Polls in Florida can be used to illustrate the idea that Trump is staging a huge comeback, or that he’s hopelessly behind. There’s one that says he’s doing pretty well there (2 points ahead of Hillary) and two others that put him far behind her in a cloud of dust (losing by 9 and 14 points).
So, what’s a person to think? As I’ve said before, the best predictors are poll averages over time. And the averages get even more predictive as one gets closer to voting day, although averages over time will tell you the general trends. The poll showing Trump doing well in Florida is somewhat of an outlier at the moment, and only time will tell if it’s the beginning of a new and possibly meaningful trend back to a competitive race there. Till then, it will function as something that Trump supporters cite in order to shore up the Trump troops.
One of the most interesting results of the Florida poll showing Trump ahead is that it also shows him as having 20% support among black voters. But it’s always good to look more closely at what a figure like that actually represents, and how large the black sample was. Unfortunately, I can’t seem to locate the most detailed report of the poll, but this piece says that it questioned 1200 likely voters. If we’re trying to figure out how many black voters were queried, let’s be very generous and say it might have been 15%, which would have been 180. Twenty percent of 180 is 36. So, that would mean that 36 black people in the poll said they would vote for Trump (and the number would have been even smaller, if less than 15% of the poll respondents were black).
You can see that we’re dealing with a large margin of error here for reporting on the black voters, because of the small numbers involved. That’s true of a lot of polls, as it turns out.
One more thing. People assume that 20% would be an awful lot of black support for the Republican, if it were true. And I agree it would be impressive, although I’m not at all sure it’s true (and it is an outlier result compared to the trend of other polls on Trump’s black support, but let’s put that aside for the purposes of this discussion). In articles I’ve seen that discuss this new Florida poll and the 20%, the figure used for Florida comparisons is Obama’s 95% to 4% showing in Florida among that group. But that seems entirely the wrong comparison to me, because Hillary Clinton is not Obama—in many senses, but what’s particularly relevant is that she’s not black.
It seems to me that a better figure to use to compare would be the stats from the last time a black person was not on the presidential ballot, which brings us to ancient times: the 2004 election. And you may be surprised to learn (as was I, by the way) that George W. Bush received 13% of black votes in Florida that year.
That would be the baseline Trump would need to beat to be doing better. And in that poll I mentioned where Trump got 20% support from blacks, the difference between 20% and 13% would be (assuming as I did earlier that 180 black voters were questioned) 36 versus 23 supporters out of 180. I wouldn’t put all that much weight on differences of that magnitude.
Now, you may say “neo, you’re just trying to put down the poll because you don’t like Trump.” No, I’m not; you won’t find a lot of false optimism on this blog for candidates I like or false pessimism for ones I don’t. That’s my goal, anyway, and I think I reach it most of the time. Furthermore, in this case I don’t like either candidate and would love for someone else to win (not gonna happen, I know). So it’s not that Trump’s defeat would fill me with joy. It most definitely would not. It’s just that neither candidate’s victory fills me with anything but despair. So one could argue that I’m probably one of the more objective folks around when looking at their chances (and yes, I know that one of them is virtually certain to win, barring a black swan of extremely major proportions)—an elephantine black swan, if you like).

