Rockey's post on PI survival also had a graph on the persistence of PIs in submitting applications.
She noted that the 1989 first-R01-equivalent folks dropped off in their grant submitting persistence around years 3-5 more than the younger cohorts.
A comment by qaz on the prior post of mine triggered a thought.
What about people who otherwise didn't really want or need a NIH R01 grant but it was a requirement for (or strongly supportive of) a tenure case?
If department expectations/preferences (for tenure or in who they hired in the first place) have changed since the late 80s, this could explain the difference in early drop-out, one-and-done rates across cohort.