In an empirical comparison on a testlet-based English assessment test, the bifactor model is preferred over the second-order and unidimensional 2PL models by both AIC and BIC¶
Claim · All claims
Evidence · 1 study · theoreticalr3·q2
Currency · newest study 2009 · not yet checked against its sources · not reviewed by a person · overdue for a recheck since 2012: tool and model accuracy changes quickly
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q2 i? When three IRT models were fitted to 20 reading comprehension items in 4 testlets from 13,508 test takers, both information criteria selected the bifactor model. → Rijmen 2009
Evidence¶
Rijmen 2009¶
Rijmen, F. (2009). Three Multidimensional Models for Testlet-Based Tests: Formal Relations and an Empirical Comparison. ETS Research Report RR-09-37. http://www.ets.org/research/contact.html doi:10.1002/j.2333-8504.2009.tb02194.x
q2 · i? · theoretical · r3
Empirical comparison fitting a unidimensional 2PL model, a second-order model, and a bifactor model to a subset of 20 reading comprehension items in 4 testlets, analyzed on a sample of 13,508 persons. Table 1 reports deviance, AIC, and BIC, and "According to both the BIC and AIC, the bifactor model is the model to be selected."
"According to both the BIC and AIC, the bifactor model is the model to be selected."
Discussion¶
Related Claims¶
- Ignoring testlet effects by fitting a unidimensional 2PL model produced mild shrinkage of item intercept and loading estimates toward zero relative to the bifactor model — related
- The second-order multidimensional IRT model for testlets is formally equivalent to the testlet model, and both are restricted bifactor models — related
- The proportionality restrictions that the testlet model imposes on the bifactor model were implausible in the applied dataset, as specific-to-general loading quotients varied far from constant within testlets — related
- Model-fit statistics disagree on the best calibration for the Oral scale in the lower elementary population but agree on a two-dimensional model for the upper elementary population — related
- The reciprocal EF–ER model fits only marginally better than a forward-only model, so model preference rests on theory as much as fit — related
- Model-complexity penalties favor Spectral BKT under student-stratified cross-validation but not under item-stratified cross-validation — related
- The assessment violates IRT assumptions of unidimensionality and local independence, which the authors identify as a likely cause of the method discrepancy — related
- In concurrent unidimensional calibration the Speaking subtest dominates the Oral scale score while Listening and Speaking correlate only moderately — related
- A multidimensional IRT model with simple structure fits testlet-based reading data better than both the general polytomous testlet model and the standard 2PL/GPCM — related
- Most Turkish IRT-based scale development studies did not check the unidimensionality or local independence assumptions — related
- Item parceling at the testlet level substantially improves model fit, allowing partial (Rasch, 2PL) or full (bifactor) recovery of CAT constructs — related
- A strict one-factor model fits poorly while four-factor, second-order, and bifactor models fit better, with the bifactor model favored by formal comparisons — related
- GLAT shows robust structural validity and reliability under a 2-parameter logistic IRT model — related
- Parceling over-fits models: at the testlet level, data from more complex models fit well with simpler models — related
- Bifactor-(S-1) models of teacher Big Five personality fit student and teacher data better than single-domain or three-facets models — related
- A bifactor model of oral language and reading fits better than one-, three-, and four-factor models in every grade from 4th through 10th — related
- Higher order cognitive skills (inference, perspective taking, comprehension monitoring) fit a bifactor structure with a general cross-language factor and language-specific factors in English and Spanish — related
- Written composition quality is best described as a single unidimensional construct across English and Spanish and across narrative and opinion genres for primary-grade dual language learners — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
bifactor-model-preferred-aic-bic-testlet-test |
title |
In an empirical comparison on a testlet-based English assessment test, the bifactor model is preferred over the second-order and unidimensional 2PL models by both AIC and BIC |
description |
In an empirical comparison on a testlet-based English assessment test, the bifactor model is preferred over the second-order and unidimensional 2PL models by both AIC and BIC |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2009 |
recheck_by |
2012 |
generated |
by process:wiki-ingest on 2026-09-26 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — rijmen-2009 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.