Skip to content
Sınaq
All guides

Assessment practice

Support exam integrity without overclaiming what signals prove

Exam integrity depends on a sensible assessment design and a fair review process. Shuffling and browser signals can support that process, but they cannot establish what a participant understood or prove misconduct on their own.

A Sınaq guide2 min read
Editorial illustration for: Support exam integrity without overclaiming what signals prove

State the rules and their purpose

Tell participants which materials, calculators, collaboration and assistance are permitted before they begin. Explain whether switching away from the exam is counted and how that information will be used. Choose rules that fit the task: an application exercise may deliberately allow reference material, while a recall check may not. A clear rule is easier to follow and review than an unstated expectation.

Design tasks that reward reasoning

Use scenarios that require applying a concept to specific facts, and refresh examples when reusing an exam. Plausible distractors can expose a misconception more usefully than trick wording. Shuffling questions or options may reduce casual copying but does not make shared content impossible. Check that options remain meaningful in any order and that explanations do not refer to a fixed letter.

Interpret signals in context

A participant may leave the tab because of a notification, an accessibility tool, an accidental gesture or an actual rule violation. A short completion time can reflect prior knowledge or guessing. Look for corroborating information and provide an opportunity to explain. Keep technical access problems separate from integrity concerns, and do not publicly label a person based on one automated event.

Use a proportionate follow-up

If a result needs verification, use a short oral explanation, a comparable supervised task or a discussion of the method, following your institution's rules. Record the evidence and decision consistently. Give participants a clear route to question an error. Collect only the monitoring information you need and restrict access to it; adding more surveillance does not automatically make an assessment more valid.

Before you publish

  • Allowed materials and collaboration rules are explicit.
  • Shuffled questions remain correct in every order.
  • A human review and challenge process exists for concerns.

Questions, answered

Does a tab switch prove cheating?
No. It is a browser event with several possible explanations. It can prompt a review when relevant rules were communicated, but requires context.
Can shuffling guarantee a fair exam?
No. It is one design option. Fairness also depends on clear tasks, appropriate timing, access, scoring and a consistent review process.

Continue reading