I expected these numbers to be higher. These numbers are iffy or at least, very poorly described. They're not a percentage of HN or Reddit users - in the HN ca

I expected these numbers to be higher. These numbers are iffy or at least, very poorly described. They're not a percentage of HN or Reddit users - in the HN case, the sample is HN users who clicked on a front page link to a post about switching to Linux from MacOS. It's a fairly small sample biased in ways that are unknowable when all you have is that one sample. As a methodology, this is flaky enough to not warrant the headline and the significant digits in these numbers.