Lockdown browsers stop the exam device, and nothing else in the room

What the lock actually covers

On the exam device, a lockdown browser is genuinely useful. It closes the casual routes: switching to a search tab, opening a stored document, messaging a friend from the same machine, pasting prepared text. For a low stakes quiz that may be most of the risk, and the tool earns its place honestly there. The failure begins when a receipt for that narrow job gets filed as an integrity strategy.

The limits are architectural, not a settings problem. The browser cannot see the phone face down beside the keyboard, the second laptop outside the camera frame, the earpiece, or the person mouthing answers from the doorway. No update will fix that, because the tool's jurisdiction ends at the edge of one screen. Honorlock's own lockdown explainer concedes the same point: browser locks cannot detect phones or other secondary devices.

Identity is a separate problem with a separate tool

A lockdown browser also cannot tell you who is sitting at the keyboard. It is entirely possible to run a perfectly locked exam taken by the wrong person, every rule obeyed on screen, the result worthless. Identity verification lives in a different layer: a photo ID shown to a webcam at startup, matched against the enrolment record, with a short environment scan that shows the room the exam will happen in.

Respondus Monitor's startup sequence, described in university guidance, bundles those steps: ID to camera, environment check, then recording with flags for review. The important design choice is that flags go to a human. Automated suspicion is a triage tool. The moment a flag alone fails a student, your integrity system has become an accusation machine with an error rate nobody published.

Assessment design closes what software cannot

Question design shrinks the value of a second device. Timed sections with randomised order and per student variants make looked up answers arrive too late or fit the wrong paper. Application questions, where the student works a scenario rather than recalling a definition, are slower to outsource to a search box or a chatbot, because the answer has to be produced in the student's own chain of reasoning under time pressure.

None of that is glamorous, and departments resist it because writing variant banks is real work. But compare the costs honestly: a proctoring stack strong enough to watch a whole room is intrusive, expensive, and still beatable, while a well built question bank keeps working in every delivery mode, including the in person hall where no browser locks anything. Integrity bought through design does not expire when the software contract does.

A sensible policy tiers the controls. Low stakes formative quizzes: lockdown browser alone, honestly labelled as a deterrent. Mid stakes: add ID check and recorded session review on flags. High stakes certification: live or blended proctoring, environment scan, and an appeals route with a human at the end of it. Writing the tiers down stops the two common failures, under protecting the exam that gates a licence, and over surveilling the quiz that was meant to teach.

Match the stack to the stakes, in writing

Tell students what each layer does and does not do. A cohort that understands the lockdown browser guards the device, and that identity and room checks guard the rest, games the system less and appeals more fairly when something goes wrong. Secrecy about controls mostly protects the vendor's marketing, not the exam.

Frequently asked questions

Keep reading: Question bank leaks: your exam is on a homework site, now what, Identity checks before the exam: the impostor is stopped at the door or not at all.