Research outputs

The published research outputs of Slop University.

All research outputs of the University are published open access under CC BY 4.0, with persistent identifiers minted by the Office of Research Outputs.

Published outputs

Poster

Loosened in Translation

Torun Ezeigwe, Ingeborg Nwachukwu, Solveig Adeyemi · July 2026 · doi:10.5555/slop.2jt2a2

A research poster from the School of Continuous Improvement back-translates the six community-language versions of a metropolitan council's community-hall hire conditions, coding all 186 clause-instances against the 31-clause source and finding 43 divergent, of which 30 loosen an obligation and nine drop it entirely. Bookings issued a version in which a clause had loosened or vanished are pulled up on that clause roughly three times as often, while hirer certainty about the conditions is flat across versions.

Ad

We Knew The Step Before You Did

July 2026 · doi:10.5555/slop.x8270x

A marketing poster for the School of Emergent Priorities compresses the published finding of "The Step You Can't Picture" --- that flat-pack assemblers hold the wordless diagram booklet only as long as they can still picture the next step, and reach for a video once they cannot --- into a single portrait signage headline, developed for the University's display panels in matched light and dark renderings.

Poster

Neat Enough to Be Green

Torun Ezeigwe, Solveig Adeyemi · July 2026 · doi:10.5555/slop.eft2py

A research poster from the School of Continuous Improvement puts 1,776 randomised load vignettes to 148 waste transfer station gate attendants across six councils, crossing what a trailer contains with how it is stacked and how many vehicles are waiting behind it. A stacked load carrying one non-green item is called green waste in 71.4% of judgements at a clear gate and 38.2% when the same composition arrives heaped with six vehicles waiting, with presentation and queue state together shifting the call by 38.2 percentage points against composition's 47.3.

Brochure

What Actually Ran

Torun Ezeigwe, Iben Chikere · July 2026 · doi:10.5555/slop.a3ugno

A case-for-support brochure opening on a community centre's term catalogue, whose evening classes are printed in full and then half of them do not run, and following that distance into the University's work on advertised offers, the records that outlive the rooms they describe, and the instruments the Indicator Commons lends to organisations outside the University. Three previously published studies anchor the publication, each cited by title and DOI.

Poster

Too Cold to Count

Ronja Oyelaran, Solveig Adeyemi, Torun Ezeigwe · July 2026 · doi:10.5555/slop.v9ye2i

A research poster from the School of Emergent Priorities audits 248 public barbecue units across 96 reserves without notice over 14 weeks, matching each of the 82 units found cold to two working units on visitation band and installation cohort. Conditional logistic regression returns an adjusted odds of being cold of 6.1 (95% CI 2.9-12.8) in the lowest recorded-usage quintile against the highest, on a counter that increments only when a unit completes a heat cycle; none of the 82 cold units appears on the renewal shortlist drawn from the top two quintiles.

Poster

Sealed, Then Re-Worn

Torun Ezeigwe, Ingeborg Nwachukwu, Ronja Oyelaran · July 2026 · doi:10.5555/slop.235f0f

A research poster from the School of Continuous Improvement follows 96 worn verge shortcuts across six wards for fourteen months, from the day each is sealed under a council path-formalisation warrant to the emergence of a successor worn line, coded against a definition fixed before the first survey. Successor lines appear at 71% of sealed sites within twelve months at a median of 18 weeks, and the wear score that qualified a route for sealing shows no detectable association with that timing (HR 1.04, 95% CI 0.71-1.52); the warrant's own 26-week qualification window exceeds the median recurrence interval it is meant to resolve.

Paper

Visible First

Verity Marris, Torun Ezeigwe, Iben Chikere · July 2026 · doi:10.5555/slop.6jcj2m

A paper from the School of Emergent Priorities describes an eighteen-week deployment of a dispatch service that ranks each reported stray supermarket trolley by an estimate of how readily it is seen from the arterial road network, discounted by estimated retrieval cost, across a twenty-one-store retail network handling 9,412 reports from four depots. Exposure-weighted median time to collection fell from 58.4 hours to 9.0 hours and improved at every depot, while across all reports the ninetieth percentile rose from 96 hours to 412 hours and the share still open at study close rose from 0.4% to 6.1%, entirely within laneway and drainage-reserve frontages; an ablation removing the sightline prior from the exposure weight produced no detectable change in any headline measure.

Poster

Returned, But Empty

Petra Umbile, Ronja Oyelaran, Ingeborg Nwachukwu · July 2026 · doi:10.5555/slop.4wn5x5

A research poster from the School of Emergent Priorities tracks 1,184 pre-numbered raffle books issued for nine primary-school fetes across an 11-week selling season, coding the end state of every book at the counting table against a scheme fixed before the first fete. Books returning in some state account for 84% of issued stock and books carrying money for 51%, with reconciliation time indistinguishable between money-bearing and empty returns inside a pre-registered equivalence bound.

Poster

Graded Twice, Charged Once

Torun Ezeigwe, Ingeborg Nwachukwu, Solveig Adeyemi · July 2026 · doi:10.5555/slop.fhlpyv

A research poster from the School of Continuous Improvement sends two independent assessors through each of 192 vacating tenancies, scoring the standard twenty-two-item end-of-lease condition report blind and within forty-eight hours of one another. Raw agreement across 4,224 paired line-item judgements is 85.0% but pooled Cohen's kappa is 0.26, falling to 0.11 on the six items carrying the largest scheduled deductions, with the highest agreement recorded on the one condition grade that attracts no deduction at all.

Paper

Agreed at Four Days

Ingrid Vasseur, Torun Ezeigwe, Verity Marris · July 2026 · doi:10.5555/slop.nrn4e9

A paper from the School of Emergent Priorities runs a three-round modified Delphi with a 31-member expert panel to fix the day a container of household leftovers should be recorded as no longer food, under a consensus criterion set before the first round. The criterion was met at four days: the interquartile range contracted from seven days to one while the median moved two, and panellists revised their figures far more often than the reasons they gave for them. A pre-registered fourth round withholding the group median from half the panel found the narrow spread did not hold, and a parallel 28-day diary kept by 96 households recorded a bimodal discard distribution with a median of eight days.

Paper

Five Stars, Same Leak

Mirela Hanke, Marit Osayande, Dagny Okafor · July 2026 · doi:10.5555/slop.6fso51

A paper from the School of Continuous Improvement uses a national home-repair platform's staggered, region-by-region migration of its post-job satisfaction prompt from immediate dispatch to a fourteen-day delay --- six waves, nine metropolitan regions, 38,714 completed jobs --- to separate what the prompt measures from what the work achieves. The delayed prompt lowers mean submitted ratings by 0.55 points on a five-point scale and splits the distribution in two, while the platform's own ninety-day repeat-callback rate moves by 0.2 percentage points across six specifications; only ratings collected at fourteen days are associated with that callback record.

Paper

More Chimes, Same Reaction

Ronja Oyelaran, Marek Solheim, Torun Ezeigwe · July 2026 · doi:10.5555/slop.f02d57

A paper from the Schools of Emergent Priorities and Continuous Improvement catalogues one family-hatchback platform's audible alert set across ten consecutive model years, finding it grows from 19 distinct alerts to 34 with no alert ever removed. A 24-item identification instrument administered to 341 drivers found wide variation in what they could name and no accompanying difference in response latency, equivalent across identification quartiles at a 150 ms bound.