Electoral College Predictions

[Finch]

It was actually Rasmussen that was least accurate.

That is probably true....But I do know Trafalgar didnt do very well in 2018 either....Both had repubs winning, a lot...Didn't happen.

 

[jaypatriot]

I'm from Texas and I can tell you, on this election cycle we Texan's are voting for Biden

You must be from Austin to say that with a level of confidence.

 

[Patriotic Voter]

It was actually Rasmussen that was least accurate.

Rasmussen is biased to the right. While almost every other poll favors the Democratic candidate, Rasmussen regularly picks Donald Trump.

 

[Perotista]

He is referring to the fallacy that polls under report republicans and over report democrats. Apparently the number of self-reported democrats and republicans never changes. This belief started with the 2012 election, with republicans believing only the "adjusted" polls that adjusted the polls more favorably for Romney. The "adjusted" polls have never been right yet, but that is not going to stop some people from believing them.

Okay, Nov 2016 party affiliation 31% Democrat, 27% Republican, 36% independent. Today or as 12 Aug 2020, 31% Democrat, 26% Republican, 41% independent.

Party Affiliation | Gallup Historical Trends

That isn't much change from pretty darn accurate finally polling from 2016. I use Gallup's figures when I do my forecasts. What Gallup doesn't do is provide the number of third party folks. But that is sort of irrelevant as you can figure that out by adding the three together to get the third party affiliation.

To me, these polls show me where things stand today, not where they'll be in November. For that I look for trends. If one goes back three months, looking for a trend one finds Biden has been between 49-51% with but a couple of days below 49%. Trump has been between 40-43%. We haven't seen the normal ups and down or any big swings for either candidate. Just a steady line more or less. That is the trend. Numbers pretty much staying the same over a three month period.

RealClearPolitics - Election 2020 - General Election: Trump vs. Biden

In 2016 over the same time period, we seen Trump go with 35 to 41% and take the lead over Clinton for a few days at the end of July. Hillary ranged from 40-44% and she regained her lead and led by 6.5 points mid August only to fall back again.

RealClearPolitics - Election 2016 - General Election: Trump vs. Clinton vs. Johnson vs. Stein

Now that is what that is. What I don't think anyone has notice, except me, Pat on the back, lol. Is that in 2016 during that time period is we had a large pool of 20% undecided voters. This year that pool of undecided's is down 8%. I think the difference is important. In 2016 Trump didn't need to convince a single Hillary supporter to come over to his side. He had that huge pool of 20% to play with and convince. This year, Trump will need to convince some Biden supporters to jump off his ship and board the Trump train. That will be much harder to do, not impossible, but much harder than convincing undecided's.

Also, Biden has played with the magic 50% number for the last three month. Hillary never rose over 44% until the last couple of weeks prior to the election and ended up at 46%. Her highest polling numbers. At least average.

He's right in the fact the polls under polled Trump's actual number by 3.9 points, but they also under polled Clinton final result by 2.8 points. Both major party candidates were numbers were over what was polled. But that is explainable due to the third party candidates which as history shows loses supporters the closer you get to an election. Both Johnson and Stein, voters were stating they'd vote for them, but once in the voting booth, decided to vote for the lesser of two evils among the two major party candidates.

Still the differences were within the margin of error.

 

[Perotista]

I don't get why people don't ignore general polls and just focus on the state ones.
That Trafalgar Group out of your state got 47 of 50 states correctly and hit the EC exactly.
I watch them like a hawk, and generally just add 2 to 3 points to Trump for the rest.

I look at the state numbers when doing my forecasts, but mostly stick with the national numbers on this site to respond to the different posts unless they're talking about a specific state.

 

[Perotista]

Disliking both candidates is almost the definition of independent. If they liked or hated either candidate, they would be on the fence. You will see many posts to the effect that opposition to Trump has grown, "now that they have seen what he's like." If it were true, it would show in the polls, but it doesn't.


Meh. There are ways to read this that do not track.

