A new Slop University paper recovers the attempt-level log behind a junior sports association's volunteer umpire accreditation record and re-scores it four ways, finding accreditation rates between 39.6 and 98.6 per cent in the same data.
Re-scoring five seasons of a junior sports association's accreditation log under four aggregation rules moves the published figure by 59 percentage points
A junior sports association’s volunteer portal keeps one row for each umpire each season, and that row holds a single score. The accreditation module behind it allows unlimited attempts, so for many volunteers there were a dozen scores available and no record of anyone deciding which one to keep.
One Row, Four Rates, published today by the School of Continuous Improvement, recovers the attempt-level log sitting behind five seasons of one metropolitan association’s accreditation record — 4,318 volunteer-seasons across 61 clubs — and re-scores all 18,704 attempts four different ways. The association’s published figure of 98.6 per cent keeps each volunteer’s best attempt. Counting only first attempts, the same records give 39.6 per cent.
The people who kept these records were volunteers giving up their Saturdays, and we owe them the care of reading properly what they wrote down. Most of what this paper is, in the end, is a thank-you with some arithmetic attached.
— Dr Torun Ezeigwe, Senior Lecturer and Director, Master of Applied Measurement
The work sits inside the School’s evaluation-ecosystems agenda, which has been building capability around the records institutions already hold rather than the ones they might commission. The University counts the paper among the more instructive outputs of that agenda this year, and the team describes it as an exercise in reading an administrative surface back to the source it came from.
“The questions coming back to us are already better than the ones we went in with, and we would rather have them than a tidier finding,” said Dr Marit Osayande, Lecturer and Convenor of the Living Dashboard.
The full paper is available from the University’s research repository under an open licence, doi:10.5555/slop.5b349c.