Problems in Application of Latent Trait Models to Tailored Testing.

Autor/inn/en	Koch, William R.; Reckase, Mark D.
Institution	Missouri Univ., Columbia.
Titel	Problems in Application of Latent Trait Models to Tailored Testing.
Quelle	(1979), (23 Seiten) PDF als Volltext kostenfreie Datei Verfügbarkeit
Sprache	englisch
Dokumenttyp	gedruckt; online; Monographie
Schlagwörter	Achievement Tests; Adaptive Testing; Educational Testing; Higher Education; Item Analysis; Item Banks; Mathematical Models; Multiple Choice Tests; Psychometrics; Research Reports; Test Construction; Test Reliability; Test Validity; Verbal Tests; Vocabulary Skills + Suchen Sie Ihr Suchwort? Achievement test; Achievement; Testing; Test; Tests; Leistungsbeurteilung; Leistungsüberprüfung; Leistung; Testdurchführung; Testen; Hochschulbildung; Hochschulsystem; Hochschulwesen; Itemanalyse; Mathematical model; Mathematisches Modell; Multiple choice examinations; Multiple-choice tests, Multiple-choice examinations; Multiple-Choice-Verfahren; Psychometry; Psychometrie; Research report; Forschungsbericht; Testaufbau; Testreliabilität; Testvalidität; Mündliche Prüfung; Aktiver Wortschatz
Abstract	Tailored testing procedures for achievement testing were applied in a situation that failed to meet some of the specifications generally considered to be necessary for tailored testing. Discrepancies from the appropriate conditions included the use of small samples for calibrating items, and the use of an item pool that was not designed to be homogeneous in content. The item pool contained 180 items concerning educational measurement that were calibrated separately by a one-parameter logistic model and by a three-parameter logistic model. The 110 undergraduate students were each tested at two sessions a week apart with both one-parameter and three-parameter tailored tests at each session. All tests were administered on a computer terminal. The results were studied for several characteristics including: goodness of fit of the observed responses to those predicted by each model; the information function of each test compared to a fifty-item traditional paper-and-pencil test; the reliabilities of the tailored tests; and the content validities of the tailored tests. A unidimensional tailored test of vocabulary was also administered, with satisfactory results. The achievement tests generally produced unsatisfactory results, presumably because of the discrepancies from appropriate conditions. (CTM)
Erfasst von	ERIC (Education Resources Information Center), Washington, DC