The vote for all other candidates went from 1% in 2012 to 6% in 2016. That 5% breaks down 0.84% increase for Jill Stein, 0.83% to write ins and other candidates, 0.62% to Evan McMullin, 2.73% to Gary Johnson. The vote for McMullin and the increase for Johnson represents 3.95 million votes, or over 3.3% of the entire vote.


Dislike is one thing. Respect is another.


Turnout has always been an issue, particularly for Democrats.


Trump does not need to do it again. He has almost 4 million defectors he can bring back. Also, his support in 2016 was somewhat timid. Now they know it can be done and they have an ax to grind. Democrats and the media have not been polite the last four years.

We look at number differently, so we'll wait until November. But all the numbers we have to day reflect today, not November.

 

[Perotista]

Not sure where that 26million number is coming from...2018 midterm 59 million dems voted to 50 million repubs...My premise is the fact that the Dems had very high turn out in 2018 and there's no indication that it is slowing down for Biden...And I disagreed that Trump will be getting all the 3rd party voters, he will be lucky to get half. However. I did agree with your premise that trump may get more votes in 2020 but only a couple more million votes or so...If I was going to make a prediction I would say Biden will edge Trump 69 mill to 65...Biden winning electoral by slim margin...But we shall see.

You're correct on Florida. I went back to RCP averages and looked again. Well sort of correct.

RealClearPolitics - Election 2020 - Florida: Trump vs. Biden

They show 5 polls where Trump leads in 1, one a tie and Biden in the other three. RCP averages have Biden ahead by 1.3 points in Florida. Now averaging all the polls, Biden has led from early April through today.

 

[ludin]

Did you actually pay attention to the polls in 2016 or just took what you heard and ran with it? Here's all the latest polls from 2016.

RealClearPolitics - Election 2016 - General Election: Trump vs. Clinton vs. Johnson vs. Stein

RCP averages had Clinton winning the popular vote by 3.3 points, in reality she won it by 2.1 points. Now considering these polls have an average margin of error of plus or minus 3 points. What those polls were telling us Hillary's lead was somewhere between 0.3 of a point to 6.3. Her finally tally of 2.1 falls well within that margin of error.

That is the number 1 issue with the polls. We do not have an national election. We have 50 separate elections based on the electorial college. The popular vote is a meaningless statistic.

Now take a look at all the polls on the list, almost all of them were within the MOE. All of them with the exception of IBD which had Trump winning the popular vote. You can't get any better results than that. 10 of 11 polls within the MOE. That's quite a record, a very good record for the polls if one knows how to read them.

The polls were wrong. Why? because they were polling using the wrong methods and statistics.

Polling data is some of the most inaccurate statistics out there.
You have sampling bias, you have people not talking you have 1000 other variables that make them unreliable.

Then you are taking a measure of a statistic that is 100% meaningless.
The only vote that matters is the electoral college vote.

That is where the polls go south.

The only semi-accurate polling is what is going on in each individual state.
WHy people keep referring to flawed data and data that doesn't matter is beyond me.

If you really want to see where the election is then you need to look at individual polls in each state and compare them.
You can't look at a national poll because national polls don't matter. Individual state matter.

 

[Jay59]

We look at number differently, so we'll wait until November. But all the numbers we have to day reflect today, not November.

Actually, we don't. It's just a hard support vs soft support issue. Trump has 95 million potential votes in his hard support, so it's a matter of turning them out. 75% is 71 million. 60% is 57 million which makes for a race. Anyone that thinks Trump will not get at least 60% of his base turning out to vote needs counseling, because they are highly motivated.

The Democrats likely maxed the anti-Trump vote in 2018. It's doubtful they manage that again, but 50 - 55 million actual votes is doable. Assuming Trump gets the bottom end of his range and Biden approachs 2018, then it comes down to the late deciding voters, who dislike both sides. In that race, Trump is the guy they trust with the economy.

 

[Finch]

That is the number 1 issue with the polls. We do not have an national election. We have 50 separate elections based on the electorial college. The popular vote is a meaningless statistic.


The polls were wrong. Why? because they were polling using the wrong methods and statistics.

Polling data is some of the most inaccurate statistics out there.
You have sampling bias, you have people not talking you have 1000 other variables that make them unreliable.

