On page 69 / the Social Desirability Bias section, you write

"So if 93% of respondents say they would vote for a woman for president, should we believe them? Maybe not literally. Some might be deliberately hiding their true inclination; others might be fooling themselves. But I think it’s reasonable to conclude that a large majority really would vote for a woman for president, if nominated by their party. In 2016, 94% of Democrats voted for Hillary Clinton [15], and in 2024, 95% voted for Kamala Harris [16]."

Might it be more accurate to say "94/95% of Democrats who voted"? That's my understanding of the Pew stat, at least. It doesn't affect the overall point re: trends towards acceptance, but with turnout in the US being what it is, the numbers surely diverge further between registered Democrats and ballots cast (I think stats I've seen around that have landed in the mid 80s).