Item Response Theory Equating with the Non-Equivalent Groups with Covariates Design

Sansivieri, Valentina (2017) Item Response Theory Equating with the Non-Equivalent Groups with Covariates Design, [Dissertation thesis], Alma Mater Studiorum Università di Bologna. Dottorato di ricerca in Scienze statistiche, 29 Ciclo. DOI 10.6092/unibo/amsdottorato/7779.
Documenti full-text disponibili:
[img]
Anteprima
Documento PDF (English) - Richiede un lettore di PDF come Xpdf o Adobe Acrobat Reader
Disponibile con Licenza: Salvo eventuali più ampie autorizzazioni dell'autore, la tesi può essere liberamente consultata e può essere effettuato il salvataggio e la stampa di una copia per fini strettamente personali di studio, di ricerca e di insegnamento, con espresso divieto di qualunque utilizzo direttamente o indirettamente commerciale. Ogni altro diritto sul materiale è riservato.
Download (1MB) | Anteprima

Abstract

We use test score equating to be able to compare different test scores from different test forms. Although it is preferable to use non-equivalent groups with anchor test (NEAT) design, it might be impossible to administer an anchor test due to test security or for other reasons. However, we still know that the groups are non-equivalent, which rules out the use of an equivalent groups (EG) design. A possibility, then, is to use non-equivalent groups with covariates (NEC) design. The overall aim of this work was to propose the use of Item Response Theory (IRT) with a NEC design. We propose the use of mixed-measurement IRT with covariates model (Tay, Newman & Vermunt, 2011; 2016) within IRT observed-score equating and IRT true-score equating to model both test scores and covariates. The proposed test equating methods are examined with simulations. The results are compared with IRT observed-score equating and IRT true-score equating methods using the EG and NEAT designs. The results from the simulations show that IRT true-score equating method doesn't work, but support the IRT observed-score equating method for which the standard errors of the equating are lower when covariates are included in the IRT model than if they are excluded. One real test dataset illustrate that the IRT observed-score equating method can be used in practice.

Abstract
Tipologia del documento
Tesi di dottorato
Autore
Sansivieri, Valentina
Supervisore
Dottorato di ricerca
Ciclo
29
Coordinatore
Settore disciplinare
Settore concorsuale
Parole chiave
Equating, Item Response Theory, Collateral Information, NEC design, Covariates
URN:NBN
DOI
10.6092/unibo/amsdottorato/7779
Data di discussione
17 Febbraio 2017
URI

Altri metadati

Statistica sui download

Gestione del documento: Visualizza la tesi

^