respondus proctoring online exam

Does Respondus Detect Cheating? How Its Flagging Works

6 min read

Respondus does not decide that you cheated. What it does is record your exam session and flag behaviours that a human then reviews. That distinction matters enormously: a flag is not a verdict, and the system is designed to be over-sensitive rather than to make judgments.

Respondus has two parts, and they do different jobs. LockDown Browser prevents you from leaving the exam window. Respondus Monitor uses your webcam and microphone to record the session and generates automated flags from what it observes.

What Respondus Monitor actually flags

The most common automated flags are behavioural rather than conclusive:

  • No face detected — you moved out of frame, or poor lighting made you hard to see
  • Multiple faces detected — another person entered the room
  • Frequent eye movement or looking away from the screen — the classic "reading notes off-camera" signal, and also what happens when you are thinking hard
  • Talking or background voices detected in the audio
  • A second person entering the frame
  • Room-scan irregularities — if the required 360-degree scan was incomplete
  • Application or tab switching — detected by LockDown Browser
  • Detection confidence dropping — for example, if you lean back, put a hand over your face, or the camera angle changes

Notice that almost every one of these has a perfectly innocent cause. That is the core characteristic of the system: it produces evidence for review, not findings.

How the instructor decides

The flags appear on a timeline alongside the recorded video. The proctor watches the moments around each flag and decides whether the behaviour is explained. A single "no face detected" flag while you lean back to think is rarely treated as anything. A cluster of talking, second-face, and off-screen eye-gaze flags at the same moment is what attracts attention.

That means your protection is the context around the flag, not the absence of flags.

How to avoid false positives

These are practical, not about avoiding detection:

  • Run the webcam check properly and complete the room scan carefully. Incomplete setup checks cause more suspicion than anything else.
  • Sit still enough to stay in frame. Leaning out of view is the most common false flag.
  • Light your face, not the background. A window behind you makes detection unreliable.
  • Use a wired or solid connection so video does not drop out and register as a gap.
  • Do not talk to yourself or read questions aloud — audio flags are easy to generate accidentally.
  • Close everything else beforehand, so you never need another application.
  • Clear the desk and the room before the scan, including notes you would never use.
  • Warn anyone in the house. A passer-by triggers a second-face flag exactly as a collaborator would.

If you are flagged wrongly

Say so promptly and calmly, and explain what was happening at the timestamp. Ask to review the specific moment. Instructors can and do dismiss flags — but a vague denial is less persuasive than a specific explanation tied to a point in the timeline. Check your institution's policy on academic-integrity hearings before assuming the flag is a foregone conclusion.

For organisations running proctored assessments

Flag-based systems are deliberately noisy, which means the reviewer's time is the real cost and false positives are a genuine operational burden rather than an edge case. If you are evaluating candidate screening at volume, the same trade-off applies in hiring: aggressive flagging on an assessment shifts work onto whoever has to interpret the flags, and discourages honest candidates who simply dislike being surveilled. Deciding what signal you actually need before turning monitoring on is the useful question. AI candidate evaluation covers how to design assessment and scoring so the review effort goes where it matters.

Always disclose that recording occurs. Silent monitoring is both a trust problem and, in some jurisdictions, a legal one.

Bottom line

Respondus does not detect cheating — it records your exam and flags behaviours like leaving the frame, eye movement, voices, or app switching, which a human then reviews. Most flags have innocent explanations, so run the webcam check properly, stay in frame, light your face, close other applications, and warn anyone at home. If you are flagged wrongly, explain the specific timestamp rather than denying generally.