We are just two weeks away from election day on November 5th. I of course do not know who will win the presidency, although I am cautiously optimistic that Kamala Harris will. But that does not count for much since I am an optimistic person by nature, tending to look for hopeful signs that the future will be better than the past, even if the odds are against it.

The national polls have been showing a small but steady lead for Harris ever since the beginning of August. When Joe Biden dropped out on July 21 and the Democratic party quickly coalesced around Harris, there was a surge in enthusiasm for her, with massive numbers of volunteers signing up and huge campaign contributions that rapidly erased the lead that creepy Donald Trump had had up to that point. People tend to expect trends to continue indefinitely but they don’t, and when the support for Harris predictably plateaued, it felt to some observers that her campaign was losing steam when it wasn’t. What we were seeing is a new equilibrium state with Harris that is much better than the previous equilibrium state when Biden was the presumptive nominee.

Her lead in the polls have not reached the 4% level where it can comfortably be expected to overcome the GOP advantage in Electoral College mathematics, so the big question being debated is how far off the polls might be in gauging actual voter sentiment. It should be borne in mind that poll reports are no longer simply aggregates of how people actually responded. There are many factors that complicate things, such as estimating the likelihood of voting by different demographics, adjusting for the fact that many people do not answer calls and so on, and polling outfits have to estimate what the preferences might be of people who don’t answer. This means that they have to increasingly use models to try and factor in these effects. Pollsters are of course aware of this and most sophisticated polls use what they learned from previous elections to refine their models this time around, especially if they were off by a large amount earlier. But the reality is that the further you get from raw data and the more that you rely on models to estimate effects, the more that systemic (as opposed to statistical) uncertainties are introduced.

The Pew Research Center says that as many as twelve kinds of adjustments need to be made to the raw data to take into account all these things.

[T]here is a growing realization among survey researchers that weighting a poll on just a few variables like age, race and gender is insufficient for getting accurate results. Some groups of people – such as older adults and college graduates – are more likely to take surveys, which can lead to errors that are too sizable for a simple three- or four-variable adjustment to work well. Adjusting on more variables produces more accurate results, according to Center studies in 2016 and 2018.

A number of pollsters have taken this lesson to heart. For example, recent high-quality polls by Gallup and The New York Times/Siena College adjusted on eight and 12 variables, respectively. Our own polls typically adjust on 12 variables. In a perfect world, it wouldn’t be necessary to have that much intervention by the pollster. But the real world of survey research is not perfect.
…

The real margin of error is often about double the one reported. A typical election poll sample of about 1,000 people has a margin of sampling error that’s about plus or minus 3 percentage points.
…

The problem is that sampling error is not the only kind of error that affects a poll. Those other kinds of error, in fact, can be as large or larger than sampling error. Consequently, the reported margin of error can lead people to think that polls are more accurate than they really are.

We also have the phenomenon of poll ‘herding’ as the election gets close. A polling company’s reputation (and future revenue) can be badly damaged by getting it wildly wrong and what people remember are the polls close to election day. So as that day approaches, if a poll initially gives a tentative result that is very different from what other polls are reporting, there is a tendency for those pollsters to look at their model and see if small adjustments can bring it closer to alignment with the rest. What is happening is that each poll is using other polls as a test of their model’s validity because that is the only test they have until the election results come in. So as the election nears, the poll results will tend to converge. But this practice throws off those who average different polls in an attempt to reduce uncertainty because that process assumes that the polls are independent. But even if all polls are independent and close to each other, there are still too many uncertainties to make any predictions about the outcome of a close race with anywhere near certainty. Finally, it is hard to gauge voter enthusiasm for a candidate, but that can greatly affect turnout among people who are usually apathetic.

But apart from the final outcome, there are some things we can predict with a considerable degree of confidence.

One is that Harris will win the popular vote (as Democrats have done in every election but one since 1992) but unless the early results show a landslide for one side or the other in the Electoral College, the election outcome will not be known for several days. Election night has now become election week and possibly even weeks. The second thing we can be confident of is that whatever the early results show, creepy Donald Trump will claim an overwhelming victory, perhaps even in all fifty states and the District of Columbia. Anything short of that will be ascribed by him to cheating. The third thing is that as the results come in for each state, there will be a ‘red wave’ (based on the votes cast on election day) and a ‘blue wave’ (based on early votes and mail-in votes), because Republicans tend to prefer in-person voting on election day while Democrats tend to vote early. The order of the waves will depend upon each state’s practices, whether the process of preparing early votes for counting starts before election day or only begins on that day. (Federal law prohibits the actual counting of votes before election day.). In states where the red wave comes first, creepy Trump and his supporters will demand that counting be stopped at once and that the blue wave is due to cheating.

If the final result shows creepy Trump losing, expect a concerted GOP effort to mount various challenges in all the states he lost, claiming that there was massive fraud. The GOP has put in a major effort around the country to challenge the eligibility of millions of voters identified as Democratic. They have recruited thousands of so-called ‘poll watchers’ to observe the voting process at precincts around the country, and these people’s opinions will likely form the basis for challenges. The GOP has changed laws in states like Georgia to make it easy to challenge voters and these will be invoked if the results are close. These claims will gain less traction if creepy Trump loses by a landslide. We have reached a sad state in this country where a Democrat has to win by a landslide in order have their election be accepted as legitimate (and even then there will be die-hard deniers) while a Republican has to win by just one vote to claim that it is a huge mandate.

One thing we are not likely to see is a repeat of January 6, 2021, for several reasons. One is that Harris is now the vice-president and thus the person who will preside over the session for the ratification of the Electoral College results so there is no chance that she will override the state certification of the electoral votes and the slates of electors based on spurious claims. Furthermore, after the last election fiasco, Congress passed a law stating that the vice-president’s role in the certification process is purely ceremonial and does not have even the semblance of the discretionary power that creepy Trump falsely claimed that Mike Pence had. The second reason is that the police and National Guard will be prepared for anything and since Joe Biden is the president, there will be no delay in calling them out to stop any violence. The third reason is that the MAGAnuts know that many of their fellow nuts who took part in the last riot have had their lives upended by serving time in prison. This may make them more leery of trying something similar. But we may have rioting by smaller groups around the country caused by furious MAGAnuts who have accepted as an article of faith that the only way that creepy Trump can lose is if there is cheating.