Sandra had given six thousand road tests. She could tell you the pass rate for every month since 2009, the most common point deductions, the intersection on Route 4 where seventeen candidates in a row had forgotten to signal.
What she couldn’t tell you — what the clipboard didn’t have a box for — was the thing she actually knew.
It arrived in the first thirty seconds. Before any scored maneuver. The candidate would adjust the mirror, put the car in gear, check over their shoulder. And Sandra would know. Not whether they’d pass. That was the test’s job, and the test was good at it. What she knew was whether they’d be okay. Whether three years from now, on a wet road in the dark with the radio on and a child asking something from the back seat, they’d make the right decision without deciding.
She’d tried to explain it once, at a department meeting. Her supervisor had asked, reasonably, what metric she was describing.
She wasn’t describing a metric.
The closest she’d come was this: some candidates drove as if the car ended at the bumper. Others drove as if they extended into the road — as if the space between them and the next vehicle was part of their body and they could feel it stretch and compress. The second group didn’t score higher. They made the same mirror checks, the same signals. On paper, no difference.
The difference was in what they were paying attention to when they weren’t paying attention to anything.
She’d had a candidate last Tuesday. Young woman, nervous. Scored 88 — solid pass, two minor deductions for lane positioning. But during the highway merge, when the truck in the adjacent lane drifted slightly, Sandra had watched the candidate’s hands. They’d adjusted before the truck moved. Not a correction. A conversation — the candidate’s body already talking to the traffic, reading the weight shift in the truck’s suspension, the slight tire wobble that said the driver was checking a phone.
None of that was on the form. The form said “highway merge: satisfactory.”
Six thousand tests. She’d filled out every box. The boxes were accurate. The pass rates predicted accident rates with reasonable fidelity. The system worked.
And Sandra knew something the system didn’t, and couldn’t, because the moment you put a box around it, it would become a thing candidates practiced for the test instead of a thing they carried in their hands. The measurement would destroy the knowledge. Not because measurement was bad. Because this particular knowledge existed only in the space the clipboard couldn’t reach.
Her supervisor hadn’t understood. He’d smiled patiently and asked her to email her suggestions for form revision.
She hadn’t emailed anything. There was nothing to revise. The form was fine. The form had always been fine. That was the whole point.