check.split) - Computing reliability estimates (check.alpha, check.angoff) - Testing the distribution of split scores (check.distribution) - Visualizing results (check.plot)

Installation

To install and load the package, run:

```r # Install from GitHub (if applicable) devtools::install_github(“cmerinos/CSEM”)

Load the package

library(CSEM)

Getting Started

Example data

set.seed(123) dhalf1 <- matrix(sample(1:5, 30, replace = TRUE), nrow = 10) dhalf2 <- matrix(sample(1:5, 30, replace = TRUE), nrow = 10)

Example: Computing MF-CSEM

csemMF(half1 = dhalf1, half2 = dhalf2, data = dataset, reliability.coef = 0.9, n.items = 6, min.score.item = 0, max.score.item = 4, conf.level = .95)

Example: Computing Thorndike-CSEM

csemthorndike(half1 = dhalf1, half2 = dhalf2, n.groups = 4)

Example: Checking Score Distributions

check.distribution(half1 = dhalf1, half2 = dhalf2, B = 2000, conf = .95)

Documentation

?CSEM help(package = “CSEM”)

Development

• Author: Cesar Merino-Soto

• Maintainer: Cesar Merino-Soto [email protected]{.email}

• License: GPL-3

Contributions

Of course, contributions always are welcome! If you find an issue, please report it or submit a pull request.