09-11-2020, 10:36 AM
(09-11-2020, 05:33 AM)Eric the Green Wrote:(09-11-2020, 12:21 AM)pbrower2a Wrote: The point is to predict the likelihood of a win by one or the other at two months out. That means that polls from the time of the Republican National Convention or earlier are excluded in this effort. The point is to show that time matters greatly in a campaign.
Calculation may be messy soon because of early voting (including much absentee voting). Some people will have voted in September. Political realities involving COVID-19 may have forced the voting to be more by mail... and people may be seeing more in-your-face political advertising than is usual. Not many times can one say that two Presidential elections are really alike, so it is hard to say which one is the most unique. This one is weird for having two opponents in their seventies, and for having the definitive outsider (Trump) facing the most blatant insider that there has ever been.
It's best to have more than one poll though. I think fivethirtyeight does a good job of evaluating polls for their accuracy, recency and partisan lean. Using more than one method is a good idea too.
I would say Trump is an insider, because he's part of the 1% oligarchy of the country and represents their interests. Biden not having a college degree and a representative of the working class is not an insider, by that measure. But he's been inside Washington politics for almost 50 years.
Yet, I still think of him as that young 30 year-old boomer-cusper elected thanks to the McGovern campaign that I worked on, so to me he's the young guard representing the younger generation, the boomers. I know we aren't that anymore, the millennials take precedence now; but at heart and in a certain way, we still are. Many of us blue boomers are young in mind and heart, while Xers like Classic are lost in the past and represent sclerotic minds. Blue boomers form a natural alliance with the up and coming millennials.
All true, as usual. At this point I am going toward probabilistic models as opposed to gaps or edges. If the 99% range of possibility is from a Trump victory by 1% to Trump by 17% , then we need pay little attention to any possibility that Trump loses even if, in the end, Trump might win the state by 2%. But consider how I saw the election of 2020 beginning on Election Day 2016. If he could have a 2010-style or 2014-style electorate, then he would get re-elected, assuming that he would be a normal President who fully satisfies his base and offers a few scraps of goodies to those who did not vote for him the first time. As with Clinton, Dubya, and Obama (go back far enough, and Eisenhower fits the pattern) he would win again with an electoral map similar to that of the one in which he got elected because the political climate would not change. Figure that Trump came close to winning Minnesota by making promises of infrastructure, and its iron miners ordinarily expect big construction and reconstruction projects to devour iron ore. Politicians win by giving people what they promise. Trump's solution was simply to add consumer costs to enhance profits for gougers or to create profits for monopolistic gougers (add tolls!), so he disappointed there. But do as you promise or even try to do as you promise, and you win over votes that swung the last time and add to that swing by getting people who get the message the next time.
Approval and disapproval polls suggest whether people like the results. Trump got into trouble with trade wars that would likely hurt American farmers. Farmers may not be as huge a constituency now as they were in prior times, but in close elections any constituency can flip a state or two. The heavily-rural states from North Dakota to Kansas are usually reliable R states in Presidential elections, and none has more than six electoral votes. But add those four states and you have 17 electoral votes, and for Trump, losing those four would be one electoral vote more than failing to hold onto Michigan. Add Iowa, the definitive agribusiness state, and one has 23. With farmers, many people rely heavily upon farm activity for their living from dealers in farm machinery and lenders to farmers to restaurant owners in small towns and wait staff who would have no income without farmers doing well. So I imagine that Trump playing it straight with infrastructure, not promoting a trade war through tariffs (OK, that did recede), and in general not acting erratically. People would get comfortable with him as President and he would win with an electoral map similar to that of 2016.
That is not happening.
But could he lose with a similar electoral map? Sure. The Millennial Generation is about 20% more D than R and the parts of the electorate over 55 who hemorrhage votes from the electorate by dying off (early-wave X, Boomers, and the Silent are all about 5% more Republican than Democratic in a nomal year. Figure that the average voter has at most a sixty year career as such, then the younger replacements in the electorate replace the older ones at a rate of about 1.6% per year and swing the difference in the electorate about 25%... which means that through that effect alone, Democrats should fare about 1.6% better in a Presidential election in 2020 than in 2016 if nothing else really changes. That would be enough to swing four states (Michigan, Pennsylvania, Wisconsin, and Florida) away from Trump. Sure, if Florida gets older population moving in such might offset that, but if Florida doesn't get Trump, then maybe Arizona or North Carolina does. On the other hand, a similar map could be that Trump loses Michigan but wins Minnesota and New Hampshire instead, which is a wash from 2016.
