Online tests you can actually trust.

Security starts with the questions

Settings help, but the strongest protection is question design. Questions that ask participants to apply an idea to a new situation are much harder to look up than recall questions. Writing your own scenarios and numbers, instead of copying well-known textbook questions, makes answers far less likely to exist online.

Changing numbers between versions of a calculation question is a simple way to make shared answers useless.

Know who is taking the test

Many integrity problems are really identity problems: duplicate entries, fake names, or someone else taking the test. Requiring email, phone, or address before starting, and turning on OTP verification so each participant confirms their email with a six-digit code, ties every submission to a verified contact.

For formal tests, turning off multiple attempts means a participant who already finished is told the quiz is complete instead of being let back in.

Use time as a control

A fair but firm time limit leaves little room for searching or messaging for answers. A whole-quiz countdown suits most tests; a per-question countdown adds pressure against looking things up and suits recall-heavy quizzes. Set limits by taking the test yourself and allowing two to three times your own time.

Make every attempt look different

Shuffling questions gives each participant a different order, which makes answer-sharing during a test much harder. Building larger question sets and rotating which questions appear across different sittings or versions extends this further, so a list of answers from one participant is of little use to another.

Before the test

Tell participants the rules in writing: the time limit, the number of attempts, that verification is required, and what happens if rules are broken. Run a short practice quiz with the same settings so nobody is surprised on the day. Participants who understand the rules and feel prepared are less tempted to cheat.

After the test

Review results before acting on them. The quiz detail view shows each participant's score and time taken. Very short completion times with high scores, or several participants with identical unusual answers, may be worth a closer look. Treat these as signals for a conversation, not proof of cheating.

When to add supervision

For the highest-stakes assessments, online settings work best alongside some supervision, such as a proctored room or a live video call. Another option is a paper version: any quiz can be exported as an offline exam paper for participants to complete in person, while the same questions are used online for practice.

Being honest about limits

No online test can be made completely cheat-proof, and it is better to say so than to promise perfection. The realistic goal is to make cheating difficult, easy to spot, and not worth the risk, while keeping the experience fair for honest participants. Good questions, sensible settings, verified identity, and clear rules together achieve that.

Example: a secure recruitment test

A company screening applicants builds a twenty-question skills test with scenario questions written from real tasks, not textbook definitions. Applicants must enter their name, email, and phone, verify their email with a one-time code, and complete the test in one attempt within thirty minutes, with questions shuffled.

The hiring team reviews scores alongside time taken, and invites the strongest candidates to a supervised second round, where a short follow-up discussion confirms the online result.

Security settings at a glance

For identity: require contact fields and turn on OTP verification. For attempts: turn off multiple attempts on anything graded. For timing: use a whole-quiz timer, or a per-question timer for recall-heavy tests, with automatic submission when time runs out. For order: shuffle questions.

Combine these with application-based questions and clear written rules, and you cover the most common ways online tests are compromised.