Wisconsin

I'm rarely this direct, but in this case I'll make an exception: Your mental model is completely wrong. If I may spell out its assumptions: Every voter is strongly engaged in the political process, and fully-informed. Each non-voter made a fully-conscious, considered decision to not vote as a protest.

I only need to go that far to make it obvious just how bonkers that mental model is. Only about 65% of eligible voters cast a vote in 2024, and that was high, fer cripessake! It's easy to find tons of surveys and person-on-the-street interviews to find out just how not-fully-informed so many people are, and how many people vote (or not) on vibes. The "uncommitted" voters in the primary were indeed the people likely to be politically-engaged and informed, and in the end a lot of them did hold their noses and vote for Harris in the general election. The numbers don't bear out that they materially affected the outcome, either way.

By contrast, in my mental model, Harris needed to motivate a lot more of the 35% of non-voters to show up at the polls by giving them a reason to make the effort. And her campaign did not, so those people engaged in exactly the sort of "not going to think about it" behavior that powers us through so much of daily life: It's only one vote, it doesn't matter. I have to work and get dinner and pick up the kids, and going to vote is a hassle. Other people will vote. It's just politics, it doesn't affect me. Nothing really changes either way.

That last rationale was the real problem with the genocide issue. Harris's messaging was muddled, at best, and didn't provide anybody a reason to make the effort to vote. A strong, vocal opposition to it would have provided more voters with a rallying point and energy to overcome the inertia and get out to the polls. (And, frustratingly, we found out after the election that the campaign knew that their messaging on Gaza was a losing issue at the time.) Of course, a more populist economic message could have provided that energy, or a promise to break from Biden policy, which wasn't working for a lot of people.

I'm not saying that my model is perfect — all models are wrong, after all — but it provides a far more plausible explanation of the 2.5 million "missing" voters than the idea that they all refrained from voting as a protest over an issue that wasn't even in the top 3 issues that they identified in surveys.