Posterior predictive model checking for conjunctive multidimensionality in item response theory

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

If data exhibit multidimensionality, key conditional independence assumptions of unidimensional models do not hold. The current work pursues posterior predictive model checking (PPMC) as a tool for criticizing models due to unaccounted for dimensions in data structures that follow conjunctive multidimensional models. These pursuits are couched in previous work investigating factors influencing dimensionality and dimensionality assessment. A simulation study investigates the model checking tools in the context of item response theory (IRT) for dichotomous observables, in which a unidimensional model is fit to data that follow a conjunctive multidimensional model. Key findings include (a) support for the hypothesized effects of the manipulated factors and (b) the superiority of certain discrepancy measures for conducting PPMC for dimensionality assessment.

Original languageEnglish (US)
Pages (from-to)672-694
Number of pages23
JournalJournal of Educational and Behavioral Statistics
Volume36
Issue number5
DOIs
StatePublished - Oct 2011

Fingerprint

predictive model
simulation

Keywords

  • dimensionality assessment
  • item response theory
  • local independence
  • multidimensionality
  • posterior predictive model checking

ASJC Scopus subject areas

  • Education
  • Social Sciences (miscellaneous)

Cite this

@article{f55377612b3b40128de3321b23bf4ddf,
title = "Posterior predictive model checking for conjunctive multidimensionality in item response theory",
abstract = "If data exhibit multidimensionality, key conditional independence assumptions of unidimensional models do not hold. The current work pursues posterior predictive model checking (PPMC) as a tool for criticizing models due to unaccounted for dimensions in data structures that follow conjunctive multidimensional models. These pursuits are couched in previous work investigating factors influencing dimensionality and dimensionality assessment. A simulation study investigates the model checking tools in the context of item response theory (IRT) for dichotomous observables, in which a unidimensional model is fit to data that follow a conjunctive multidimensional model. Key findings include (a) support for the hypothesized effects of the manipulated factors and (b) the superiority of certain discrepancy measures for conducting PPMC for dimensionality assessment.",
keywords = "dimensionality assessment, item response theory, local independence, multidimensionality, posterior predictive model checking",
author = "Roy Levy",
year = "2011",
month = "10",
doi = "10.3102/1076998611410213",
language = "English (US)",
volume = "36",
pages = "672--694",
journal = "Journal of Educational and Behavioral Statistics",
issn = "1076-9986",
publisher = "SAGE Publications Inc.",
number = "5",

}

TY - JOUR

T1 - Posterior predictive model checking for conjunctive multidimensionality in item response theory

AU - Levy, Roy

PY - 2011/10

Y1 - 2011/10

N2 - If data exhibit multidimensionality, key conditional independence assumptions of unidimensional models do not hold. The current work pursues posterior predictive model checking (PPMC) as a tool for criticizing models due to unaccounted for dimensions in data structures that follow conjunctive multidimensional models. These pursuits are couched in previous work investigating factors influencing dimensionality and dimensionality assessment. A simulation study investigates the model checking tools in the context of item response theory (IRT) for dichotomous observables, in which a unidimensional model is fit to data that follow a conjunctive multidimensional model. Key findings include (a) support for the hypothesized effects of the manipulated factors and (b) the superiority of certain discrepancy measures for conducting PPMC for dimensionality assessment.

AB - If data exhibit multidimensionality, key conditional independence assumptions of unidimensional models do not hold. The current work pursues posterior predictive model checking (PPMC) as a tool for criticizing models due to unaccounted for dimensions in data structures that follow conjunctive multidimensional models. These pursuits are couched in previous work investigating factors influencing dimensionality and dimensionality assessment. A simulation study investigates the model checking tools in the context of item response theory (IRT) for dichotomous observables, in which a unidimensional model is fit to data that follow a conjunctive multidimensional model. Key findings include (a) support for the hypothesized effects of the manipulated factors and (b) the superiority of certain discrepancy measures for conducting PPMC for dimensionality assessment.

KW - dimensionality assessment

KW - item response theory

KW - local independence

KW - multidimensionality

KW - posterior predictive model checking

UR - http://www.scopus.com/inward/record.url?scp=80053945061&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80053945061&partnerID=8YFLogxK

U2 - 10.3102/1076998611410213

DO - 10.3102/1076998611410213

M3 - Article

AN - SCOPUS:80053945061

VL - 36

SP - 672

EP - 694

JO - Journal of Educational and Behavioral Statistics

JF - Journal of Educational and Behavioral Statistics

SN - 1076-9986

IS - 5

ER -