Impersonation at test centres: the ID check is the exam before the exam

The document is only half the check

A convincing document can be genuine: borrowed from a lookalike, bought from the real candidate, or issued in the impersonators own name against a fraudulent application. Checking that the document is real therefore settles less than staff assume. The comparison that matters is document photograph to living face, taken slowly, in good light, with the candidate looking at the checker, not at a camera over their shoulder.

Train checkers on the comparison itself: ears, hairline, the proportions photographs preserve and disguises change. Give them explicit permission to take time and to escalate. An impersonator relies on queue pressure more than on document quality, because a busy desk at a full session start is a desk that waves people through. Stagger arrival times before blaming the checkers.

Capture the admission evidence every time

Photograph or scan at admission, under the programs published policy, and store the image with the session record. A retained admission photo ends most later disputes in one comparison, and its absence turns an answer similarity case into an argument about memory. The same applies to the signature or biometric the program collects: collected at admission, compared at return from breaks, because substitution mid session is a known variant, not a paranoid invention.

Where the program uses biometrics, publish what is collected, why, how long it is kept, and the alternative for candidates who cannot or will not use it. A biometric with no stated retention or no alternative route creates legal exposure out of proportion to its security value, and in several jurisdictions it creates the exposure whether or not it ever catches anyone.

The professional impersonator is a repeat customer: same face, different names, across centres and dates. No single desk can see that pattern. The program can, by comparing admission images across its own sessions under a clear policy and legal basis, and by sharing confirmed impersonation findings through the channels its rules allow. One confirmed repeat face usually explains a cluster of results that analytics had been circling for months.

Repeat faces are the pattern to hunt

Audit centres on admission quality the way you audit them on everything else: test the desk with programme issued scenarios, review a sample of admission records against session results, and count how many admissions lack a usable image. A centre with immaculate paperwork and no usable admission photos is telling you where its effort goes.

Refusal is a decision with consequences, so it needs a script, a second staff member, and a record written at the time. The candidate gets a factual statement: the document presented could not be matched to the presenter, the session fee position, and the review route. What the candidate does not get is an accusation of fraud at a public desk, which escalates the room and prejudges the review.

When the desk refuses admission

Review refusals centrally and quickly, because some will be wrong: changed appearance, poor document photos, medical reasons. A program that overturns its own wrong refusals fast and apologises cleanly loses nothing. A program that defends every refusal defends the wrong ones with the same energy as the right ones, and appeal bodies notice the pattern.

Keep reading: How do testing programs spot collusion when two candidates answer alike?, How should exam programs handle AI assistants a candidate can run on a second screen?.

Frequently asked questions