Comparing Student Attitudes of Cheating Behaviors in the Physical and Online Environments with an Emphasis on AI Usage

Abstract

This study was conducted to compare students’ beliefs about the seriousness of cheating behaviors in the physical and online environment and to analyze how these beliefs relate to self-reported cheating behaviors.  Given the recent advances of artificial intelligence (AI) and its growing presence in the college classroom, specific emphasis is placed on cheating behaviors related to this technology.  Using a quantitative descriptive approach, a survey was distributed to 328 undergraduate students at a small suburban college in the United States. Using a 5-point Likert scale, students were asked to rank the perceived seriousness of 25 cheating behaviors related to the physical and online classroom environments and to report how frequently they engaged in each of the behaviors. Fifteen of the cheating behaviors were comparable across both environments and an additional 5 behaviors specific to each environment were included. The study aimed to address the following research questions: (RQ1) Is there a difference between the perceived seriousness of academic cheating behaviors in the physical classroom compared to the online classroom? (RQ2) Do students with online experience rank cheating behaviors in the online environment as more serious than those students without online experience? (RQ3) Do students self-report higher levels of cheating online than in the physical classroom?  (RQ4) Do students who perceive certain cheating behaviors as more serious forms of cheating, self-report less cheating of those behaviors?  (RQ5) Is AI perceived as a more acceptable academic cheating behavior compared to other cheating behaviors? (RQ6) Are students self-reporting the use of AI as frequently as other cheating behaviors in the online and physical classrooms?  For RQ1 a Wilcoxon signed rank test revealed that environment does matter with 6 out of 15 cheating behaviors being perceived to be more serious in the online environment and 1 behavior perceived to be more serious in the physical environment.  For RQ2, a  Mann-Whitney U test revealed that experience does matter with 14 out of 20 behaviors ranked as more unacceptable for students that had online experience.  The Wilcoxon signed rank test showed that students were more likely to self-report cheating for 6 of the 15 behaviors in the physical environment so for RQ3 there is not evidence that students cheat more often online.  Spearman’s rank correlation was used to address RQ4 revealing that overall, students who rank behaviors as more extreme, self-report those behaviors less often. For RQ5 it was found that using AI for homework assignments was considered generally more acceptable than most other cheating behaviors and for RQ6, students were found to be reporting the usage of AI as frequently or more frequently compared to most other cheating behaviors.