Item Response Theory Scaling with Heterogeneous Populations.

Autor/in	Blais, Jean-Guy
Titel	Item Response Theory Scaling with Heterogeneous Populations.
Quelle	(1993), (49 Seiten) PDF als Volltext kostenfreie Datei Verfügbarkeit
Sprache	englisch
Dokumenttyp	gedruckt; online; Monographie
Schlagwörter	Academic Achievement; Adolescents; Cross Cultural Studies; Data Analysis; Diversity (Student); Elementary School Students; Equated Scores; Estimation (Mathematics); Foreign Countries; International Studies; Item Response Theory; Junior High School Students; Junior High Schools; Mathematics Tests; Preadolescents; Primary Education; Reference Groups; Sampling; Scaling; Science Tests + Suchen Sie Ihr Suchwort? Schulleistung; Adolescent; Adolescence; Adoleszenz; Jugend; Jugendalter; Jugendlicher; Cultural comparison; Kulturvergleich; Auswertung; Estimation; Mathematics; Schätzung; Ausland; Internationaler Studiengang; Item-Response-Theorie; Junior High Schools; Student; Students; Sekundarstufe I; Schüler; Schülerin; Pre-adolescence; Präadoleszenz; Primarbereich; Reference group; Bezugsgruppe; Scale construction; Skalenkonstruktion
Abstract	Tools used in scaling proficiency scores from the Second International Assessment of Educational Progress (IAEP) are described. The second IAEP study, conducted in 1991, was an international comparative study of the mathematics and science skills of samples of 9- and 13-year-old students from 20 countries. This paper focuses on part of the second stage of data analysis, work done in creating a unique scale for all the participating populations, that is, creating reference populations, scaling methodology, and linkage of 9- and 13-year-old populations. All populations contributed to four combined reference populations, superpopulations, one for each age group by mathematics and science combination. The item response theory scaling model is the three-parameter logistic model. Linking was accomplished through a small set of common items included in the 9- and 13-year-old assessments in each subject area. Results show that even if on an item-by-item basis the equating gives results that are not ideal, when all items are taken into account, student proficiency scores estimated with both sets of item parameter estimates can be considered to be on the same scale. Results further indicate that no information has been lost as a result of using one set of item parameter estimates over another. Fourteen tables and 16 figures present analysis data. (SLD)
Erfasst von	ERIC (Education Resources Information Center), Washington, DC