On the other hand there are other similar studies that reach similar conclusions, and specifically try to control for aerodynamics e.g. [1] which says
> The weak positive relationship between vehicle registration year and splat rate suggests that newer vehicles are more efficient at sampling insects than older vehicles.
i.e. they saw more insects on newer cars compared to older ones in the same time period.
In general ecology studies aren't like lab physics, you can't control every possible confounding variable; the systems are too complicated and studies ex-situ have their own limitations. But refusing to engage with the data we do have because it's not perfect isn't going to help you make better decisions, and doesn't represent some moral high ground.
[1] https://cdn.buglife.org.uk/2022/05/Bugs-Matter-2021-National...
While the retraction brings into question the anecdotal evidence for the windscreen phenomenon ( https://en.wikipedia.org/wiki/Windshield_phenomenon ), there are other studies with other sampling approaches that support the global insect population collapse ( https://en.wikipedia.org/wiki/Decline_in_insect_populations ).
Although the funniest one was driving through a cloud of moths on the A9 one summer about 30 years ago in my little Nissan, which hoovered up enough of them to choke the air filter and die on the next (fairly long and steep) hill. They were hell to get off the windscreen too.