3PL: A Useful Model with a Mild Estimation Problem

Autor/in	Partchev, Ivailo
Titel	3PL: A Useful Model with a Mild Estimation Problem
Quelle	In: Measurement: Interdisciplinary Research and Perspectives, 7 (2009) 2, S.94-96 (3 Seiten)Infoseite zur Zeitschrift PDF als Volltext Verfügbarkeit
Sprache	englisch
Dokumenttyp	gedruckt; online; Zeitschriftenaufsatz
ISSN	1536-6367
Schlagwörter	Stellungnahme; Item Response Theory; Models; Computation; Guessing (Tests); Problems + Suchen Sie Ihr Suchwort? Item-Response-Theorie; Analogiemodell; Erraten; Problemsituation
Abstract	This author has always felt a bit uneasy about the three parameter logistic (3PL) model because it seems about as prone to discovering guessing when guessing does not take place, as the Rasch model is reluctant to admit that guessing might ever take place at all. However, the author has attributed this to the prior "magic" employed when estimating item parameters--namely, the less information on guessing that is available in the data, the more does the prior determine the estimates. The perception that one's ability to actually detect guessing might be a matter of parameterization, which is a fundamental identification problem with the models, certainly does not make life easier. In this article, the author talks about 3PL as a useful model with a mild estimation problem. The author stresses that, when plagued by insufficient data, the 3PL model has great problems with the easy items but does better with the more difficult ones. The author's practical concerns in estimating IRT models are curiously similar to the insights in Maris and Bechger--curiously, because they tend to originate from other, more mundane sources. (Contains 2 figures.) (ERIC).
Anmerkungen	Psychology Press. Available from: Taylor & Francis, Ltd. 325 Chestnut Street Suite 800, Philadelphia, PA 19106. Tel: 800-354-1420; Fax: 215-625-2940; Web site: http://www.tandf.co.uk/journals
Erfasst von	ERIC (Education Resources Information Center), Washington, DC
Update	2017/4/10