Model-in-the-Loop Engineering · Capstone brief

ENG-09 · Human factors in the loop

When you design the operator interface for a model-assisted decision, do operators over-trust the score — and can you measure that with a short, ethical user test?

The question

When you design the operator interface for a model-assisted decision, do operators over-trust the score — and can you measure that with a short, ethical user test?

System / materials

Teacher-approved UI mock or sim. Test uses classmates as operators only with consent and no deception about risk. No real machinery motion without interlocks.

Expected failure modes

Leading questions. Blaming operators instead of design. Testing on dangerous equipment.

Done looks like

UI sketch, test protocol, trust/over-trust observations, and one design change grounded in results.

Five C's

CT: measuring trust, not likability. CR: ethical protocol. CO: peer operator run. CM: before/after UI notes. CZ: consent and hazard boundaries.

Mentor role

Human-factors or safety engineer reviews protocol at Checkpoint 1 and findings at Checkpoint 2. School-supervised.

Rubric calibration

R1: scoped UI, consent. R2: protocol reproducible. R3: baseline UI tested. R4: over-trust case shown. R5: design change linked to data. R6: no deceptive risk.

Two ways this goes wrong

(a) “Users liked it.” (b) Real machine motion with untrained operators.

Credit lane fit

Lane A immediately. No verified credit claim.