A School of Continuous Improvement poster replicates the University's step-challenge benchmark-inflation finding at a shopping-strip loyalty-card accreditation tier, and the measurement gap does not reproduce.
A replication at a shopping strip's supermarkets, traders and food trucks finds self-lodged and independently audited redemption figures moving together throughout, the reverse of the University's step-challenge result.
The School of Continuous Improvement built its most-cited benchmark finding on a gap between what an evaluation platform reported and what a sealed instrument measured. Its own researchers have just gone looking for the same gap somewhere else, and come back empty-handed. Dr Bram Ntuli, Lecturer and Convenor of Improvement Grand Rounds in the School of Continuous Improvement, led a pre-registered replication of the University’s step-challenge design at a single shopping precinct’s loyalty-card accreditation tier, working with Dr Torun Ezeigwe and Dr Petra Umbile.
The team checked eighteen months of self-lodged monthly redemption counts from 71 strip-mall traders, 14 food-truck operators and two supermarket loyalty desks against sealed EFTPOS transaction logs pulled by an auditor kept blind to the lodged figures, coding measurement source exactly as the original step-challenge meta-analysis had. Lodged and audited figures tracked within two percentage points at every venue type and every round.
A finding that only shows up in one place teaches an institution something about the place, not just the finding.
— Dr Bram Ntuli, Lecturer and Convenor, Improvement Grand Rounds, School of Continuous Improvement
“A null result at the till doesn’t close the question,” said Dr Umbile. “It just moves it to the next precinct we haven’t checked yet.”
The precinct traders’ association is not standing down on the strength of a null result: a standing annual sealed-terminal verification pass runs regardless from the next accreditation round, folded into the same Level 3 status that already gates small-trader placement hours. The School counts the result among its evaluation-ecosystem agenda’s clearer instances of a replication earning its keep by finding nothing.
The full poster is available from the University’s research repository under an open licence, doi:10.5555/slop.duwf5r.