Case: The Two-Control Refrigerator

Norman's two-compartment fridge had two controls labeled "freezer" and "refrigerator." It was nearly impossible to set the temperatures.

Why: The controls suggest a false conceptual model (Fig 1.10A) — two compartments + two controls implies each control independently sets the temperature of its named compartment (each with its own sensor and cooling unit). Wrong. In reality (Fig 1.10B): there is one thermostat and one cooling unit; one control sets the thermostat, the other sets the proportion of cold air sent to each compartment — which is why the controls interact. The sensor's location is unknowable.

Compounding factors:

• The The System Image (controls + manual) communicates an incorrect model.
• No immediate feedback: the manual says allow 24 hours for temperatures to stabilize → you'd need a lab notebook and controlled experiments to set a fridge.

Lesson: Even when Norman knew the correct model, he still couldn't set it, because the design hid which control did what. A wrong/false conceptual model is worse than a complex-but-true one. Decades of reader thank-you letters; zero response from GE, which still ships the same design.

Confidence

1.0 — detailed worked example.

Sources: ["DOET ch.1 'Conceptual Models' p.28-31 (Fig 1.9-1.10)"]