Then you are taking a measure of a statistic that is 100% meaningless.
The only vote that matters is the electoral college vote.

That is where the polls go south.

The only semi-accurate polling is what is going on in each individual state.
WHy people keep referring to flawed data and data that doesn't matter is beyond me.

If you really want to see where the election is then you need to look at individual polls in each state and compare them.
You can't look at a national poll because national polls don't matter. Individual state matter.

1) You are correct, we are the only democracy that doesn't recognize a popular vote. Which is crazy to me, but that is another debate.

2) From what I understand polls have changed their polling methods to poll individual states more closely (along with a national poll.) They still do a national poll because it would literally be impossible for a candidate to win the electoral college and lose by 10+ percentage points in the national poll, so there is somewhat of a correlation that can be studied. Trump will need to be with in 3 to 4 points in the popular vote to win....

 

[Ethel2]

Actually, we don't. It's just a hard support vs soft support issue. Trump has 95 million potential votes in his hard support, so it's a matter of turning them out. 75% is 71 million. 60% is 57 million which makes for a race. Anyone that thinks Trump will not get at least 60% of his base turning out to vote needs counseling, because they are highly motivated.

The Democrats likely maxed the anti-Trump vote in 2018. It's doubtful they manage that again, but 50 - 55 million actual votes is doable. Assuming Trump gets the bottom end of his range and Biden approachs 2018, then it comes down to the late deciding voters, who dislike both sides. In that race, Trump is the guy they trust with the economy.

There was no pandemic in 2018, and the plague is THE most important issue to voters this year. Given the way the donkey has mismanaged the plague I think that this year the democrats will turn out in droves to vote, either in person or by mail. This election will be a referendum on how the donkey has mismanaged the pandemic.

There are probably very few late deciding voters and since the economy has tanked (yes-the economy has tanked for the average person out there although the stock market has done OK) they will vote against the buffoon who did this to them.

 

[Linc]

I look at the state numbers when doing my forecasts, but mostly stick with the national numbers on this site to respond to the different posts unless they're talking about a specific state.

At 48.14% Clinton to 46.04% Trump, I’d say few expect 5.82% to go 3rd party in 2020. With people locked in on the POTUS race, down-ballot races this time will actually have an effect on bringing out more voters, imo.

The swing states in large part have both Senate and POTUS races that are too close to call, along with dozens of House races. (Though you do it — tyvm) Except for a few states, the days of ticket-splitting seem to be over.

As you know, I’m all for expanding the map, both ways.

 

[ludin]

1) You are correct, we are the only democracy that doesn't recognize a popular vote. Which is crazy to me, but that is another debate.

This is 100% not correct. We recognized the popular vote just not a national one. We recognized 50 individual popular votes for president.
Which i think is a better measure than 1 national vote. It means that the person running for president has to appeal to a wide range of people across the country.

2) From what I understand polls have changed their polling methods to poll individual states more closely (along with a national poll.) They still do a national poll because it would literally be impossible for a candidate to win the electoral college and lose by 10+ percentage points in the national poll, so there is somewhat of a correlation that can be studied. Trump will need to be with in 3 to 4 points in the popular vote to win....

[/QUOTE]

They polls are still not as accurate. even at the state level. the reason is the polls is only as accurate as the person collecting the data.
There are over sampling sizes sampling bias's all over the place.

Polls have become more political than they have objective.
which is why anymore I simply don't get my excitement level going over what the last poll says.

 

[NWRatCon]

1) You are correct, we are the only democracy that doesn't recognize a popular vote. Which is crazy to me, but that is another debate.

2) From what I understand polls have changed their polling methods to poll individual states more closely (along with a national poll.) They still do a national poll because it would literally be impossible for a candidate to win the electoral college and lose by 10+ percentage points in the national poll, so there is somewhat of a correlation that can be studied. Trump will need to be with in 3 to 4 points in the popular vote to win....

I completely agree, although I would state it somewhat differently: There is a 3-4% bias in the EC toward the GOP, so Biden needs to win the popular vote by 4+% in order to win the EC.

