How to prevent cheating in online exams and mock tests
How students cheat in online tests, which controls suit which stakes, the rules to publish, and the warning signs to check in results before you publish ranks.
On this page 9 sections
To prevent cheating in online exams, start with how students actually cheat: a second phone, answer-sharing groups, questions passed from an early slot to a later one, someone else taking the test, and screenshots posted while the test is live. Then set controls to match what each test is worth, publish clear rules before the test, and check results for warning signs before you publish ranks. This guide covers each step for coaching institutes running mock tests and scholarship exams.
How students cheat in online exams
In a hall, an invigilator watches the room. Online, the student controls the room and the device, so the methods change:
| Method | What it looks like | What to watch for |
|---|---|---|
| A second device | A phone beside the laptop, used to search or message during the test | Scores far above a student's usual level, especially on questions that can be looked up |
| Answer-sharing groups | A WhatsApp or Telegram group comparing answers while an all-India mock is live | Near-identical answer sheets, including the same wrong answers |
| Slot-to-slot leaks | Students in an early slot passing questions to friends in a later slot | Later slots scoring noticeably higher on the same paper |
| Impersonation | A stronger student, or a paid "solver", takes the test on someone else's account | A result out of line with the student's history and classwork |
| Screen capture | Questions photographed or recorded and posted while the test is live | Your questions appearing in groups or channels during the test window |
| Pre-test leaks | The paper or answer key reaches students before the test | Many students answering the hardest questions correctly and quickly |
No online check catches everything. Nothing on the test device can see a second phone on the desk, which is why the review after the test matters as much as the controls during it. If the paper itself may have leaked, see our guide to preventing question paper leaks.
Match controls to the stakes
Not every test deserves the same response. A daily practice quiz that only helps the student learn needs light controls. A scholarship test that decides fee waivers needs strong ones.
| Test type | Stakes | Sensible controls |
|---|---|---|
| Daily practice quiz | Low | A different question order for each student; no monitoring |
| Weekly or sectional test | Medium | Different question sets of equal difficulty, section timers, alerts when a student leaves the test screen |
| All-India mock with ranks | Medium to high | A fixed start window, account-sharing prevention, a review of results before ranks are published |
| Scholarship or admission test | High | All of the above, plus identity checks and proctoring, or an in-centre round for the shortlist |
Two rules apply at every level. Release answer keys and solutions only after the test has closed for everyone, including the last slot. And keep timing realistic: limits tighter than the real exam's just punish honest students on slow connections.
Set the rules before the test
- Publish the rules: the test window, what counts as unfair means, what is monitored and what happens when rules are broken.
- Ask for a commitment. Have students confirm, before they start, that they will attempt the test alone, without other devices or help.
- Say that results are reviewed before ranks are published, and that any flagged result is discussed with the student first.
- Tell parents of under-18 students, especially if tests are proctored.
- Explain why it matters. Copying in a practice test only hides a student's weak topics from them and from you.
Shared logins make impersonation easy, so treat account sharing as a cheating risk too. See how to stop account sharing.
Warning signs to check before you publish ranks
Many kinds of cheating leave traces in the results. Before ranks go out, look for:
- near-identical answer sheets, especially ones that share the same wrong answers;
- hard questions answered correctly in implausibly little time, again and again;
- scores far above a student's own history or classwork;
- one batch, centre or slot scoring well above the others on the same paper;
- your questions or answers appearing in student groups or channels during the test window.
A warning sign is a reason to look closer, not proof. Hold ranks for flagged accounts, review them, and talk to the student before any penalty. Apply the same policy to everyone, including toppers, and keep a note of each decision.
If you use proctoring, watch for false flags
Webcam proctoring, live or recorded, can suit high-stakes tests, but its automated flags are prompts for review, not verdicts. A peer-reviewed study of automated proctoring software found that students with darker skin tones were flagged for review more often. Everyday Indian conditions add their own false flags: poor lighting, shared rooms, family members walking past, weak 4G.
- Have a person review every flag, and let the student explain before any penalty.
- Tell students in advance what is recorded, why, who sees it and when it is deleted. Recordings are personal data.
- For under-18 students, proctoring is also a form of monitoring a child, so read our guide to verifiable parental consent under DPDP before you roll it out.
Design tests where cheating matters less
- Ask application questions rather than pure recall: multi-step numericals, data interpretation and case-based questions are hard to look up in the time available.
- Refresh your question bank every cycle, and retire any question that has leaked.
- Make practice about learning. Show students analysis they will want, such as accuracy by topic and time spent per question, so an honest attempt is worth more to them than a copied score.
- Reserve verified ranks for supervised tests. If a rank will appear in your marketing or decide a scholarship, run that test under stronger controls or add an in-centre round for the shortlist.
For the exam-day side, such as capacity and support when thousands start at once, see our guide to running an online exam for thousands.
Key takeaways
- Start from how students cheat: second devices, answer-sharing groups, slot-to-slot leaks, impersonation and screen capture.
- Match controls to what each test is worth; practice quizzes need little, scholarship tests need a lot.
- Publish the rules, including what is monitored, before the test.
- Check results for warning signs before publishing ranks, and treat each flag as a question, not proof.
- If you proctor, have a person review every flag, and handle recordings, especially of minors, with care.
Where Upclass fits
For mock tests in your own app, see our test series software. For courses, Upclass gives institutes tests, live classes and branded apps, and VidSafe adds screen- and camera-recording detection and account-sharing prevention. See the LMS for coaching institutes page.
Frequently asked questions
How do online exams prevent cheating?
Through a mix of rules, test design and review rather than any single tool. Institutes publish rules before the test, give students different question orders or sets, use fixed start windows and section timers, and prevent account sharing. High-stakes tests may add proctoring or an in-centre round. Afterwards, results are checked for warning signs, such as identical wrong answers, and a person reviews each flag before any action.
How to prevent students from cheating in online exams?
Explain the rules and what is monitored before the test, and tell students that results are reviewed before ranks are published. Use a fixed window and section timers, release solutions only after everyone has finished, and check results for warning signs such as identical wrong answers or sudden jumps in score. For practice tests, remind students that copying only hides their weak areas.
How to conduct online exam without cheating?
Plan it in three stages. Before: build a fair paper, test the app on low-end phones, and publish the rules and what is monitored. During: open a fixed start window, keep section timers, prevent account sharing and, for high stakes, proctor or supervise. After: check results for warning signs, review each flag with the student, then publish ranks and solutions.