Implementation of the Omega (?) Index to Detect Large-Scale Systematic Cheating

Autor/in	Vista, Alvin
Titel	Implementation of the Omega (?) Index to Detect Large-Scale Systematic Cheating
Quelle	In: European Journal of Educational Research, 8 (2019) 4, S.1307-1322 (16 Seiten)Infoseite zur Zeitschrift PDF als Volltext kostenfreie Datei Verfügbarkeit
Sprache	englisch
Dokumenttyp	gedruckt; online; Zeitschriftenaufsatz
ISSN	2165-8714
Schlagwörter	Cheating; Identification; Computer Software; Standardized Tests; Item Response Theory; Multiple Choice Tests; Achievement Tests; International Assessment; Reading Achievement; Reading Tests; Grade 4; Foreign Countries; Progress in International Reading Literacy Study + Suchen Sie Ihr Suchwort? Prellen; Identifikation; Identifizierung; Standadised tests; Standardisierter Test; Item-Response-Theorie; Multiple choice examinations; Multiple-choice tests, Multiple-choice examinations; Multiple-Choice-Verfahren; Achievement test; Achievement; Testing; Test; Tests; Leistungsbeurteilung; Leistungsüberprüfung; Leistung; Testdurchführung; Testen; Leseleistung; Lesetest; School year 04; 4. Schuljahr; Schuljahr 04; Ausland
Abstract	Cheating detection is an important issue in standardized testing, especially in large-scale settings. Statistical approaches are often computationally intensive and require specialised software to conduct. We present a two-stage approach that quickly filters suspected groups using statistical testing on an IRT-based answer-copying index. We also present an approach to mitigate data contamination and improve the performance of the index. The computation of the index was implemented through a modified version of an open source R package, thus enabling wider access to the method. Using data from PIRLS 2011 (N=64,232) we conduct a simulation to demonstrate our approach. Type I error was well-controlled and no control group was falsely flagged for cheating, while 16 (combined n=12,569) of the 18 (combined n=14,149) simulated groups were detected. Implications for system-level cheating detection and further improvements of the approach were discussed. (As Provided).
Anmerkungen	Eurasian Society of Educational Research Association. Ataturk District 70.Str. No:15 Gaziantep, Turkey 27260. Tel: +90-342-2116792; Fax: +90-342-2116677; e-mail: editor@eu-jer.com; Web site: http://www.eu-jer.com
Erfasst von	ERIC (Education Resources Information Center), Washington, DC
Update	2020/1/01