Using Deep Learning to Choose Optimal Smoothing Values for Equating.

Test developers typically use alternate test forms to protect the integrity of test scores. Because test forms may differ in difficulty, scores on different test forms are adjusted through a psychometrical procedure called equating. When conducting equating, psychometricians often apply smoothing methods to reduce random error of equating resulting from sampling. During the process, they compare plots of different smoothing degrees and choose the optimal value when using the cubic spline postsmoothing method. This manual process, however, could be automated with the help of deep learning-a machine learning technique commonly used for image classification. In this study, a convolutional neural network was trained using human-classified postsmoothing plots. The trained network was used to choose optimal smoothing values with empirical testing data, which were compared to human choices. The agreement rate between humans and the trained network was as large as 71%, suggesting the potential use of deep learning for choosing optimal smoothing values for equating.

• Publication year

  2025

• Venue

  Applied Psychological Measurement

• Publication date

  2025-08-23

• Fields of study

  Medicine, Computer Science

• Identifiers
  DOI 10.1177/01466216251363244PMID 40881830PMCID PMC12374957
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar, PubMed

• No claims are published for this paper.

• No concepts are published for this paper.

• Machine Learning and Small Data

  2021cited by this paper
• A New Statistic for Selecting the Smoothing Parameter for Polynomial Loglinear Equating Under the Random Groups Design

  2020cited by this paper
• Automated Item Generation with Recurrent Neural Networks

  2018cited by this paper
• A Comparison of Strategies for Smoothing Parameter Selection for Mixed-Format Tests Under the Random Groups Design

  2018cited by this paper
• Validation of Automated Scoring of Science Assessments.

  2016cited by this paper
• An Introduction to Machine Learning

  2015cited by this paper
• Evaluation of Two New Smoothing Methods in Equating: The Cubic B-Spline Presmoothing Method and the Direct Presmoothing Method.

  2009cited by this paper
• Comparison of Parametric and Nonparametric Bootstrap Methods for Estimating Random Error in Equipercentile Equating

  2008cited by this paper
• Test Equating, Scaling, and Linking: Methods and Practices

  2004cited by this paper
• Smoothing Methods for Estimating Test Score Distributions

  1991cited by this paper
• Effectiveness of Analytic Smoothing in Equipercentile Equating

  1984cited by this paper
• The Standard Error of Equipercentile Equating

  1981cited by this paper
• Likelihood of a model and information criteria

  1981cited by this paper
• A strong true-score theory, with applications.

  1965cited by this paper

• No citing papers are available for this paper.