Your data with the context you don’t like.

FACT: 2003 - age-60 pilots? 100% Part 135.
FACT: Your own study says ≈80% of all accidents (any age) were Part 135.

Which means the so-called “U-shape” at 60+ isn’t about age at all, it’s about the dataset being entirely Part 135.

Or you could just say, “Darn, it was 2003, there were no Part 121 pilots over 60, and 80% of the incidents were Part 135 at all ages.”

!