Right now, based upon the most reliable polling I can find (State polling is notoriously inaccurate), I have Biden at 289 electoral votes - keeping Florida and North Carolina as toss-ups. I think Biden wins Arizona, Wisconsin, Michigan and Pennsylvania. With Florida, he's at 318. If, miraculously, North Carolina swings blue, 333. I am not including either Nebraska's or Maine's 2nd Districts, because they are just too small to make reliable predictions. Trump's best chance at winning is taking Florida and Pennsylvania away from Biden, which would make it a close election - potentially even tied. I don't think Trump can pull Arizona and Nevada back from Biden, although Biden's margin is only about 5 points there.

I don't use the RCP average, as Perotista knows, because they take any old poll, and some are just crap. That's why I rely on FiveThirtyEight, because they analyze the reliability of pollsters and eliminate the chaff. Then I go to 270toWin to chart my predictions.

 

[Ethel2]

I completely agree, although I would state it somewhat differently: There is a 3-4% bias in the EC toward the GOP, so Biden needs to win the popular vote by 4+% in order to win the EC.

Right now, based upon the most reliable polling I can find (State polling is notoriously inaccurate), I have Biden at 289 electoral votes - keeping Florida and North Carolina as toss-ups. I think Biden wins Arizona, Wisconsin, Michigan and Pennsylvania. With Florida, he's at 318. If, miraculously, North Carolina swings blue, 333. I am not including either Nebraska's or Maine's 2nd Districts, because they are just to small to make reliable predictions. Trump's best chance at winning is taking Florida and Pennsylvania away from Biden, which would make it a close election - potentially even tied. I don't think Trump can pull Arizona and Nevada back from Biden, although Biden's margin is only about 5 points there.

I don't use the RCP average, as Perotista knows, because they take any old poll, and some are just crap. That's why I rely on FiveThirtyEight, because they analyze the reliability of pollsters and eliminate the chaff. Then I go to 270toWin to chart my predictions.

I think that analysis is spot on. I also use fivethirtyeight. I can't see the donkey winning Pennsylvania again, nor Michigan or Minnesota. Arizona is more iffy. The key state, again, is Florida. Without Florida the donkey cannot win the election. Mike Bloomberg has just pledged 100M to the Biden campaign in Florida. Hopefully that will help swing it blue.

The bottom line is that however Biden wins it, he must win it. We simply cannot afford another four years of the donkey screwing around with the pandemic. He will kill a million citizens before he is done if he is re-elected. You know, before the pandemic I could almost hold my nose and tolerate him; now, because he is responsible for the deaths of hundreds of thousands of Americans, there is some real hate for him. I do not tolerate mass murderers, and he falls into that category, for me.

 

[BlueTex]

I think that analysis is spot on. I also use fivethirtyeight. I can't see the donkey winning Pennsylvania again, nor Michigan or Minnesota. Arizona is more iffy. The key state, again, is Florida. Without Florida the donkey cannot win the election. Mike Bloomberg has just pledged 100M to the Biden campaign in Florida. Hopefully that will help swing it blue.

The bottom line is that however Biden wins it, he must win it. We simply cannot afford another four years of the donkey screwing around with the pandemic. He will kill a million citizens before he is done if he is re-elected. You know, before the pandemic I could almost hold my nose and tolerate him; now, because he is responsible for the deaths of hundreds of thousands of Americans, there is some real hate for him. I do not tolerate mass murderers, and he falls into that category, for me.

What effect do you think all the absentee voting will have on polling this year? A significant number of voters will be casting their votes weeks before Nov 3rd.

 

[Ethel2]

What effect do you think all the absentee voting will have on polling this year? A significant number of voters will be casting their votes weeks before Nov 3rd.

I don't think absentee voting will affect the polling, but it might affect the results since more democrats will be voting by mail than republicans. Polls simply ask which candidate is preferred, not how someone plans to cast their ballot.

Am I missing something?

 

[BlueTex]

I don't think absentee voting will affect the polling, but it might affect the results since more democrats will be voting by mail than republicans. Polls simply ask which candidate is preferred, not how someone plans to cast their ballot.

Am I missing something?

I'm just wondering if the pollsters are accounting for who has already cast a vote and if that could impact the polls.

