One strong source of selection bias is that odds are the commuters any of us know are either our co-workers or our neighbors. Both groups are self-selecting for a particular commute distance, be that long or short.
e.g. my co-workers work at facilities usually on the far outskirts of a metropolis or else buried down in the crappy part of town. And they’re highly paid folks who tend to be live in the more expensive suburbs. And who don’t mind driving extra far because they only do it once or twice a week and usually at off-peak times.
That group is strongly self-selected for long commutes.
My neighbors when I lived in typical far-edge gated McMansionville suburbia-bordering-ruralia were likewise self-selected for longish commutes. Sure, one guy was a doc who drove 3 miles to the local small town hospital. But his pal the other doc on the block drove 35 miles in to the big city hospital. The density of high-dollar jobs was very low out where we lived, so most of the high earners drove a long way. And with those house prices, high earners were the only kind there were.
That group was also strongly self-selected for long commutes.
Conversely, folks who live in middle-sized cities and small towns that aren’t suburbs will likewise be strongly self-selected for neighbors and co-workers with 3-5 mile commutes.
For me personally, the word “commute” doesn’t mean “the travel process to get to work”. That’s called “getting to work”. By definition, it isn’t a “commute” unless it’s a long drive, long train ride, etc. IOW, if it takes 15 minutes, it isn’t a “commute”. But I admit that may be my own personal (mis-)definition of the word.