The distinction between approval and disapproval estimates mass assessments of performance including a response to any "clear and present danger" and avoiding the political disaster of offending parts of the electorate that can shift against him. Some people care about whether a climate of corruption prevails or does not... OK, they might still vote for Romney even if Obama acts with integrity because they expect much the same level of integrity. Maybe avoiding scandal does not get one re-elected, but getting ensnared can get one defeated. If the binary choice were between Mitt Romney and Donald Trump, then I would vote for Mitt Romney because I hate cronyism and corruption in principle and because those vices in government do me no good. This time we get Joe Biden, who has the weight of Obama corruption weighing him down (OK, that is more like a buoy than like an anchor). The "clear and present danger" is a dangerous respiratory infection that so far kills on the scale of a bungled war. On that I need say little more. I am not going to predict whether what Trump has said about Americans in the Armed Services, past and present, will do to his vote. It will not help him.
Approval and disapproval numbers have had no meaningful trend for about half a year, so they are likely stable enough to suggest that Trump is in electoral trouble. Shaking such off in a year is far more possible than shaking off such in a month. The 4000-foot increase in elevation from east to west s easier to between Omaha and the Wyoming state line is far easier on I-80 than is the similar rise from near sea level to the near-summit of Mount Diablo even if the rise is similar... it's the angle that is steeper. Trump had a vehicular climb analogous to the ascent of I-80 in Nebraska a year ago. It is now more like the ascent of Mount Diablo.
What I await is one pollster's fifty-state results which come from sources different from telephone polling such as interactive results from the Internet. A few years ago such would have been unorthodox. It may be reliance upon landlines that is becoming less reliable in polling. That pollster's methods do not require averaging of polls that have different assumptions for the states. Remember: a poll for a 2010-style or 2014-style electorate would suggest that Trump is doing very well, and one that has a 2006-style or 2018-style electorate suggests that Trump would do badly.
Who votes matters greatly, and the pollster who gets the election closest to being right is the one who has an idea of who will vote. A compiler of polls can combine polls from someone who presumes a 2010-style electorate (like Rasmussen) and one that expects a 2018-style electorate (like PPP or Quinnipiac). Oddly there are some foreign pollsters, especially Canadian and British, who seem promising due to disinterest. (For those who don;t know what the word disinterest means it expresses the idea that someone making a determination has more interest in the process than in the result and will let the facts make the decision. One wants a judge with disinterest in a criminal or civil case, that is, not having a personal interest in the case and not being prejudiced. If the suit is between a contractor and a government agency I don't want the judge to have a known family relationship to the contractor or to anyone working in any capacity for that government agency. Judges can recuse themselves from a case when they find that something has compromised the judge's objectivity, such as an offer of a bribe from a defendant.
Disinterest by me in this election? No. I am as big a partisan hack as there is. How this one goes may shape much in my life, including my longevity and whether I can find any satisfying meaning or purpose in life. I'm trying to avoid saying that Trump will lose because he is so awful that no sane person could ever vote for him. People did vote for him in adequate numbers in critical places that he won, and some changes of assumptions would mandate my assessment that he would be re-elected. The last thing that he has ever wanted to be is a loser -- and that stares him in the face if he does not have sycophants telling him how great he is and how much America loves him. He will lie, he will cheat, and he will steal to have a chance to win; that is predictable from his character. He may have the most troubled Presidency since Herbert Hoover, but at least Hoover had some integrity and had no risk of a prison term. His good buddy Putin may have coached him on how to get re-elected through means abnormal in America but normal in, for example, Belarus.
The ideal subject of totalitarian rule is not the convinced Nazi or the dedicated Communist but instead the people for whom the distinction between fact and fiction, true and false, no longer exists -- Hannah Arendt.