Does Commitment to a No-Cheating Rule Affect Academic Cheating?
Last registered on August 29, 2019

Pre-Trial

Trial Information
General Information
Title
Does Commitment to a No-Cheating Rule Affect Academic Cheating?
RCT ID
AEARCTR-0004612
Initial registration date
August 23, 2019
Last updated
August 29, 2019 8:45 AM EDT
Location(s)
Region
Primary Investigator
Affiliation
University of Erlangen-Nuremberg
Other Primary Investigator(s)
PI Affiliation
Deutsche Bundesbank
PI Affiliation
University of Munich
Additional Trial Information
Status
Completed
Start date
2013-02-01
End date
2019-07-31
Secondary IDs
Abstract
Educators around the globe often require students to commit to academic integrity by signing a no-cheating declaration. This paper evaluates how such no-cheating declarations affect academic cheating. Exploiting data from a field experiment with undergraduate students, we identify cheating by comparing the similarity in multiple-choice answers of seat neighbors and counterfactual neighbors. Our main finding is that students plagiarize more after having signed a no-cheating declaration. This effect is driven by students of below-average ability. Regarding channels, we find evidence suggesting that requesting a commitment to a no-cheating rule weakens the social norm of academic integrity and triggers psychological reactance.
External Link(s)
Registration Citation
Citation
Cagala, Tobias, Ulrich Glogowsky and Johannes Rincke. 2019. "Does Commitment to a No-Cheating Rule Affect Academic Cheating?." AEA RCT Registry. August 29. https://doi.org/10.1257/rct.4612-1.0.
Sponsors & Partners

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information
Experimental Details
Interventions
Intervention(s)
In two written, 60-minute undergraduate exams at the business school of a German university, we randomly allocated students to one of two main treatment groups: A control condition and a commitment treatment. All the students in a given hall received the same treatment. The only difference between the control group and the commitment treatment was that students in the commitment treatment signed a declaration of compliance with a no-cheating rule. We placed this declaration on the cover sheet of the exam materials. It read:

"I hereby declare that I will not use unauthorized materials during the exam. Furthermore, I declare neither to use unauthorized aid from other participants nor to give unauthorized aid to other participants.''

The declaration was printed below a form in which students in all treatments had to fill in their names and university IDs. The salient location was meant to direct the students' attention to the declaration immediately before the beginning of the exam.

We also implemented a treatment with close monitoring of students (but no commitment). In this paper, we use the monitoring treatment to substantiate that our methods can identify cheating. In particular, in the spirit of previous work highlighting that close monitoring can eliminate academic cheating, we increased the monitoring intensity in the monitoring treatment to a level that we expected would eliminate plagiarism. In the empirical analysis, we then test whether, as expected, this type of treatment variation nullifies or, at least, sharply reduces the amount of cheating detected by our methods.
Intervention Start Date
2013-02-01
Intervention End Date
2019-07-31
Primary Outcomes
Primary Outcomes (end points)
Indicators for identical, identical correct, and identical incorrect answers to (multiple choice) exam questions among seat neighbors
Primary Outcomes (explanation)
In our main estimations, we use as outcomes indicators for identical answers among actual and counterfactual neighbors. Actual neighbors are students sitting next to each other in the same seat row during the exam. Counterfactual neighbor pairs are students not sitting next to each other in the exam. The pairs result from randomly drawing students either from the same hall or the same seat row.
Secondary Outcomes
Secondary Outcomes (end points)
- Share of identical answers among actual and counterfactual seat neighbors
- Perceived sanction for cheating in exam
- Perceived detection probabilities when cheating in exam
- Perceived share of other students cheating in exam
- Perceived social norm of behaving in accordance with academic integrity in exam
Secondary Outcomes (explanation)
All outcomes capturing perceptions were collected in a post-exam online survey among exam participants. The perceived social norm of behaving in accordance with acedemic integrity in exams was elicited as the perceived share of other students cheating in a hypothetical scenario with a detection probabilitiy of zero.
Experimental Design
Experimental Design
We implemented the field experiment in two written, 60-minute undergraduate exams at the business school of a German university, both of which took place in several lecture halls. The exams covered "principles of economics" (first exam) and "principles of business administration" (second exam). Both exams were compulsory for students in their first semester and were part of the curriculum for a bachelor's degree.

As for the design of the examination questions, each exam included 30 multiple-choice problems consisting of four statements. Only one of the four statements was correct. The students' task was to mark the correct statements on an answer sheet. All multiple-choice problems had the same weight, and the set of exam questions came in only one version. In a given exam, every student answered the same questions appearing in the same order.

