As I think I have remarked before, I now check with a site called electoral – vote.com each day to see what is happening in the shifting polls. This site has a useful chart called “tipping point state” which lists all of the states and the District of Columbia in order of the margin that Biden or Trump has in the latest polls with the Biden states listed first. The chart goes from dark blue (certainly for Biden) through lighter blue to barely blue to white (tied) and then to barely red, darker red, and deep red. For each candidate, the state that would tip him into victory is identified with a little pointing finger icon.
Today the situation is this: to win, Trump must first take
all of the dark red states, which only give him 78 electoral votes, then the
lighter red states which bring him up to 120 electoral votes, then the four
barely red states (Arkansas and Iowa, in which he leads by 2%, and North
Carolina and Texas in which he leads today by 1%). Then he must take all five
states that are barely blue (Georgia, Florida, New Hampshire, Nevada, and Ohio)
which Biden currently leads by anywhere from 1% to 4%, and then Trump must take
Arizona where Biden leads by 5% and finally, in order to get over 270, Trump
must take Pennsylvania where Biden leads by 5%. He must completely run that
table without missing a single state in order to get over 270.
As things now stand, unless there is some significant shift
in the polls, I simply don’t see how he is going to manage that.
7 comments:
That is encouraging. What continues to appall me, however,is the realization that about 40% of the adults in this country still support Trump. Hillary may have been wrong when she said that Trump supporters were the deplorables, but it's not too far from the mark to say that all the deplorables are Trump supporters. But my God, how could there be so many of them?
For statistically thoughtful modeling of the upcoming election, I recommend two additional sites:
FiveThiryEight: https://projects.fivethirtyeight.com/2020-election-forecast/
&
Economist 2020 Election forecast: https://projects.economist.com/us-2020-forecast/president
Nate Silver at 538.com also has a very good Twitter feed in which he makes a lot of micro-points concerning why 538.com gives Trump relatively good odds. The fact that a 2- or 3-point margin of error is still very much possible even with regard to the poll's "now" figures, Trump's EC advantage of about 2 points (as in 2016), and the fact that the time between now and election day has historically seen substantial polling changes (normally in the direction of making the race tighter) is behind just about all of Trump's now-24% chance of winning. 538 is more hedge-y than other models but that's not a bad thing!
No reason to get worked up over events you cannot control. We'll know the general outline of the votes in just a few weeks, even if the final tallies, certifications, and court judgments take a few more weeks.
I don't see Biden winning Arizona. The only Democratic presidential candidate to win Arizona in the past 70 years was Clinton over Dole in 1996 (in an election where Perot presented another option). Arizona is still a bastion of conservative politics in the Southwest. They barely elected Sinema to the Senate two years ago, and she's the most right-wing of Senate Democrats. I also don't think Biden has any real chance of winning anywhere in the South—he is simply not going to win Georgia, Florida, or North Carolina.
The only states that are really even in play are a few states in the Upper Midwest (Michigan, Wisconsin, Minnesota) and Pennsylvania. Biden should have an advantage in Pennsylvania because he was born there and has campaigned for years as being from Scranton, even though his family left when he was young and his entire political life has been in neighboring Delaware. (He's also had an association with UPenn since leaving the White House.) So Biden will probably win, based on narrowly winning Pennsylvania. If, for whatever reason, Biden loses Pennsylvania, Trump will win.
"[U]nless there is some significant shift in the polls, I simply don’t see how he is going to manage that."
The ups and downs of polls are a distraction. Each poll depends on how pollsters try to shape samples of respondents whom they believe will be representative of actual voters that will successfully cast votes.* So-called "likely voters" may or may not follow through. The forecasters also use polls with the assumption that there is no aspect of the election that is fundamentally different than in past elections. We have seen that that assumption may not have been valid in a number of the past cycles (Watergate scandal, 9/11, first black major-party nominee, first female major-party nominee, first nominee with absolutely no past government service, etc). No election since WWII has had anything like COVID-19.
Further, all the polls continue to show Trump has a chance of winning, despite Biden's apparent advantage. Don't forget that almost all the forecasters were wrong in their concluding that Clinton would win in 2016—they ignored the fact that the polls had been showing Trump had a chance. The nearly universal mistake is to assume that ten polls showing candidate A leading over candidate B 52%-48% equals ten polls showing candidate A leading over candidate B 100%-0%.
*for example, from recent Marist poll of Pennsylvania voters:
The samples were then combined and balanced to reflect the 2017 American Community Survey 5-year estimates for age, gender, income, race, and region. Assistance was provided by Luce Research for data collection. Results are statistically significant within ±3.6 percentage points. There are 1,039 registered voters. The results for this subset are statistically significant within ±3.8 percentage points. There are 771 likely voters defined by a probability turnout model which determines the likelihood respondents will participate in the November 2020 election based upon their chance of vote, interest in the election, and past election participation.
FWIW if it makes you feel any better, Allan Lichtman has been predicting Trump's defeat for more than a month now using a method that he claims ignores the polls.
Others comments have already mentioned 538, but I committed the sin of typing the note below before I had thoroughly reviewed them.
The top of the line election analysis (justified by a long track record) is from Nate Silver and company at 538 (which is the name of their website). They analyze the quality and track the record of each pollster and weigh them accordingly. They provide state by state adjusted polling averages along with modeled predictions for each state and the net result.
Their model currently gives Trump a 25% chance of winning the election.
All of that analysis depends on what is meant by "likely voters" - the standard definition or those likely to make it to a voting booth or mail in a ballot, and not the more relevant definition: those who are not "disappeared" and whose vote actually get counted.
Of course there is the generally rosy picture for Biden in the polls but that's just traditional indicators..
Nevertheless, a raft of bad news sets Trump back -- a rough week for Trump, maybe we say that Joe Biden resumes his status as the favorite.
The figures who guided Trump's political rise in 2016 have been much diminished, because of criminal indictment (Steve Bannon), criminal prosecution (Roger Stone), domestic ruptures (Kellyanne Conway).. Trump’s campaign does not have many strategists.
Donald Trump could still stage a comeback (again). Trump has clawed back a sizeable chunk of support. Right now the most likely outcome of the election is still that Mr Trump loses, but for all his volatility and incompetence, Trump is the default choice — even the safe choice — for a critical mass of white voters and business owners. Trump will be hard to beat even if he fights fair, which he almost certainly will not do. The Democrats are in for a desperate fight.
Post a Comment