A School of Continuous Improvement paper finds that crossing a neighbourhood forum's Trusted Neighbour vote threshold causes a real rise in duty-roster volunteering, and reports the threshold's own adoption as a specification elsewhere in the University.
A twenty-five-vote cutoff set for account integrity now predicts duty-roster hours, and the number has become University policy
Crossing a vote threshold that a neighbourhood forum set for account-integrity reasons, and never meant to steer anyone, changes how much unpaid duty-roster work members actually do. A paper published today by the School of Continuous Improvement reports a regression-discontinuity study of the threshold, and finds the change survives even when the badge it unlocks is invisible.
Dr Joost Nwosu, Postdoctoral Fellow, led the study with Dr Bram Ntuli, Lecturer and Convenor, Improvement Grand Rounds, and Dr Thandiwe Solberg, Senior Lecturer and Convenor, Demo Quarter. Members of 214 affiliated local forums who cross the network’s twenty-five-vote Trusted Neighbour cutoff volunteer for measurably more duty-roster hours in the following eight weeks than members who fall just short, a jump that holds through five weeks when a display fault hid the badge from every other member. The threshold itself has not moved across eight semesterly review cycles, over which the network’s forum count grew roughly fivefold.
The vote count was never supposed to mean anything beyond the badge. Once we looked at what happened after it, that stopped being quite true.
— Dr Joost Nwosu, Postdoctoral Fellow
The release extends the School’s practice of testing an incidental rule for what it actually causes before anyone asks it to. The Indicator Commons has since adopted the same threshold, unrevised, as the reference vote-threshold for badge-gated duty rosters elsewhere in the University’s volunteer systems. “A number that started as a fraud check is now a specification,” said Dr Solberg. “Nobody voted on that; it travelled on its own.”
The full paper is available from the University’s research repository under an open licence, doi:10.5555/slop.mn6z54.