Lately I’ve seen lots of articles and comments about Trump’s appeal to Democrats who might cross over to vote for him. It makes a certain amount of intuitive sense, actually. After all, Trump is not a doctrinaire conservative, to say the least. One could even say he’s a RINO of sorts, albeit an idiosyncratic RINO who is very conservative on certain issues and somewhat of a liberal on others.
Plus, the topic that seems to be firing people up about Trump is immigration, and although worries and concerns about immigration (illegal, legal, and refugee-esque) are more common among Republicans than Democrats, they are hardly limited to Republicans. So all in all, some Democratic support for Trump makes sense.
Democratic support alone wouldn’t automatically mean a victory over Hillary, strangely enough, either for Trump or for any other Republican candidate, unless that Democratic support is (to coin a phrase) absolutely HUGE. Short of that, for victory to occur, it would be necessary that Trump or any other candidate not lose Republican voters, as well. And of course we know that many Republican voters are not keen on Trump (same for Rubio, same for Christie, same for quite a few others).
Right now neither I nor anyone else can say what will happen in November of 2016. We’re limited to attempts to measure what’s going on in the here and now. Given that restriction, I must say that I’m skeptical about the those reports of widespread Democratic support for Trump—not because I don’t think it could be true, but because of the quality of the evidence. I already wrote about this a few weeks ago here, but since then a more recent poll on the subject has a lot of people buzzing.
There are lot of things about that newer poll that give me pause, however, even more than most polls. The first is that the pollsters showed respondents a Trump ad first, and were testing the results of the ad. That means that the ad was fresh in their minds, which certainly would be likely to create a positive bias in the questions that followed. The second is that there is no way to compare Trump’s crossover voters in the polls to the number of crossover voters other GOP contenders might get. More? Less? We don’t know; because the pollsters only asked about Democratic support for Trump (because they were actually testing the ad). The third is that this was an internet and partially app-based poll; they can sometimes be quite valid or sometimes very skewed, and I haven’t seen anything about how these particular subjects were selected in this particular poll.
The fourth is that there were 916 “likely voters” in the poll. I haven’t found any word on how that broke down in terms of number of Republican vs. Democrat respondents, but let’s just say that half were Democrats (458). Twenty percent of those said they would cross over for Trump, which means that about 91 of them said that. That’s a rather small “n” on which to pin an entire theory that Democrats are going for Trump. Plus, if Trump were to run against Hillary, 14% of Republicans in the poll said they would vote for Hillary, which would be 64. So the net difference among the entire voter pool of 916 between Democrats-for-Trump and Republicans-for-Hillary would be 27.
What’s more—and this is what especially gives me pause about current measurements of the “Democrats-for-Trump” phenomenon—as I wrote a while back, in more conventional polls with a more conventional design, in a posited race against Hillary no Democratic advantage for Trump over the other candidates appears. Nor, by the way, is there an advantage for Trump with black voters, another demographic that many people say goes for Trump more than for the other Republicans.
So far, I’ve not seen any conventionally-designed polls that find an advantage for Trump with such groups. In fact, the opposite appears to be true. For example, the very latest national poll from Fox, which I have already discussed here, had some details that I didn’t get into in that earlier post but which are relevant.
If you look closely at questions 22-25 in that poll, which was taken January 4-7 and involved a sample of 1006 respondents queried by telephone (cell and landline), you will see that Trump does slightly worse against Hillary among Democrats and among black voters than the other leading GOP candidates do. Take a look if you don’t believe me—and these are typical of results I’ve seen in earlier polls.
In a matchup against Hillary, Cruz gets 11% of those identifying as Democrats, whereas Hillary gets 6% of people who say they are Republicans. Rubio gets 12% of Democrats against Hillary’s 5% of Republicans, a trifle better. Bush (remember him?) gets 10% of Democrats to Hillary’s 7% of Republicans, a tiny bit worse. And Trump gets 9% of Democrats to Hillary’s 8% of Republicans, which is a bit worse, although they all cluster rather closely together and the differences are not so very significant.
Against Hillary, Cruz gets 5% of the black vote, but Rubio gets 9% of the black vote. Could be significant, I suppose. Bush gets 6% of the black vote against her. And Trump? 4% of the black vote. Again—except perhaps for Rubio—they all are very similar, but Trump does slightly worse.
You can say it’s all a bunch of garbage. You can say that things will change because there’s lots more time (and I’d agree with you there). You can say people are lying to the pollsters, but I think that’s just another way of saying “I’ll throw out what I don’t like and keep what I do like.” If people are lying, we’re all just guessing, so don’t cite statistics about 20% of Democrats supporting Trump from a different poll where they could just as easily be lying. And I can think of a very good reason why people would lie and say that—because they are Democrats who think Trump would be a weaker nominee for the GOP than the others would and therefore they want Trump to be nominated, so they want to encourage Republicans to think of him as a winner who will draw the Democratic vote and defeat Hillary.
Is my theory right? Haven’t a clue; it certainly could be wrong. But I’m airing it to show that there’s no problem thinking up ways to discredit polls if you’ve got a mind to. As for me, I tend to think that polls tell us something real, and that most people don’t lie to pollsters (and that only a very tiny minority do), but that the best way to evaluate a poll is to look at sample and methodology, and then look at average trends in many polls over time. The trends I see so far with Trump are that he doesn’t appeal to Democrats or black voters more than the other GOP candidates do, and perhaps even slightly less. That could change. I can certainly imagine why he might end up appealing more to those groups, but so far I just don’t see it.
[NOTE: People also often bring up crowd enthusiasm when speaking of Trump. In general, I don’t think crowd enthusiasm matters in terms of election results, for reasons I already stated back in 2012 and also here.]