My figure was based on miles travelled, not trips. Further, your assumption that local travel would severely skew the data seems at best unsupported. Think about all those trips taking the kids to school/soccer practice/etc. Having a lot of people in the car is going to radically skew the average number of occupants up.
> Just like how the average person has around 1 testicle, this data is also multimodal
Then why are you essentially arguing that people with two testicles are a niche case?
The overwhelming majority of car trips that happen are going to be car trips that make sense to be car trips. That's not evidence of driving's economic sensibility, it is selection bias. The question is of all trips where a person needs to go a certain distance, how often is flying the cheaper option, and for long trips it is a lot.
The transportation to/from the airport at either end can often be significant, as many parts of the US are not accessible via public transit. And rural destinations often have no final-destination options other than an expensive rental car. It is common to incur these costs.
But, that is not the topic of conversation above -- we were talking about trips 600+ miles in distance. These are almost exclusively not commutes.
Averages don't necessarily describe your whole data set. Just like how the average person has around 1 testicle, this data is also multimodal :) ... People commute alone, but they go on vacation with friends/family.