A bootstrap generalization of modified parallel analysis for IRT dimensionality assessment

Holmes Finch, Patrick Monahan

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

This article introduces a bootstrap generalization to the Modified Parallel Analysis (MPA) method of test dimensionality assessment using factor analysis. This methodology, based on the use of Marginal Maximum Likelihood nonlinear factor analysis, provides for the calculation of a test statistic based on a parametric bootstrap using the MPA methodology for generation of synthetic datasets. Performance of the bootstrap test was compared with the likelihood ratio difference test and the DIMTEST procedure using a Monte Carlo simulation. The bootstrap test was found to exhibit much better control of the Type I error rate than the likelihood ratio difference test, and comparable power to DIMTEST under most conditions. A major conclusion to be taken from this research is that under many real-world conditions, the bootstrap MPA test presents a useful alternative for practitioners using Marginal Maximum Likelihood factor analysis to test for multidimensional testing data.

Original languageEnglish
Pages (from-to)119-140
Number of pages22
JournalApplied Measurement in Education
Volume21
Issue number2
DOIs
StatePublished - Apr 2008

Fingerprint

Statistical Factor Analysis
factor analysis
Research
methodology
statistics
simulation
performance
Datasets

ASJC Scopus subject areas

  • Education
  • Developmental and Educational Psychology
  • Psychology (miscellaneous)

Cite this

A bootstrap generalization of modified parallel analysis for IRT dimensionality assessment. / Finch, Holmes; Monahan, Patrick.

In: Applied Measurement in Education, Vol. 21, No. 2, 04.2008, p. 119-140.

Research output: Contribution to journalArticle

@article{eb678fe10f1c4b81b6de96acde0fab70,
title = "A bootstrap generalization of modified parallel analysis for IRT dimensionality assessment",
abstract = "This article introduces a bootstrap generalization to the Modified Parallel Analysis (MPA) method of test dimensionality assessment using factor analysis. This methodology, based on the use of Marginal Maximum Likelihood nonlinear factor analysis, provides for the calculation of a test statistic based on a parametric bootstrap using the MPA methodology for generation of synthetic datasets. Performance of the bootstrap test was compared with the likelihood ratio difference test and the DIMTEST procedure using a Monte Carlo simulation. The bootstrap test was found to exhibit much better control of the Type I error rate than the likelihood ratio difference test, and comparable power to DIMTEST under most conditions. A major conclusion to be taken from this research is that under many real-world conditions, the bootstrap MPA test presents a useful alternative for practitioners using Marginal Maximum Likelihood factor analysis to test for multidimensional testing data.",
author = "Holmes Finch and Patrick Monahan",
year = "2008",
month = "4",
doi = "10.1080/08957340801926102",
language = "English",
volume = "21",
pages = "119--140",
journal = "Applied Measurement in Education",
issn = "0895-7347",
publisher = "Routledge",
number = "2",

}

TY - JOUR

T1 - A bootstrap generalization of modified parallel analysis for IRT dimensionality assessment

AU - Finch, Holmes

AU - Monahan, Patrick

PY - 2008/4

Y1 - 2008/4

N2 - This article introduces a bootstrap generalization to the Modified Parallel Analysis (MPA) method of test dimensionality assessment using factor analysis. This methodology, based on the use of Marginal Maximum Likelihood nonlinear factor analysis, provides for the calculation of a test statistic based on a parametric bootstrap using the MPA methodology for generation of synthetic datasets. Performance of the bootstrap test was compared with the likelihood ratio difference test and the DIMTEST procedure using a Monte Carlo simulation. The bootstrap test was found to exhibit much better control of the Type I error rate than the likelihood ratio difference test, and comparable power to DIMTEST under most conditions. A major conclusion to be taken from this research is that under many real-world conditions, the bootstrap MPA test presents a useful alternative for practitioners using Marginal Maximum Likelihood factor analysis to test for multidimensional testing data.

AB - This article introduces a bootstrap generalization to the Modified Parallel Analysis (MPA) method of test dimensionality assessment using factor analysis. This methodology, based on the use of Marginal Maximum Likelihood nonlinear factor analysis, provides for the calculation of a test statistic based on a parametric bootstrap using the MPA methodology for generation of synthetic datasets. Performance of the bootstrap test was compared with the likelihood ratio difference test and the DIMTEST procedure using a Monte Carlo simulation. The bootstrap test was found to exhibit much better control of the Type I error rate than the likelihood ratio difference test, and comparable power to DIMTEST under most conditions. A major conclusion to be taken from this research is that under many real-world conditions, the bootstrap MPA test presents a useful alternative for practitioners using Marginal Maximum Likelihood factor analysis to test for multidimensional testing data.

UR - http://www.scopus.com/inward/record.url?scp=41849085226&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=41849085226&partnerID=8YFLogxK

U2 - 10.1080/08957340801926102

DO - 10.1080/08957340801926102

M3 - Article

AN - SCOPUS:41849085226

VL - 21

SP - 119

EP - 140

JO - Applied Measurement in Education

JF - Applied Measurement in Education

SN - 0895-7347

IS - 2

ER -