Watching Works; Honor Pledges Can Backfire: What Research Says About Preventing Cheating

Summary for Educators

Based on Tobias Cagala, Ulrich Glogowsky, and Johannes Rincke

“Detecting and Preventing Cheating in Exams: Evidence from a Field Experiment” Journal of Human Resources | Vol. 59, No. 1 | January 2024 | pp. 210–241
University of Wisconsin Press | Project MUSE

🔵 THE BIG IDEA

What actually prevents students from cheating? Cagala, Glogowsky, and Rincke moved beyond self-reported dishonesty by developing a method for detecting answer copying among neighboring students during undergraduate multiple-choice exams. Under normal monitoring, they estimated that at least 7.7 percent of row-wise neighboring pairs engaged in plagiarism, with academically weaker pairs more likely to cheat. The experimental findings are particularly important for educators: close monitoring eliminated detected cheating. But another seemingly sensible strategy produced the opposite result. Asking students to sign an honesty declaration before the exam doubled cheating relative to the control condition. Follow-up experiments suggested that the declaration may have inadvertently weakened the existing social norm favoring academic integrity. The leadership lesson is provocative: anti-cheating practices should be judged by evidence, not intuition. Some simple interventions work remarkably well; others can backfire.

🔵 KEY TAKEAWAYS FOR EDUCATORS

• Make monitoring visible. Students in the study responded strongly when monitoring became more intensive; close monitoring eliminated detected cheating in the experimental setting.

• Don't assume honor pledges work. Requiring an honesty declaration sounds sensible, but in this experiment it doubled cheating relative to the control group.

• Protect existing norms. If most students already believe cheating is unacceptable, an intervention that emphasizes dishonesty may inadvertently signal that cheating is more common than students assumed.

• Pay attention to opportunity. The researchers detected copying by comparing answer similarities among neighboring and non-neighboring students, underscoring how assessment conditions themselves can facilitate misconduct.

• Support struggling students before the exam. Academically weaker student pairs were more likely to engage in cheating, suggesting that prevention should include earlier academic intervention—not merely stronger enforcement on test day.

• Evaluate what actually works. Academic-integrity policies should be treated like other educational interventions: examine outcomes, question assumptions, and modify practices when evidence contradicts expectations.

◻️ WHY IT MATTERS

Cheating is notoriously difficult to study because students may not accurately report their own misconduct. This research is valuable because the authors developed statistical methods to identify copying through patterns in multiple-choice responses and then tested prevention strategies in an actual university setting. Although the findings come from higher education and should not automatically be generalized to K–12 schools, they challenge a common assumption: good intentions don't guarantee good interventions. As schools reconsider academic integrity in the age of generative AI, leaders need evidence about what changes student behavior—not simply policies that sound persuasive.

🟢 LEADERSHIP ACTION STEPS

✔ Strengthen active supervision during assessments by ensuring teachers can circulate, observe student behavior, and minimize obvious opportunities for copying.

✔ Review honor pledges and integrity statements rather than assuming they reduce misconduct; examine whether they reinforce an established positive norm or inadvertently suggest cheating is widespread.

✔ Design assessment environments that reduce unnecessary opportunities for dishonesty while preserving a respectful atmosphere of trust.

✔ Support students who are academically struggling before high-stakes assessments, addressing one potential pressure associated with dishonest behavior.

✔ Measure the effectiveness of your academic-integrity strategies by examining actual patterns of misconduct and adjusting practices when the evidence warrants it.

🟡 LEADER REFLECTION

Which of our school's anti-cheating practices do we know actually change student behavior—and which do we continue using simply because they seem as though they should work?

Original Article

Detecting and Preventing Cheating in Exams: Evidence from a Field E...

Tobias Cagala, Ulrich Glogowsky, and Johannes Rincke

“Detecting and Preventing Cheating in Exams: Evidence from a Field Experiment” Journal of Human Resources | Volume 59, Number 1 | January 2024 | pp. 210–241
DOI: 10.3368/jhr.0620-10947R1

Prepared with the assistance of AI software

OpenAI. (2026). ChatGPT (GPT-5.6 Sol) [Large language model].