# A rostering app's fairness score won't agree with itself

> A new Slop University paper resubmits an unchanged casual-shift roster to a rostering platform's fairness-compliance engine four times over twelve weeks and finds the score rarely agrees with itself.

A fairness-compliance score is meant to be the last word on whether a
week's roster treated its casual staff evenly. A new Slop University paper
resubmits the exact same roster to the scoring engine four times over
twelve weeks and finds the last word keeps changing.

For the School of Continuous Improvement, the result lands less as a
surprise than as confirmation of a suspicion the School has been circling
for a while: that a compliance score's authority and its accuracy are not
the same achievement.

> We've learned, again, that the School is better at building instruments
> that satisfy an audit than instruments that survive being asked twice.
>
> --- Senior Lecturer Renke Sabel, Convenor of the Indicator Commons, School
> of Continuous Improvement

"Once a manager has a number, the roster underneath it stops being anyone's
job to reread," said Lecturer Marit Osayande, Convenor of the Living
Dashboard. "What we don't yet know is what would make someone look again."

The team worked with a regional retail-and-hospitality group operating
three dozen stores, taking one week's already-finalised roster and putting
it back through the rostering platform's compliance engine at intervals
across a school term, before asking the managers who had received each
week's verdict what they had done with it.

The full paper is available from the University's research repository
under an open licence, doi:10.5555/slop.o8swbm.

Source: <https://slop.university/news/2026-07-20-wont-agree-with-itself/>\
Licence: [CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)\
Cite as: Swift, B. (2026). _Slop University_ [Artwork]. https://doi.org/10.5281/zenodo.21199776
