ERIC Number: EJ1048931
Record Type: Journal
Publication Date: 2015
Pages: 19
Abstractor: As Provided
ISBN: N/A
ISSN: ISSN-0895-7347
EISSN: N/A
Erasure Analyses: Reducing the Number of False Positives
McClintock, Joseph Clair
Applied Measurement in Education, v28 n1 p14-32 2015
Erasure analysis is the study of the pattern or quantity of erasures on multiple-choice paper-and-pencil examinations, to determine whether erasures were made post-testing for the purpose of unfairly increasing students' scores. This study examined the erasure data from over 1.4 million exams, taken by more than 600,000 students. Three different methods of calculating erasures and two methods for analyzing erasures were explored. In the present study's dataset, the distribution of the mean number of erasures was positively skewed, and student ability and student race were associated with higher percentages of erasures. Techniques that do not consider these factors are likely to be biased toward flagging more classes than would be expected by chance. A simple technique for reducing the number of false positive flags is proposed.
Descriptors: Multiple Choice Tests, Cheating, Methods, Computation, Ability, Race, Evaluation Methods
Routledge. Available from: Taylor & Francis, Ltd. 325 Chestnut Street Suite 800, Philadelphia, PA 19106. Tel: 800-354-1420; Fax: 215-625-2940; Web site: http://www.tandf.co.uk/journals
Publication Type: Journal Articles; Reports - Research
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A
Grant or Contract Numbers: N/A