I was refuting his anecdotal claim with mine - I meant this to illustrate the futility of using anecdotal evidence, but the words didn't convey the sentiment as well as I intended.

I fully understand there is most likely some of normal distribution of sizes, but that's a fact that the author seems to have largely ignored as well.

On an unrelated note, the site seems to be down.

Here is some data to help someone answer such questions: http://dined.io.tudelft.nl/en,dined2003,102. It shows that, for Dutchmen, the difference in _average_ hand length between males and females is about 2cm. Looking at the standard deviation and across ages, one gets (at one sd) a variation between long male and short female hands of (I guesstimate) about 5cm. Thumb reach distance will be, say, about half that. I would guess elderly users not only have shorter hands, but also have less dexterity, and of course there are plenty of people outside of the +/- one stadard deviation range (37%, IIRC), so it would not surprise me at all to see difference in 'reach' of over an inch, even in small grouos of users.

And to add to this: it may be acceptable for Apple that 10% of people can't use the device optimally, but it certainly won't be acceptable if 50% of people can't use the device optimally. So you have to err on the smaller side of caution.