According to the exam regulations in the department, students who cheat (e.g., by copying answers from neighbors or using mobile phones) fail the exam. It is also part of the exam regulations that supervisors in exams announce standardized examination rules by reading them aloud. As part of the announcements, supervisors highlight that cheating is prohibited and that detected cheaters would fail the exam. They also emphasize a list of actions counting as cheating attempts, including copying answers from neighbors, using unauthorized materials, and not switching off mobile phones. In the experiment, we made sure that the supervisors made the announcements as planned.

Importantly, the setting we study is one in which the level of monitoring is rather low. Commonly, up to 200 students take exams in lecture halls with up to 800 seats, supervised by only two to four members of the university staff (depending on the size of the hall). Moreover, if a supervisor files a case of attempted cheating, this leads to a significant hassle during the exam and to additional paperwork with the department's examination board after the exam. As a result, the supervising staff has little incentive to monitor students effectively.

In the experiment, the seating arrangement was as follows: Row-wise, a student was sitting in every second seat (i.e., any two students were separated by an empty seat). Column-wise, students were sitting in every second column (i.e., any two rows with students were separated by an empty seat). The fact that the row-wise distance between two students was smaller than the column-wise distance or the diagonal distance suggests that students more likely copied answers from neighbors in the same row than from students sitting in the front or the back.

Also of note is that the university does not have an honor code. Furthermore, in the years before the experiment, the department did not use any form of commitment requests to prevent cheating in exams.

The main purpose of the field experiment is to test how commitment affects cheating in exams. To that end, we randomly allocated students from two strata (gender and high-school GPA as a proxy for ability) to one of two treatment groups: A control condition and a commitment treatment. All the students in a given hall received the same treatment. We, thus, exclude spillovers between treatments, which substantiates the stable unit treatment value assumption. We also randomly assigned students to seats within the lecture halls and made sure that they took their preassigned seats.

The only difference between the control group and the commitment treatment was that students in the commitment treatment signed a declaration of compliance with the no-cheating rule. We placed this declaration on the cover sheet of the exam materials. It read:

"I hereby declare that I will not use unauthorized materials during the exam. Furthermore, I declare neither to use unauthorized aid from other participants nor to give unauthorized aid to other participants.''

The declaration was printed below a form in which students in all treatments had to fill in their names and university IDs. The salient location was meant to direct the students' attention to the declaration immediately before the beginning of the exam.

To further our understanding of the nature of commitment, two aspects of the commitment request are worth noting. First, by letting students sign the declaration, we changed the degree of commitment to an existing no-cheating rule relative to the control group, but neither varied the existence nor the content of the rule itself. In particular, the declaration did not introduce additional information regarding the rule. Instead, the public announcements, which were identical across treatments, laid out the rules by stating that cheating was prohibited and by highlighting the consequences of cheating. Second, the declaration was not morally loaded but neutral in the sense that it did not refer to any ethical norm.

We also implemented a treatment with close monitoring of students (but no commitment). The monitoring treatment helps us to substantiate that our methods can identify cheating. In particular, in the spirit of previous work highlighting that close monitoring can eliminate academic cheating, we increased the monitoring intensity in the monitoring treatment to a level that we expected would eliminate plagiarism. In the empirical analysis, we then test whether, as expected, this type of treatment variation nullifies or, at least, sharply reduces the amount of cheating detected by our methods.

