Students' perceptions of both the certainty and the deterrent effect of potential consequences of cheating

Extensive research indicates that cheating among undergraduate students is a serious problem, and we have initiated a long-term investigation to identify and validate concrete approaches for reducing the frequency of cheating. We have previously presented results from that study and have (among other things) described factors that influence the frequency and definitions of cheating among engineering undergraduates and presented student opinions on what actions might prevent cheating. However, we have not reported our findings regarding the relationship between the level of assessment and various consequences of cheating on a student's decision to cheat. In this paper, results from 695 student surveys will be presented to describe our findings. In the survey, students are presented with three scenarios representing distinct opportunities for cheating (cheating on a final examination, copying solutions from another student's homework, and adding false references to a term paper). Each of these scenarios represents a different level of assessment. For each separate scenario, there are questions about three possible consequences to cheating - shame, loss of colleague's respect (i.e., embarrassment), and being caught (i.e., the threat of formal sanctions). For each of the three possible consequences, the student is asked for level of agreement with two statements: 1) a statement about being personally affected by the consequence and 2) a statement that the consequence would prevent the student from cheating. Responses to these scenario-specific questions will be compared and related to other questions from the 139-item survey, especially (1) the student's self-reported frequency of engagement in that specific behavior during college, (2) the student's categorization of the behavior as cheating, unethical, or neither, and (3) the student's self-reported frequency of high school cheating. Our results show that student responses are clustered according to level of assessment rather than consequence. This may indicate that the type of consequence has less to do with a student's decision to cheat than does the level of assessment. Several hypotheses to describe these findings will be explored, and practical approaches to reduce the level of cheating based on this work will be proposed.Peer Reviewedhttp://deepblue.lib.umich.edu/bitstream/2027.42/55261/3/2003 ASEE PACES1.pd