What Is Academic Collusion? Where the Line Is and What the Data Shows
Collusion is the form of academic misconduct hardest to define cleanly. The line between encouraged collaboration and prohibited collusion depends on instructor policy — and shifts dramatically between courses. This guide covers what counts and what the data shows.
TL;DR
Academic collusion is unauthorised collaboration on individual assessments. The AMI's D5 dimension measures it. Nigeria (75), India (72), Malaysia (72), Indonesia (72), South Korea (70), Turkey (69), Iran (69) lead globally. The line between collaboration and collusion depends on instructor policy.
TL;DR
Academic collusion is unauthorised collaboration on work meant to be done individually. The AMI's D5 dimension scores it. Nigeria (75), India (72), Malaysia (72), Indonesia (72) lead globally. The boundary between permitted collaboration and prohibited collusion depends on instructor policy and is often genuinely unclear.
Definition
Academic collusion is unauthorised collaboration on academic work that is intended to be assessed as the individual work of one student. The misconduct is the misrepresentation that the submitted work reflects only the named student's effort when it actually involved one or more others.
The "unauthorised" qualifier is critical. Many forms of collaboration are explicitly permitted or even required — group projects, peer review, study groups. The misconduct emerges when individually-assessed work has been produced jointly without disclosure.
Common forms
Identical or near-identical submissions
The clearest form. Two or more students submit work with substantial overlapping content (calculations, paragraphs, code).
Shared problem-set solutions
Students work through problem sets together and arrive at substantively identical solutions. Common in STEM disciplines where unique algorithmic solutions are limited.
Joint essay writing
Students collaborate on essay structure, argument, and content while submitting as individuals. The most common form in humanities and social sciences.
Code copying
Students share programming assignments — copying full solutions or substantial sections. Code-similarity detection tools (Moss, JPlag) are widely deployed in CS courses but not universally.
Exam answer comparison
Students sharing exam answers in real time (texting during exams, sharing screens in online exams). Crosses into a more serious offence than asynchronous collusion.
The collaboration-collusion boundary
The most genuinely difficult area of academic misconduct policy. The boundary depends on:
Instructor and institutional policy
Some instructors permit substantial discussion of problem sets; others require independent work. Some explicitly forbid showing one's work to others; others encourage peer learning.
Assignment type
Group projects authorise full collaboration. Take-home exams typically prohibit it. Routine problem sets sit in between, with policy varying by instructor.
Course culture
Some disciplines have strong study-group traditions. STEM courses at major research universities often involve heavy informal collaboration that may or may not be formally permitted.
Cultural norms
Educational systems vary in how they treat collaboration. East Asian study-group cultures (스터디 그룹 in Korea, 勉強会 in Japan) treat intensive group study as the norm; the line to prohibited collusion can be culturally ambiguous.
What the AMI data shows
D5 scores on a 0–100 scale across the 39-country set:
| Top D5 scores | Score |
|---|---|
| Nigeria | 75 |
| India | 72 |
| Malaysia | 72 |
| Indonesia | 72 |
| South Korea | 70 |
| Turkey | 69 |
| Iran | 69 |
| United States | 68 |
| Brazil | 65 |
| Philippines | 65 |
| Lowest D5 scores | Score |
|---|---|
| Singapore | 50 |
| Argentina | 50 |
| Colombia | 52 |
| Germany | 52 |
| Thailand | 55 |
| Australia | 55 |
The highest D5 scores cluster in countries with:
- Large class sizes at major institutions (Nigeria, India, Indonesia, Malaysia)
- Strong group-study cultures (Korea, Turkey)
- Limited assessment design distinguishing individual from collective work
The United States' D5 score of 68 is notable — high for a Q1 country. McCabe survey data has consistently shown elevated US collusion rates in STEM coursework with group-problem-set cultures.
Detection methods
Submission similarity
Plagiarism detection tools used internally — comparing student submissions against each other within the institutional repository. Effective for detecting verbatim or near-verbatim overlap.
Code-similarity tools
Moss (Stanford), JPlag, and similar tools detect structural similarity in code submissions. Widely used in CS courses.
Statistical analysis
Comparing answer patterns across students can detect collusion even when wording differs. Less common; requires careful interpretation given that students working similar problems will produce similar work even without collusion.
Real-time exam monitoring
Online proctoring software can detect screen sharing and texting during online exams. Effective for synchronous collusion in exam settings.
Why D5 scores vary
Survey data availability
The McCabe / ICAI surveys captured collusion data for some countries with strong direct measurement. Regional extrapolation for others compresses variance.
Cultural variation in study practice
Group-study practices that produce elevated D5 are not universal. Nordic and German university cultures historically have less group-study emphasis; Asian and Latin American cultures more.
Assessment design
Universities with assessment designs that distinguish individual from collaborative work (oral examinations, individual project defences, varied question sets) have lower effective D5 even with similar underlying student behaviour.
Sources
- McCabe, D. L. (ICAI / Rutgers) collusion survey data
- Moss, JPlag code-similarity documentation
- AMI v1.5 methodology document
- Country-specific integrity literature
Full methodology | Download dataset
Related
Frequently asked questions
What is academic collusion?
Academic collusion is unauthorised collaboration on work intended to be assessed individually. The 'unauthorised' part is critical — collaboration is often permitted or encouraged, but specific assessments require individual work. Submitting jointly-produced work as your own individual work crosses into collusion.
Where is the line between collaboration and collusion?
The line is set by instructor and institutional policy. Discussing concepts with peers is usually permitted; producing identical or near-identical answers is usually collusion. The boundary is often unclear when assignments are completed in study groups — students may discuss approach, methodology, and even partial solutions while still being expected to submit independent final work.
Which countries have the highest academic collusion rates?
On the AMI's D5 dimension, Nigeria (75) scores highest, followed by India (72), Malaysia (72), and Indonesia (72). South Korea (70), Turkey (69), and Iran (69) also score high. The lowest scores are in Singapore (50), Argentina (50), and Colombia (52). Group-study cultures and large class sizes drive elevated scores.
How to cite this article
APA: Booth, F. (2026). What Is Academic Collusion? Where the Line Is and What the Data Shows. Academic Misconduct Index. https://academicmisconductindex.com/blog/what-is-collusion-academic
BibTeX: @misc{booth2026what, author={Booth, Francisco}, title={What Is Academic Collusion? Where the Line Is and What the Data Shows}, year={2026}, url={https://academicmisconductindex.com/blog/what-is-collusion-academic}}
Francisco Booth
Independent researcher, founder of the Academic Misconduct Index
Related posts