A School of Continuous Improvement research poster restages a shopping-centre car park's clearance-rate threshold at twelve primary-school pickup zones and finds the relationship it predicted does not reproduce, while permit renewal keeps applying the threshold regardless.
A preregistered replication across twelve primary-school car parks finds no relationship, and permit renewal keeps scoring every one against the retail threshold anyway
A threshold built to flag an overflowing shopping-centre car park has not travelled well to the school gate. A new research poster from the School of Continuous Improvement reports a preregistered replication of the retail clearance-rate relationship at twelve primary-school pickup car parks, and finds the threshold that predicted overflow at the mall predicts nothing at all once the same design is restaged at the school run.
Led by Dr Renke Sabel, Senior Lecturer and Convenor of the Indicator Commons, with Dr Marit Osayande and Dr Bram Ntuli, the study locked its coding rules and analysis plan before the school term began, matching boom-gate exit timestamps against verge-side dashcam counts of vehicles queued onto the street across a full pickup term. Where the original retail relationship ran strongly (r = 0.71), the replication returns nothing of the kind (r = 0.07), and every one of the twelve car parks still clears the compliance threshold at least once a week regardless.
An indicator that predicts nothing at the school gate is still, technically, an indicator. Nobody has yet told the compliance office that distinction matters.
— Dr Renke Sabel, Senior Lecturer and Convenor, Indicator Commons
Permit renewal has not paused to take account of the replication. “The threshold was built for a different queue, on a different clock, and it is still the number every one of our twelve car parks is renewed against,” said Dr Marit Osayande, Lecturer and Convenor of the Living Dashboard. “A site-specific figure is now before the Commons, for whenever the next intake round gets to it.”
The University counts a failed replication of one of its own findings among the more useful things a research poster can do this term, not least because so few of its indicators are ever tested against a setting they weren’t built for. The full poster is available from the University’s research repository under an open licence, doi:10.5555/slop.7v853y.