Alan Abramowitz lays down what is probably the most reasoned assessment of where things are with regards public opinion as it stands now, with regards 2016: basically, the big picture looks exactly like back in 2012. On the one hand, this is broadly true…but also misleading. Let’s think about the moving parts for a minute.
What the findings show is that, within each demographic subgroup, the attitudes towards Clinton and Trump are subsumed by their partisanship. Republicans will not swing towards Clinton. Democrats will not swing towards Trump. However, this is also a bit misleading: no election is ever won by drawing in the voters from the opposite side, barring colossal failures–and even then, it is not clear if real large scale defections have ever happened. Even in 1936, Republicans voted Alf Landon, not FDR, after all. So who are the voters who might be persuadable, who might behave differently? And potentially, more importantly, how many of them are there?
The first step we might want to take is drop from the sample EVERYONE who is a partisan. We don’t need to ask what the Demcorats and/or Republicans think, at least for not who they prefer relative to the other. We already know the answer. We also have a fairly good idea what the vaguely defined group titled “independents” think: in aggregate, they almost split halfway, at least until when the election day is approaching. We want to know what the subsets within these “independents” think–but that we cannot do, without the raw data. The numbers we will be dealing with, unfortunately, will be quite small, which will affect the statistical power of our analysis, but that is part of the point: the differences that will decide the election–not just this election, but any generic election–will be smallish margins, not the big national figures in absence of something catastrophic.
More important is that the real determinant of the outcome will not be whether the voters of any demographic subgroup change their preferences, but whether they bother to show up. Here, there are some potentially significant facts which, unfortunately, do not crop up in the surveys. First, in 2012, the turnout among African Americans was unusually high, almost on par with the whites, and in some subdemographics, e.g. those without college education, actually higher than whites. We will not see this happen, if we can guarantee one thing now. Clinton campaign towards African Americans is built on the usual Democratic Party machinery, steering the subset of the African American voters who do vote quite consistently–essentially, Churchgoing old ladies. Among the larger population, unfortunately, this is a rather small proportion. If the African American turnout drops significantly, this will cut into the Clinton margin enough, at least, to make the race much closer, although probably not enough to flip the race. This is complemented on the Republican side by the prospects that Trump might be able to raise the turnout among the so-called Missing White Voters, the less educated, less affluent whites. Again, the picture here is mixed. It appears that Trump performs better among them, in terms of support, than Romney did, but the difference is slim. The real problem for Romney, of course, was not that those who did vote did not support him, but the turnout among them was quite low–thus the “missing” part. The realistic boost in turnout that Trump can bring about among them, however, does not seem particularly high–but there is no data that would permit gauging this with any kind of reliability. (I have attempted to cobble together a methodology that extrapolates the numbers from primaries, but lack of previous data in corresponding situations–1980 and 2008 seem the only comparables, and both are too unusual elections to draw too much from them–means that I wouldn’t put even two cents on them.) Of course, all these are potentially dwarfed by the likely reaction among the more conventional Republican voters: they will not turn towards HRC, but they could easily mimic the behavior of the “low information voters” (which, ironically, they are in this context) by staying home in extraordinary numbers. Do note that the dismal poll numbers seen for Trump, when they do show up, is mostly the product of the lukewarm support he enjoys among the Republicans, even if they are punctuated by other polls showing that they are still willing to stand behind the Republican Party, even if its ticket is headed by Trump. This seems like a recipe for an unusually low tunrout that could change everything.
I think, on one hand, everything is staying as it always has been: if people do show up, they will behave as they did in 2012, demographic by demographic. But the real question is about turnout, of which we have precious little data or even theory to build around.
PS. In a way, I have laid out another version of the old characterization of Trump candidacy, that he has a high floor, but low ceiling. But this has been true throughout the primary campaign: Trump never really broke through his ceiling–just that his opponents could not get their act together (and to be fair, Trump has been very skilled in exploiting their inability to coalesce). In the general election, the story is similar: the very best that he can do is to squeak out a narrow victory by mobilizing all the Republicans plus a handful of unconventional voters who would not have turned out but for his candidacy. This is risky because, compared to Trump’s Republican opponents, Democrats start from a very high floor themselves, in form of the broadly favorable demographics that stack the electoral deck significantly in their favor. Trump enjoys, however, a great advantage in form of the Clintons, who seem determined to exclude from their coalition a lot of voters who, in return for some concessions, might pad the Democratic votes significantly–a large subset of the Sanders supporters. Clinton may still triumph in November, if only because Trump’s gamble, that he could still somehow maintain the Republican coalition (especially the affluent Republican women) fails, not so much by their defection to Clinton but to “I ain’t voting.”
The whole affair, from the Democratic side, seems way too smug and overconfident to me. Trump is engaged in a huge gamble, for relatively small reward and very high risk. It is improbable that he will win–although, to be fair to him, without this risk, it is also improbable that the Republicans had a serious chance at the presidency (whether he could have gotten the chances that he has to bring out such MWV’s he will cobble together without royally pissing off the usual Republican voters is the big, giant question. There used to be the talk of “hippy punching” to prove your anti-communist credentials. In this case, one does have to punch some elites to prove your populist credentials.) Precisely because it seems so difficult for Trump to win, Clinton is entrenching her base and refusing to expand her coalition. But would she have done otherwise, had a more conventional Republican been the nominee? Republicans ARE handicapped by their demographic disadvantages so that it will have been a difficult uphill climb, even if through a different route, for any other candidate. I think, paradoxically, we are set for a stage where things can really surprise. While improbable, I suspect a Trump surprise is more likely than one’d think–more likely, in fact, than would have been possible with any other Republican candidate, especially in face of the smug and complacent approach that the Clintons are taking. The odds are not good enough to bet real money. But, with any other Republican candidate, there wouldn’t have been odds enough for even fake money.