 

[Jay59]

There was no pandemic in 2018, and the plague is THE most important issue to voters this year.

Only indirectly. The plague's impact on the economy is the most important.

Given the way the donkey has mismanaged the plague

That asking to much. Trump did not mismanage COVID, though some Governors did.

I think that this year the democrats will turn out in droves to vote, either in person or by mail. This election will be a referendum on how the donkey has mismanaged the pandemic.

Do you have a reason other than the usual Orange Man Bad

There are probably very few late deciding voters

It's true that there are fewer than usual, but it's still in the ballpark of 25 - 30 million.

and since the economy has tanked (yes-the economy has tanked for the average person out there although the stock market has done OK) they will vote against the buffoon who did this to them.

The average person does not blame the president for the economy tanking. They blame COVID protocols.

One interesting factoid is that the Thursday night NFL game was the most watched sporting event since the Super Bowl. Yet, there was a thread about how the viewership of the NFL was down 30% from last year. Not having sports to watch is a very tangible impact to a huge number of people. Suddenly, there is a buffet of sports. NHL and NBA playoffs, baseball pennant races and the opening of the NFL schedule. This does not have a direct impact on the vast majority of the country, but it does a great deal to lift spirits.

 

[Ethel2]

Only indirectly. The plague's impact on the economy is the most important.


That asking to much. Trump did not mismanage COVID, though some Governors did.


Do you have a reason other than the usual Orange Man Bad


It's true that there are fewer than usual, but it's still in the ballpark of 25 - 30 million.


The average person does not blame the president for the economy tanking. They blame COVID protocols.

One interesting factoid is that the Thursday night NFL game was the most watched sporting event since the Super Bowl. Yet, there was a thread about how the viewership of the NFL was down 30% from last year. Not having sports to watch is a very tangible impact to a huge number of people. Suddenly, there is a buffet of sports. NHL and NBA playoffs, baseball pennant races and the opening of the NFL schedule. This does not have a direct impact on the vast majority of the country, but it does a great deal to lift spirits.

The donkey has TOTALLY mismanaged the pandemic-from the start. He does not get ALL of the blame: China, the WHO, the CDC (who screwed up the initial testing materials by allowing them to be contaminated), and the governors also get some of the blame. Cuomo, for example, failed to protect nursing home residents who were repatriated to their facilities. Thats his fault. Some of the governors reopened their states too early (at the prodding of the donkey), and they should have resisted doing that.

But the Orange Buffoon KNEW about how serious this could be and failed to address it properly. Although he does not have the authority to close down states he has a lot of leverage to influence the governors to do so. Had the states closed down just TWO WEEKS earlier 90% of the deaths could have been prevented in this country. That is what SHOULD have happened. I saw it coming and bought a bunch of masks and hand sanitizer back in Feb. If I can see it coming, why didn't the white house donkey see it coming? He was late enacting the DPA. He modeled (and continues to model) crappy behavior by not wearing a mask and encouraging others to follow his lead. He encourages the use of unproven and dangerous methods to treat this to the detriment of the citizens of the country he is sworn to protect. He continues to downplay the seriousness of the pandemic, failing to listen to the advice of those who know more than he does about it. This country has about 5% of the world's population and has 25% of the deaths; you cannot sugar coat that. There is no national plan to deal with this. Sound like total mismanagement to you? It does to me.
Covid has caused the economy to tank for the average American, and thats on the donkey too. Although neither of us can speak for the "average person", my guess is that the "average person" blames the president for the current state of affairs, and that is reflected in the polls.
Of all these facts the most important is that the donkey KNEW this was coming and failed to react. The primary job of the president is to protect the health and safety of the citizens of this country. T Rump failed miserably in that regard, and for that reason he should be prosecuted, not re-elected.

 

[Perotista]

That is the number 1 issue with the polls. We do not have an national election. We have 50 separate elections based on the electorial college. The popular vote is a meaningless statistic.


The polls were wrong. Why? because they were polling using the wrong methods and statistics.

Polling data is some of the most inaccurate statistics out there.
You have sampling bias, you have people not talking you have 1000 other variables that make them unreliable.

