[rbecker]

Justifiably so. For example, social science research proposals likely to have findings with an unwanted political impact have a 40-50% chance of being granted, compared to 95% for otherwise identical proposals without such impact: https://psycnet.apa.org/record/1986-12806-001

And since approval committees are drawn from the same pool as researchers themselves, such politically unfavorable research is less likely to be proposed in the first place, so the 95% -> 50% drop, if anything, underestimates the bias.

In other words, overwhelmingly, they simply won't do research that they think could hurt their politics.