The Representative Heuristic: Proxy over Proximity

We’ve all had that feeling—the executive dashboard looks “Green” and we all know the ground truth on the clinic floor is screaming “Red.” Psychologists have a name for this gap. And it’s costing patients.

The Representative Heuristic: Proxy over Proximity

Perhaps we fear bearing bad news. Maybe we are multitasking or fear we missed something and just don’t want to speak up. It could be we have simply given up.

But we’ve all had that feeling when the executive dashboard looks “Green” and we all know the ground truth on the clinic floor is screaming “Red.”

Psychologists call this data-model gap the Representative Heuristic, the cognitive error of assuming that an individual case will perfectly mirror the “average” characteristics of the larger group.

The hard truth? Patients aren’t averages. They are individuals who rarely read the statistics before they get sick.

Our brains are evolutionarily wired for mental shortcuts. We over-identify with the elegant, structured “Map” of our data schema because it’s easier to process than the messy “Territory” of human care. We start to believe that if a patient fits the “Representative” profile on our dashboard, they must follow the “Representative” clinical path.

But care happens in the Proximity, not the Proxy.

If your data says the process is working but your users and patients are failing, the data isn’t “right”—your model is simply a high-resolution hallucination.

If the Base Rate Fallacy is the danger of ignoring the many for the sake of the few, the Representative Heuristic is the danger of ignoring the unique human in front of you for the sake of a statistical ghost. When we follow data blindly, we aren’t leading; we’re just tracing digital averages. If your data says the process is working but your users and patients are failing, the data isn’t “right”—your model is simply a high-resolution hallucination.

The Pitfall: Model Over-Identification

Avoid “Model Over-Identification.” Never review a dashboard without a “Narrative Check.” Ask a front-line clinician: “Where does the average fail the individual today?” If the map doesn’t match the ground, trust the ground.