As for the implementation details of close monitoring, they were as follows: In the monitoring treatment, we allocated additional supervisors to the lecture halls such that, on average, one supervisor monitored only 8.4 students, a significant decrease relative to the 44.2 students per supervisor under baseline monitoring (in control and commitment). Importantly, in all halls, supervisors remained at specific predefined spots throughout the exam. In the control and the commitment group, supervisors took positions in the front of the hall. In the monitoring treatment, the spots where supervisors located were evenly distributed all-over the hall.
Experimental Design Details
We implemented the field experiment in two written, 60-minute undergraduate exams at the business school of a German university, both of which took place in several lecture halls. The exams covered "principles of economics" (first exam) and "principles of business administration" (second exam). Both exams were compulsory for students in their first semester and were part of the curriculum for a bachelor's degree. As for the design of the examination questions, each exam included 30 multiple-choice problems consisting of four statements. Only one of the four statements was correct. The students' task was to mark the correct statements on an answer sheet. All multiple-choice problems had the same weight, and the set of exam questions came in only one version. In a given exam, every student answered the same questions appearing in the same order. According to the exam regulations in the department, students who cheat (e.g., by copying answers from neighbors or using mobile phones) fail the exam. It is also part of the exam regulations that supervisors in exams announce standardized examination rules by reading them aloud. As part of the announcements, supervisors highlight that cheating is prohibited and that detected cheaters would fail the exam. They also emphasize a list of actions counting as cheating attempts, including copying answers from neighbors, using unauthorized materials, and not switching off mobile phones. In the experiment, we made sure that the supervisors made the announcements as planned. Importantly, the setting we study is one in which the level of monitoring is rather low. Commonly, up to 200 students take exams in lecture halls with up to 800 seats, supervised by only two to four members of the university staff (depending on the size of the hall). Moreover, if a supervisor files a case of attempted cheating, this leads to a significant hassle during the exam and to additional paperwork with the department's examination board after the exam. As a result, the supervising staff has little incentive to monitor students effectively. In the experiment, the seating arrangement was as follows: Row-wise, a student was sitting in every second seat (i.e., any two students were separated by an empty seat). Column-wise, students were sitting in every second column (i.e., any two rows with students were separated by an empty seat). The fact that the row-wise distance between two students was smaller than the column-wise distance or the diagonal distance suggests that students more likely copied answers from neighbors in the same row than from students sitting in the front or the back. Also of note is that the university does not have an honor code. Furthermore, in the years before the experiment, the department did not use any form of commitment requests to prevent cheating in exams. The main purpose of the field experiment is to test how commitment affects cheating in exams. To that end, we randomly allocated students from two strata (gender and high-school GPA as a proxy for ability) to one of two treatment groups: A control condition and a commitment treatment. All the students in a given hall received the same treatment. We, thus, exclude spillovers between treatments, which substantiates the stable unit treatment value assumption. We also randomly assigned students to seats within the lecture halls and made sure that they took their preassigned seats. The only difference between the control group and the commitment treatment was that students in the commitment treatment signed a declaration of compliance with the no-cheating rule. We placed this declaration on the cover sheet of the exam materials. It read: "I hereby declare that I will not use unauthorized materials during the exam. Furthermore, I declare neither to use unauthorized aid from other participants nor to give unauthorized aid to other participants.'' The declaration was printed below a form in which students in all treatments had to fill in their names and university IDs. The salient location was meant to direct the students' attention to the declaration immediately before the beginning of the exam. To further our understanding of the nature of commitment, two aspects of the commitment request are worth noting. First, by letting students sign the declaration, we changed the degree of commitment to an existing no-cheating rule relative to the control group, but neither varied the existence nor the content of the rule itself. In particular, the declaration did not introduce additional information regarding the rule. Instead, the public announcements, which were identical across treatments, laid out the rules by stating that cheating was prohibited and by highlighting the consequences of cheating. Second, the declaration was not morally loaded but neutral in the sense that it did not refer to any ethical norm. We also implemented a treatment with close monitoring of students (but no commitment). The monitoring treatment helps us to substantiate that our methods can identify cheating. In particular, in the spirit of previous work highlighting that close monitoring can eliminate academic cheating, we increased the monitoring intensity in the monitoring treatment to a level that we expected would eliminate plagiarism. In the empirical analysis, we then test whether, as expected, this type of treatment variation nullifies or, at least, sharply reduces the amount of cheating detected by our methods. As for the implementation details of close monitoring, they were as follows: In the monitoring treatment, we allocated additional supervisors to the lecture halls such that, on average, one supervisor monitored only 8.4 students, a significant decrease relative to the 44.2 students per supervisor under baseline monitoring (in control and commitment). Importantly, in all halls, supervisors remained at specific predefined spots throughout the exam. In the control and the commitment group, supervisors took positions in the front of the hall. In the monitoring treatment, the spots where supervisors located were evenly distributed all-over the hall.
Randomization Method
Randomization done in office by a computer
Randomization Unit
Individual student
Was the treatment clustered?
No
Experiment Characteristics
Sample size: planned number of clusters
The treatment was implemented hall-wise. In our main specifications, we compare actual neighbor pairs to pairs of counterfactual neighbors from the same seat row and account for row clusters. The number of clusters in those specifications is 81.
Sample size: planned number of observations
766 individual students in the first exam. 353 of those students took part in the second exam. We use all observations from both exams. The effective sample size in our estimations is, however, much larger, as we contruct a large number of counterfactual neighbors pairs.
Sample size (or number of clusters) by treatment arms
First exam: 333 students in control, 208 in commitment, and 225 in monitoring.
Second exam: 204 students in control, and 149 in commitment
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB
INSTITUTIONAL REVIEW BOARDS (IRBs)
IRB Name
IRB Approval Date
IRB Approval Number
Post-Trial
Post Trial Information
Study Withdrawal
Intervention
Is the intervention completed?
No
Is data collection complete?
Data Publication
Data Publication
Is public data available?
No
Program Files
Program Files
Reports and Papers
Preliminary Reports
Relevant Papers