I have been reading my new numbers guru, Rachel Bitecofer, and here are some takeaways. Much of this is not surprising, but some is, at least to me. You can check out her analysis here. All of this rests on the well-known but often under-appreciated fact that voter turnout in American elections is astonishingly low – maybe 60% of eligible voters in presidential years and [usually] 35-40% in off years. One might ask whether, with turnout like that, American voters deserve a democracy, but that is a discussion for another day.
Remember, even in solidly Republican House districts, with an off-year turnout of 35-40% there are large numbers of non-voting Democrats [often clumped together in big cities, college towns, and the like.] Suppose in a Republican district with 400,000 eligible voters, the entire electorate breaks 55-45 for the Republicans, a seemingly unbeatable +10 advantage for the Rs. This means there are 220,000 Republicans and 180,000 Democrats in the district. If only 40% turn out, and there is equal enthusiasm on both sides, a ten point Republican victory means that 88,000 voted R and 72,000 voted D, a 16,000 margin. But there are still 108,000 non-voting Ds! If 15% of them can be motivated to get off their asses and vote, the Democrats win narrowly. Common sense suggests that it may be easier to motivate 16,200 non-voting Ds than it is to turn 8,100 Rs to the D side.
OK, with that as background, let me summarize what Bitecofer says she has found.
First: Contrary to Conventional Wisdom, the big Democratic House victories in 2018 were not driven by voter concern about health care, nor were they driven by Republican defections to the Democrats. The victories were the result of an enormous surge in the turnout of reliably Democratic segments of the electorate driven by hatred of Trump. This may sound unsurprising, but it has enormous implications for the choice of candidates up and down the ticket, including at the very top.
There is a hunger out there in the hearts and minds and stomachs of scores of millions of Americans for a chance to vote against Trump. I saw this the day after Inauguration Day in 2017 when I attended the Women’s March in Washington.
Second: the single most significant determinant of the relative popularity of candidates for the Democratic nomination is the astonishing, unfathomable, too easily overlooked sheer ignorance of the American electorate. Bitecofer has some fascinating and rather complex analysis of the role of name recognition in the poll results we have all been seeing.
I recall many years ago, when I was living in Massachusetts, reading a news story about name ID among voters. Various Democratic and Republican politicians had name IDs in the 60s, 70s, or 80s, but topping the list, unsurprisingly, was Teddy Kennedy, the senior Senator, who had the astronomical name ID score of 95%. The author of the story oohed and aahed about this astonishing figure, but all I could think was, “Dear God! One in twenty people I pass on the street have never heard of Teddy Kennedy. What alternate universe do they live in?”
As Bitecofer shows, “Statistically, low levels of name recognition have massive impacts on polling data. It is not possible to compare favorability of low and high name ID candidates.” Now, political preferences are relatively difficult to change, but name recognition changes considerably as more low information voters turn their attention to what we junkies think about obsessively every day.
A thought for the day