A Sparse Latent Class Model for Cognitive Diagnosis

Yinyin Chen, Steven Culpepper, Feng Liang

Research output: Contribution to journalArticle

Abstract

Cognitive diagnostic models (CDMs) are latent variable models developed to infer latent skills, knowledge, or personalities that underlie responses to educational, psychological, and social science tests and measures. Recent research focused on theory and methods for using sparse latent class models (SLCMs) in an exploratory fashion to infer the latent processes and structure underlying responses. We report new theoretical results about sufficient conditions for generic identifiability of SLCM parameters. An important contribution for practice is that our new generic identifiability conditions are more likely to be satisfied in empirical applications than existing conditions that ensure strict identifiability. Learning the underlying latent structure can be formulated as a variable selection problem. We develop a new Bayesian variable selection algorithm that explicitly enforces generic identifiability conditions and monotonicity of item response functions to ensure valid posterior inference. We present Monte Carlo simulation results to support accurate inferences and discuss the implications of our findings for future SLCM research and educational testing.

Original languageEnglish (US)
JournalPsychometrika
DOIs
StateAccepted/In press - Jan 1 2020

Fingerprint

Latent Class Model
Identifiability
Educational Models
Social Sciences
Research
Personality
Learning
Bayesian Variable Selection
Latent Process
Psychology
Model Diagnostics
Latent Variable Models
Response Function
Variable Selection
Social sciences
Monotonicity
Monte Carlo Simulation
Likely
Valid
Testing

Keywords

  • Bayesian variable selection
  • identifiability
  • sparse latent class models

ASJC Scopus subject areas

  • Psychology(all)
  • Applied Mathematics

Cite this

A Sparse Latent Class Model for Cognitive Diagnosis. / Chen, Yinyin; Culpepper, Steven; Liang, Feng.

In: Psychometrika, 01.01.2020.

Research output: Contribution to journalArticle

@article{a08e2e0edb384a57a1ee09ba4500eb94,
title = "A Sparse Latent Class Model for Cognitive Diagnosis",
abstract = "Cognitive diagnostic models (CDMs) are latent variable models developed to infer latent skills, knowledge, or personalities that underlie responses to educational, psychological, and social science tests and measures. Recent research focused on theory and methods for using sparse latent class models (SLCMs) in an exploratory fashion to infer the latent processes and structure underlying responses. We report new theoretical results about sufficient conditions for generic identifiability of SLCM parameters. An important contribution for practice is that our new generic identifiability conditions are more likely to be satisfied in empirical applications than existing conditions that ensure strict identifiability. Learning the underlying latent structure can be formulated as a variable selection problem. We develop a new Bayesian variable selection algorithm that explicitly enforces generic identifiability conditions and monotonicity of item response functions to ensure valid posterior inference. We present Monte Carlo simulation results to support accurate inferences and discuss the implications of our findings for future SLCM research and educational testing.",
keywords = "Bayesian variable selection, identifiability, sparse latent class models",
author = "Yinyin Chen and Steven Culpepper and Feng Liang",
year = "2020",
month = "1",
day = "1",
doi = "10.1007/s11336-019-09693-2",
language = "English (US)",
journal = "Psychometrika",
issn = "0033-3123",
publisher = "Springer New York",

}

TY - JOUR

T1 - A Sparse Latent Class Model for Cognitive Diagnosis

AU - Chen, Yinyin

AU - Culpepper, Steven

AU - Liang, Feng

PY - 2020/1/1

Y1 - 2020/1/1

N2 - Cognitive diagnostic models (CDMs) are latent variable models developed to infer latent skills, knowledge, or personalities that underlie responses to educational, psychological, and social science tests and measures. Recent research focused on theory and methods for using sparse latent class models (SLCMs) in an exploratory fashion to infer the latent processes and structure underlying responses. We report new theoretical results about sufficient conditions for generic identifiability of SLCM parameters. An important contribution for practice is that our new generic identifiability conditions are more likely to be satisfied in empirical applications than existing conditions that ensure strict identifiability. Learning the underlying latent structure can be formulated as a variable selection problem. We develop a new Bayesian variable selection algorithm that explicitly enforces generic identifiability conditions and monotonicity of item response functions to ensure valid posterior inference. We present Monte Carlo simulation results to support accurate inferences and discuss the implications of our findings for future SLCM research and educational testing.

AB - Cognitive diagnostic models (CDMs) are latent variable models developed to infer latent skills, knowledge, or personalities that underlie responses to educational, psychological, and social science tests and measures. Recent research focused on theory and methods for using sparse latent class models (SLCMs) in an exploratory fashion to infer the latent processes and structure underlying responses. We report new theoretical results about sufficient conditions for generic identifiability of SLCM parameters. An important contribution for practice is that our new generic identifiability conditions are more likely to be satisfied in empirical applications than existing conditions that ensure strict identifiability. Learning the underlying latent structure can be formulated as a variable selection problem. We develop a new Bayesian variable selection algorithm that explicitly enforces generic identifiability conditions and monotonicity of item response functions to ensure valid posterior inference. We present Monte Carlo simulation results to support accurate inferences and discuss the implications of our findings for future SLCM research and educational testing.

KW - Bayesian variable selection

KW - identifiability

KW - sparse latent class models

UR - http://www.scopus.com/inward/record.url?scp=85077683865&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85077683865&partnerID=8YFLogxK

U2 - 10.1007/s11336-019-09693-2

DO - 10.1007/s11336-019-09693-2

M3 - Article

AN - SCOPUS:85077683865

JO - Psychometrika

JF - Psychometrika

SN - 0033-3123

ER -