At least this time around.
Here's the NYT report (gift link):
My questions:
• What proportion of the votes for Hong were supporters of the Republican candidate Tom Tiffany — voters who thought Hong would be a weak opponent for him? It was an open primary, and I saw some open encouragement of this behavior. Tiffany obviously had his nomination sewn up. You can see that David Crowley only won by about 3,000 votes. How much larger would his margin of victory have been without those insincere strategic votes for Hong?
• Why did Hong concede so quickly, with such a narrow margin? Why not wait at least until all the votes are in? Why not ask for a recount? She's entitled to a recount if the margin is 1% or less, but it's only paid for by the state if it's less than 0.25%. To drag out the process after the primary would help Tiffany. That's good reason not to indulge in a recount, but why not wait until all 100% of the votes are counted? She conceded when it was less than 95%.
• Why were the polls so bad?! The final Marquette poll, from July 22–27, released July 29, had Hong at 38% and Crowley at 7%. There was also another candidate at 16% (Barnes), but he dropped out. Still, giving all the Barnes votes to Crowley would only put Crowley at 23%. That 38% for Hong looked awfully strong. But I wouldn't assume the polls got it wrong. The polls inspired a lot of people to attack Hong, making her look unusually foolish and driving a lot of Democrats into voting for what is practical and strategic and not for the radical who excited them so much.
• Speaking of exciting radicals in Wisconsin, these votes were falling into place as a highly visible and emotional street protest was escalating in Madison after the death of Corey Ruiz. Though Ruiz had apparently attacked the police with a knife, Hong quickly denounced his killing as "a state-sanctioned… execution." Those protests went on for 2 weeks, mostly during the period when there were no new polls. How much did those impassioned protests turn Wisconsin voters against Hong?
***