There is a huge vested interest in portraying elections in the US as very close, right up to election day. Each side likes to do so to prevent complacency among their supporter and to nudge people to vote, contribute money, and volunteer to do campaign work. The media loves it because it draws viewers and generates ad revenue. And there is a vast network of election pundits who are kept in business by blathering away about every little nugget of news to say how they think it will affect the election, even though they have no idea. Then there are the political consultants and media operatives who become more important in close elections, as campaign seek to squeeze out every advantage.

In this race, the Harris-Walz campaign keeps hammering on the theme that the race is nail-bitingly close and that all stops must be pulled out to prevent creepy Donald Trump from winning. It does not seem to matter what the polls say. The creepy Trump campaign has a bit of a problem in that their candidate likes to boast that he is hugely popular even to the extent that he would win California, which tends to depress turnout. To compensate, they are suggesting, with no evidence at all of course, that voting systems are corrupt and that mail-in ballots are a scam, and that Democrats have a huge operation in place to steal the election from creepy Trump and thus they need to be vigilant and vote in huge numbers.

Opinion polls are of course the main mechanism used to gauge the state of the race. There are many polls done by many different organizations and these naturally will give different results, not necessarily out of deliberate bias but because they use different methodologies, a phenomenon known as the ‘house effect’. What the media will do is give great publicity to whichever latest poll shows the closest gap and not publicize much those that seem to indicate that one candidate is drawing away. Polls that seem to show a tie seem to make them absolutely giddy with excitement.

As a news consumer, it is better to not pay too much attention to the breathless headlines that accompany some new poll and instead look at poll averages and see how they change with time, though even there there is a house effect depending on which polls are counted in the average and over what period they are taken. Averages are provided by The Economist, The Guardian, YouGov, Real Clear Politics, and 538. I play closest attention to Real Clear Politics because its house effect favors Republicans. In the US, national polls have limited value since it is the results in the swing states that really matter but they are not entirely valueless. But Democrats tend to need a +4% margin nationally in order to have a break-even chance of winning the presidential election.

Bear in mind that how poll samples are weighted are based on models that use data from past elections and those might not take into account subtle changes (such as the likely voting of different demographic groups), or hard to measure items like voter enthusiasm, or new factors. In this election, we are going to see for the first time what effect the overthrow of the Roe v. Wade decision in 2022 will have on the presidential race. We already saw that it had some effect in the mid-term elections in 2022 but now the full consequences of that are being seen in all its cruelty, coupled with the anti-woman rhetoric of weird JD Vance. How that will affect voter demographic and turnout is hard to predict.