The limits of measurement: problems in measuring trends for low-performing countries.

Autor/inn/en	Martin, Michael O.; Mullis, Ina V. S.; Foy, Pierre
Titel	The limits of measurement: problems in measuring trends for low-performing countries.
Quelle	Aus: McElvany, Nele (Hrsg.); Holtappels, Heinz Günter (Hrsg.): Empirische Bildungsforschung. Theorien, Methoden, Befunde und Perspektiven. Festschrift für Wilfried Bos. Münster u.a.: Waxmann (2013) S. 189-204 PDF als Volltext Link als defekt meldenVerfügbarkeit
Sprache	englisch
Dokumenttyp	online; gedruckt; Sammelwerksbeitrag
ISBN	3-8309-2888-2; 978-3-8309-2888-1
Schlagwörter	Bildungsforschung; Empirische Forschung; TIMSS (Third International Mathematics and Science Study); Fehler; Kritik; Messung; Schüler; Internationaler Vergleich; Bildungsforschung; Leistungsbeurteilung; Empirische Forschung; Schüler; Fehler; Leistungsbeurteilung; Schülerleistung; Messung; Messverfahren; Internationaler Vergleich; Kritik; IEA (International Association for the Evaluation of Educational Achievement); IGLU (Internationale Grundschul-Lese-Untersuchung)
Abstract	This research uses results from PIRLS 2006 and TIMSS 2007 to examine indicators of the match and mismatch between the demands of the assessments (in terms of the difficulty of the assessment items) and the distribution of student achievement in each country. The study uses percent correct indicators, including the average percentage of items answered correctly and the percentage of students scoring at no better than chance level, as indicators of match and mismatch. It also considers match and mismatch using scale based indicators derived from the TIMSS and PIRLS achievement scaling. The study includes (in an appendix) a description of a simulation illustrating how a floor effect could result in bias due to overestimating achievement in countries with very low achievement levels. Based on these results, the study suggests criteria for signaling when measurement has reached its limits and discusses options available to countries with a mismatch between the level of development of their education systems and the demands of the TIMSS and PIRLS assessments. These include administering the assessment in a higher grade (e.g., sixth grade instead of fourth grade) and adopting a less demanding assessment - prePIRLS - instead of, or as well as, PIRLS. (DIPF/Orig.).
Erfasst von	DIPF \| Leibniz-Institut für Bildungsforschung und Bildungsinformation, Frankfurt am Main
Update	2014/2