When writing the title to this piece, I got a sense of deja vu. I was pretty sure that I had written the piece that I am about to write before. On looking back through my archive I found that I had indeed written a similar piece on 11 September, but entitled ‘Why you shouldn’t read to much into polls which show Trump in the lead’. This is a similar article.
Earlier this week, polling released by ABC News and The Washington Post suggested that Trump had moved into a one point lead in national polls for the Presidency. Now it was expected that Trump would gain some support following James Comey’s disclosure that the FBI were re-opening their investigation into Hillary Clinton’s use of a private email server during her time as Secretary of State, however the result of this poll was particularly remarkable because just a couple of weeks earlier it had found a twelve point lead for Clinton. I would dispute both these findings. It is almost impossible for support for candidates, in what has generally been a pretty close election, to swing by thirteen points in the space of two weeks — it simply does not happen. It strikes me that for the poll to turn up such a swing, then there must be something seriously wrong in their methodology. Now before you say so, this isn’t simply me not wanting to accept that Trump may be in the lead. I expressed the same doubts when the same poll showed that Clinton led by twelve points. It was simply such an outlier from other polls and the polling average, then it surely can’t have been correct.
One of the reasons for this huge swing could be that supporters of candidates often stop responding the polls in quite the same number when their candidates are having a bad day. For Trump, this was following the release of the Access Hollywood tape and following a couple of relatively poor debate performances, when his campaign was arguably at its lowest ebb. For Clinton, this was when it was announced that the FBI was re-opening its investigation into her use of a private email server. This means that it is hard to build up a picture of what it actually going on by simply looking at a single poll in isolation.
In my opinion, it would be far more productive to look at the polling averages, which tend to be pretty accurate in terms of predicting the final result. Currently the RealClearPolitics polling average gives Clinton a 2.2 percent lead in a four way race, which seems to be a pretty plausible reflection of where the race currently stands.
In addition, it is worth comparing the polls in this election, with what the polls said at the same stage in previous elections.
In 2012, Mitt Romney led Barack Obama by one point (49–48) with one week to go in the campaign. In 2004, John Kerry led George W. Bush by one point (also 49–48) with one week to go until election day. Both Romney and Kerry lost. Even if Trump is one point ahead at this stage, it does not mean that he is going to win. If anything, what it might do is increase turnout on the Democratic side, with many other unenthusiastic voters coming out to vote in order to prevent a Trump Presidency.
Although this election clearly isn’t over yet, Clinton remains the likelier victor. Given past history, you should certainly be taking the polls with a pinch of salt at this stage.