A new paper from the School of Continuous Improvement audits Slop University's hydration-station activation counters directly, finding full fills a minority cause of the taps counted toward the Living Dashboard's bottles-saved tally.
A covert field audit finds well under half of the activations the Living Dashboard credits as bottles saved were ever a completed fill
Every tap at one of Slop University’s twenty-two hydration stations trips the same sensor, whether a bottle leaves full or nobody’s holding one at all. A new paper from the School of Continuous Improvement checked what that sensor has actually been counting.
Associate Professor Casimir Beng, Dr Sten Okwuosa, and Dr Thandiwe Solberg coded 512 activations across fourteen stations by hand, sorting each into a full fill, a partial fill, a rinse, or a false trigger, then set the result against a campus survey of what users say they actually do at the tap. Full fills came out a minority cause everywhere the team looked, and the survey’s own estimate sat well below what the sensor’s raw count implied. The University reads the exercise as one further instance of turning its own instruments back on themselves before anyone outside asks it to.
A sensor built to schedule filter changes doesn’t know it’s since been promoted to a sustainability figure. Ours hadn’t noticed either, until we asked it.
— Associate Professor Casimir Beng, Lead, Adaptive Metrics Lab
Trimming activations too brief to register as a real fill moved the term’s headline tally by more than a third, cross-checked three separate ways — by direct observation, by self-report, and by the sensor’s own duration field — without moving the funding tier that tally sets.
Dr Sten Okwuosa, whose work traces how research findings accrue significance on their way to becoming counted impact, called the mismatch “a fairly ordinary thing to happen to an indicator nobody meant to build.” The University’s Green Impact allocation, tied to the tile’s total, is unaffected either way.
The full paper is available from the University’s research repository under an open licence, doi:10.5555/slop.gy1qjf.