Then you are taking a measure of a statistic that is 100% meaningless.
The only vote that matters is the electoral college vote.

That is where the polls go south.

The only semi-accurate polling is what is going on in each individual state.
WHy people keep referring to flawed data and data that doesn't matter is beyond me.

If you really want to see where the election is then you need to look at individual polls in each state and compare them.
You can't look at a national poll because national polls don't matter. Individual state matter.

Have it your way, but being 1.1 points within the actual results, I think is pretty darn good. Now I agree with your last sentence, state polls are more important than the national one. Although the national one can point out trends nationwide where state polls can't

 

[Perotista]

Actually, we don't. It's just a hard support vs soft support issue. Trump has 95 million potential votes in his hard support, so it's a matter of turning them out. 75% is 71 million. 60% is 57 million which makes for a race. Anyone that thinks Trump will not get at least 60% of his base turning out to vote needs counseling, because they are highly motivated.

The Democrats likely maxed the anti-Trump vote in 2018. It's doubtful they manage that again, but 50 - 55 million actual votes is doable. Assuming Trump gets the bottom end of his range and Biden approachs 2018, then it comes down to the late deciding voters, who dislike both sides. In that race, Trump is the guy they trust with the economy.

Everything we look at today is as of today. I think we need to keep that in focus. If the election was today, Biden wins and wins easily. But it isn't. You had a huge 25% of all Americans who disliked both Trump and Clinton which included 54% of independents. This is the main reason I call 2016 the anti election. People voting for the candidate they least wanted to lose, not win, but least wanted to lose.

Besides those numbers, on election day 2016 60% of all Americans viewed Trump negatively vs. 58% who viewed Clinton negatively. There really wasn't that many, a minority of Americans who wanted one or the other to win. So do some of those numbers carry over to 2020, Trump's numbers? Biden isn't nearly as disliked as Clinton was. So I guess we'll see.

 

[Perotista]

At 48.14% Clinton to 46.04% Trump, I’d say few expect 5.82% to go 3rd party in 2020. With people locked in on the POTUS race, down-ballot races this time will actually have an effect on bringing out more voters, imo.

The swing states in large part have both Senate and POTUS races that are too close to call, along with dozens of House races. (Though you do it — tyvm) Except for a few states, the days of ticket-splitting seem to be over.

As you know, I’m all for expanding the map, both ways.

It's still early, that we have to keep in mind. But I fully expect Pennsylvania, Wisconsin and Michigan to return to the Democratic fold. Arizona looks like a lost cause to both Trump and McSally. Ohio, North Carolina, Florida are t traditional tight swing states which could go either way. Georgia and Texas also remain close with Trump ahead.

We did see some tightening of the race in Wisconsin, Minnesota and Michigan over the last couple of weeks, but in the last few days, Biden has regain his lead. Kenosha probably the cause. But Trump wasn't and isn't political savvy enough, have the right amount of finesse to play the law and order card properly and effectively. He's like a bull in a China shop, always over doing things as he charges ahead without paying attention to what else is surrounding him. The violent protest could have been god sent to Trump, except he doesn't know how to handle it properly to get folks to switch from Biden to him.

 

[NWRatCon]

I'm just wondering if the pollsters are accounting for who has already cast a vote and if that could impact the polls.

That's a very interesting question. In the past, some pollsters asked "have you voted?" This year, that is going to be a high number. Can't be "undecided" if you've already voted.

I think with all the churn about absenteeism, that voters, particularly Dems, will be voting early. I won't be surprised if those mailed ballots show early results in a lot of States. I think it will be Trump, not Biden, who will be sweating the late returns.

 

[Ethel2]

That's a very interesting question. In the past, some pollsters asked "have you voted?" This year, that is going to be a high number. Can't be "undecided" if you've already voted.

I think with all the churn about absenteeism, that voters, particularly Dems, will be voting early. I won't be surprised if those mailed ballots show early results in a lot of States. I think it will be Trump, not Biden, who will be sweating the late returns.

It would seem that if someone has already voted then they would poll for one candidate or another. All it would do is make the poll more reliable since their vote is cast-but it won't change anything.