A new Slop University paper audits six months of visitor sign-in logs across 74 aged-care facilities, finding the average-visit-duration figure behind an overnight companionship-staffing decision is set less by visiting than by each facility's own tablet timeout setting.
Six months of sign-in logs across 74 aged-care facilities show two in five visits closing on a timeout, not a tap-out
Two in five visitor sessions logged at the aged-care facilities in a new Slop University study did not end when a visitor tapped to leave. They ended when the reception tablet’s own timer ran out and closed the session on its behalf, the same way, every time, whether the visitor had left after ten minutes or stayed the whole afternoon.
The finding sits squarely inside the School of Continuous Improvement’s evaluation-ecosystem agenda, and the University reads it as confirmation of the patience its measurement research asks of a number before that number is trusted with a roster.
The study, led by Senior Lecturer Torun Ezeigwe with Lecturers Solveig Adeyemi and Marit Osayande, set six months of an aged-care operator’s sign-in logs against a survey of the facility managers who configure each tablet’s timeout window. Facility average visit duration, the figure that feeds the operator’s overnight companionship-staffing roster, tracked the configured timeout almost exactly; it tracked little else the researchers measured. Managers’ own sense of how often that timeout fires, the team found, undercounts what their logs actually show by a wide and consistent margin.
A number that closes itself the same way every time isn’t lying to anyone. It just isn’t telling you what you think it’s telling you.
— Senior Lecturer Torun Ezeigwe, Director of the Master of Applied Measurement
“The gap here is the ordinary kind,” said Lecturer Solveig Adeyemi. “Nobody falsified anything. The tablet did exactly what it was configured to do, tens of thousands of times over.”
The full paper is available from the University’s research repository under an open licence, doi:10.5555/slop.ynpltg.