ABSTRACT
This chapter aims to extend the erasure detection index (EDI) to detect tampering at the class, school, or district level. It provides a brief description of the EDI, presents an extension of this index that easily allows for its application to group-level data, and conduct a simulation study to demonstrate the Type I error rate and power of the group-based EDI in a variety of contexts. In a school-based setting, reasonable groups for which tampering might be explored include class, school, and district, all of which are typically known in advance. For a fixed number of tampered classes, the general pattern of results is highly similar to those found when detecting classes. Most erasure-based methods to detect test tampering are based on empirical distributions of counts of total or wrong-to-rights (WTR) erasures. The common K-12 dataset included G. Rasch-based item parameters, which could have been used to estimate examinee ability and determine the probability of WTR erasures.
