How it is actually tested
Updated
On this page7
The children
Aged 42 to 51 months, which is three and a half to just over four years old, split into bands: 30 percent aged 42 to 44 months, 40 percent aged 45 to 48, and 30 percent aged 49 to 51. Testing begins with a panel of 50 and can extend across further groups to 200.
Why that narrow age range
It is the age at which manual dexterity has developed enough to attack a closure but comprehension has not developed enough to follow a printed instruction. Younger children lack the strength and coordination; older ones can read and reason. The regulation targets the worst case.
Two periods, and a demonstration between them
Children get a first period to open the package unaided, then a demonstration, then a further period. The headline 85 percent figure is effectiveness without a demonstration, which is the harder measure and the realistic one.
The seniors
A hundred adults: 25 percent aged 50 to 54, 25 percent 55 to 59, and 50 percent aged 60 to 70, with 70 percent of the panel female. Five minutes to open and resecure, then one minute with an identical package.
The one-minute test is the real one
The five-minute period lets somebody work out the mechanism; the minute with a second package measures whether they can then do it. That is much closer to how a package is used at home, and it is where awkward designs fail.
This costs money and takes time
Recruiting panels of small children and older adults to a protocol is a laboratory service with a real price and a lead time. That cost is the reason certified packages are sold as specific tested combinations rather than assembled freely from parts.