Skip to main content
Slop University grades every returned seed packet's reliability

Slop University grades every returned seed packet's reliability

A Slop University paper grades every seed packet returned to a community seed library into one of five reliability kinds, and reports the Adaptive Metrics Lab making the schema mandatory network-wide.

A season of 2,412 returned packets across forty-six garden clubs ends with a mandatory five-kind grading standard and no exceptions.

Every packet returned to a Slop University-affiliated seed library now passes through a four-step credentialing process before it goes back on the shelf. A new paper from the School of Continuous Improvement reports the process’s first full season: 2,412 packets, forty-six garden clubs, and a five-kind reliability grade for every one of them.

Dr Bram Ntuli, with Associate Professor Casimir Beng and Dr Joost Nwosu, built the grading scheme from a germination trial, a label check against a photo-voucher record, and a “segregation index” that catches a hybrid variety’s seed being saved and handed back as if it would grow true. Two independent volunteer graders assess every packet, with the team reporting substantial agreement between them.

A commons only works if the thing coming back is what it says it is, and until now nobody had actually checked.

— Dr Bram Ntuli, Lecturer and Convenor, Improvement Grand Rounds

The paper’s more delicate finding concerns what happened once clubs could see each other’s grading results: pass rates rose across the network within a single season, while the packets themselves, by the team’s own measure, did not improve to match. “A number everybody can compare is a number everybody eventually manages,” said Associate Professor Beng, the Adaptive Metrics Lab’s lead. The School frames the season as the first confirmation that its grading instincts apply as well off campus as on it. The Lab has adopted the five-kind schema as the mandatory standard for every affiliated club’s return table regardless, retiring a packet outright if it fails to clear the top grade twice.

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