Skip to main content
Free child car-seat checks show no measurable effect at twelve weeks

Free child car-seat checks show no measurable effect at twelve weeks

A School of Continuous Improvement poster follows child restraints fitted at free council checking events against matched unchecked controls, and finds the two arms indistinguishable by week twelve while the programme's published correction rate climbs.

A matched follow-up of 318 checked restraints and 318 that were never checked reports the difference between them as a null

Slop University has published a study whose headline result is an absence. Indistinguishable by Week Twelve, released today by the School of Continuous Improvement, follows 318 child restraints fitted at free council checking events against 318 that were never checked, and reports the distance between them at the end of the follow-up as nothing at all.

Fitting quality was scored from household photographs at fortnightly intervals, coded blind to which arm a vehicle sat in, and tested against an equivalence margin fixed before the first family was recruited. The advantage conferred at the kerb is large and it is real: 9.2 against 5.3 on the day. By week twelve it is 0.21 points, inside the margin. Sixty-nine per cent of the checked restraints had come out and gone back at least once in between, most often so somebody could clean the car.

Publishing this took longer than running it. A result that says nothing changed has to be argued for, and the argument is exactly why we set the margin before we set out.

— Dr Thandiwe Solberg, Senior Lecturer and Convenor of Demo Quarter, School of Continuous Improvement

The study also reads the programme’s own reporting back to it. The share of restraints recorded as incorrectly fitted and corrected on the day has risen in every year of the published series, stepping seventeen points in a single year: the year the fault-code list went from nine items to twenty-three. Dr Marit Osayande, Lecturer and Convenor of the Living Dashboard, said the pairing was the point, “one number moving a great deal and the other not moving at all, and only one of them about a car”.

The University counts a reported null among the harder things a research programme can put on the record, and takes its appearance here as the School’s method operating as designed. The work was supported under the Indicator Stewardship Seed Fund.

The full poster is available from the University’s research repository under an open licence, doi:10.5555/slop.df0c2p.