So, I decided to look at some of the polls that are coming out of Florida over the last few days to get an idea of how accurate they might be. However, there is one poll that we can (for a lack of a better word) call “bullshit” on.
So, to set a baseline, let’s look at pre-Election Day voting numbers in Florida. It is important to note that these are hard numbers, meaning that there is no “sampling error”, because it isn’t a sample of a population, but the actual vote that has happened to this point. Additionally, polls are supposed to give us a “snapshot” of a particular moment in time. Since the polls I will be referencing in this piece were just conducted, we can take the numbers from this exact moment. Therefore, if the poll is an accurate reflection of “this moment in time”, there should only be a difference of the “margin of error” between the polls and the actual numbers. Luckily, we can match these numbers up to determine the true accuracy. Again, that is important to remember.
As of right now, 58.8% of votes cast have been done during early voting, with 41.2% happening through vote by mail. Of those who have done early voting, registered Democrats have a 2.5% advantage over Republicans (40.1% D, 37.6% R, 22.0% NPA/OTH). As far as vote by mail, Republicans have slight advantage of 2.7% (38.4% D, 41.2% Rep, 20.4% NPA/OTH). Alright, remember those numbers.
With those fresh in our memory, let’s look at the CBS/YouGov poll released today. The poll states that the race is dead even at 45%-45%. However, those are not the numbers we are going to look at, for now.
According to the poll, of those who have cast a ballot in pre-Election Day voting, 47% did it through early voting, with 53% doing it through vote-by-mail? See the problem? These numbers are not only the opposite of what the hard numbers show, but they are also 12% off target! The margin of error for this poll is +/- 3.6 percent. However, comparing the survey numbers to the hard numbers shows that the poll is way beyond the margin of error. Remember, a poll is a “snapshot”, and that snapshot should reflect the current numbers. However, they do not.
In this case, the survey favors those who have voted by mail, who, as was mentioned early, are more likely to be Republican. Therefore, the poll will favor Donald Trump over Hillary Clinton. Unfortunately, the poll does not tell the preference of those who voted before Election Day. If it did, we could possibly further examine their error.
Unfortunately, other polls released are not as recent, and I have not archived all of the early voting numbers (which was stupid on my part).
Still, the point stands. This poll showing a tie does have a major problem, which shows it favoring Trump. And if the poll is “tied”, yet has this error, it is pretty safe to say that Hillary is probably winning the State of Florida.