Most people are friends with people like themselves?
The selection bias of extrapolating an entire country (not to mention one as large and diverse as usa) by using your friend group should be very obvious.
I'm just pointing out obvious ways in which you have selection bias here.
FWIW googling, allegedly 1 in 10 people in rural america do not have a license. Or at least that is the claim floating around the internet. So yes, i think figuring out if this is a real issue should include both checking on rural and urban people, along with many other demographics. It is not obvious how this situation differs between urban/rural.
The selection bias of extrapolating an entire country (not to mention one as large and diverse as usa) by using your friend group should be very obvious.