ABSTRACT
Screening or monitoring for test irregularities, such as systematic cheating on statewide assessment programs, is one of two important aspects of test security procedures. Manually checking examinees’ responses in each testing room for such irregularities is a daunting task given the very large number of testing rooms throughout a given state and the short turnaround time required for reporting examinees’ scores. In this regard, the use of statistical methods to screen test irregularities is a viable alternative from a standpoint of efficiency. Statistical methods for detecting cheating on tests can be broadly categorized as individual-level (Angoff, 1974; Frary, Tideman, ampentity Watts, 1977; Bay, 1994; Holland, 1996; Wollack, 1997; Lewis ampentity Thayer, 1998; Sotaridona, van der Linden ampentity Meijer, 2006; van der Linden ampentity Sotaridona, 2006; Belov ampentity Armstrong, 2010) or group-level (Jacob ampentity 2003; Sotaridona ampentity Choi, 2007; Wollack ampentity Maynes, 2011; Belov, 2011; Skorupski ampentity Egan, 2012; Shu, Henson, ampentity Luecht, 2012; Wibowo, Sotaridona, ampentity Hendrawan, 2013).
