That would probably make it seem, though, that the buses actually don't arrive (on average) every 10 minutes, since you'd oversample the buses that take longer than 10 minutes.
After all, the point of the bus arrivals isn't in service of the bus (or driver) but of the passengers. Observed average wait time at each bus stop is a better measure. The even better measure would be average wait time weighted by number of passengers [1].
[1] which is tougher to measure empirically, or even model, than just average wait time for that one person, since it requires counting passengers boarding, not just bus arrival times.
Better to consider each bus stop as an asset to invest in, the more valuable it is, the more people you can serve.
Perhaps you misunderstood my point, which was more about data and statistics, as is the article itself, rather than transportation.
A similar argument could apply to the article's example of "average class size", where that's a valid statistic when observed by a teacher (or facilities manager), but misleading to a potential student. Something like "average size of a freshman's classes" would be more meaningful to a prospective student, and "oversampling" would not be a valid complaint there, either.
> instead of waiting at the bus stop, because of what happened to them yesterday at the bus stop.
It sounds like you're suggesting that there's an even better measure than the two I proposed, rather than the original measure being better. If so, I don't dispute that there could be many more, as I never claimed "best".
In this instance, though, measuring people who never show up to the bus stop in the first place is impossible, and even measuring those who showed up but abandoned waiting (i.e. never boarded) is impossible without additional instruments (whereas, presumably, electronic fare collection equipment could closely enough approximate counting boardings).