Election Prediction

Still though, getting 48 States out of 50 right is pretty good.

Also interesting to note, Intrade gives Rasmussen an A- grade (91% accuracy):


And Rasmussen has Romney winning the election :confused:

You sure have a funny way of looking at the world.

Due to the American electoral system Rasmussen actually has Obama winning, or with a better chance to win to be more accurate, despite the recent national poll.

http://www.rasmussenreports.com/public_content/archive/2012_electoral_college_scoreboard

If you track Rasmussen vs other polls they tend to be R +3.5 or so.


I didn't know Rasmussen gave those rankings. Do you have a link?
 
I don't believe he is at all correct on how the Times uses its data. Not even close. In fact the NYT has full transparency on its poll procedure and analysis and it does not do what Dick says it does:

http://www.nytimes.com/2012/10/31/us/politics/how-the-polls-were-conducted.html?_r=0

Even if he was, the 538 model uses hundreds and thousands of polls, not a single poll.

So do you believe Dick is correct and Romney will win by a landslide?

I don't think Morris was questioning the poll's transparency.

It seemed to me he was questioning the way the weighting was done. The Times does acknowledge the weighting. The issue Morris has raised is whether it was done properly.

As for my prediction? I believe Romney will win but I'm not so sure about a landslide.
 
You sure have a funny way of looking at the world.

Due to the American electoral system Rasmussen actually has Obama winning, or with a better chance to win to be more accurate, despite the recent national poll
Sorry, I meant to say Rasmussen has Romney winning the popular vote, not electorals. And only 4 times in history has a POTUS won popular vote, and not the election (last time was Gore in 2004)

See here: http://en.wikipedia.org/wiki/List_of_United_States_presidential_elections_by_popular_vote_margin
 
I don't think Morris was questioning the poll's transparency.

It seemed to me he was questioning the way the weighting was done. The Times does acknowledge the weighting. The issue Morris has raised is whether it was done properly.

As for my prediction? I believe Romney will win but I'm not so sure about a landslide.

Except the Times does not acknowledge the weighting in the way Morris describes it.

So if you don't think it will be a landslide, then you must think Morris is wrong.
 
I don't think Morris was questioning the poll's transparency.

It seemed to me he was questioning the way the weighting was done. The Times does acknowledge the weighting. The issue Morris has raised is whether it was done properly.

As for my prediction? I believe Romney will win but I'm not so sure about a landslide
Here's my conclusion after we see who wins the election:

1. If Obama wins easily it restores my faith in polls somewhat.
2. If Romney loses Dick Morris should be taken off the air permanently.
3. If Romney wins I lost all faith in polls and/or the media. And Morris can stay
 
Sorry, I meant to say Rasmussen has Romney winning the popular vote, not electorals. And only 4 times in history has a POTUS won popular vote, and not the election (last time was Gore in 2004)

See here: http://en.wikipedia.org/wiki/List_of_United_States_presidential_elections_by_popular_vote_margin

You are talking about a daily trackin poll rasmussen runs. A daily result. It is transitory data at best. But the only thing that actually matters in a presidential election is electoral votes. And all of the main tracking sources and conglomeration sites that I follow have BO winning the electoral college.
 
Here's my conclusion after we see who wins the election:

1. If Obama wins easily it restores my faith in polls somewhat.
2. If Romney loses Dick Morris should be taken off the air permanently.
3. If Romney wins I lost all faith in polls and/or the media. And Morris can stay

I think you need some more options. The polls are not calling for Obama to win easily, in fact that are calling it super close.

If Romney wins big, and the data behind his win fits Morris' analysis than he will look pretty damned good.
 
You are talking about a daily trackin poll rasmussen runs. A daily result
No, Rasmussen has had Romney up by a few percentage points in popular vote for a few weeks now

But the only thing that actually matters in a presidential election is electoral votes. And all of the main tracking sources and conglomeration sites that I follow have BO winning the electoral college
I dont disagree. But my point still stands, only 4 times in history has a candidate won popular vote, and not the overall election
 
the problem is..there are just too many polls. all with conflicting data


If you think about it...more polls mean more data. And thus people who combine and analyse the data should get more reliable predictions than any one poll. Very simple stuff.

In this case data really cannot conflict, as long as the pollsters are honest.
 
538 actually does a far more sophisticated analysis than yours and their numbers are pretty clear. And the vast majority of pollsters don't actually report the way you say they do. There is no "if" statement like you propose. In fact they report their numbers categorized as "likely voters" or "registered voters" and ask questions about how likely the people are going to vote, and how strongly they feel about their convictions.

I can't remember the last time I saw a poll report the way you suggest they do. Where do you get this stuff?

You're funny. Incorrect, but at least you're funny.

If you don't see the details of the polls, you're probably not looking in the right places.
 
If you think about it...more polls mean more data. And thus people who combine and analyse the data should get more reliable predictions than any one poll. Very simple stuff.

In this case data really cannot conflict, as long as the pollsters are honest.

Let's try this again, shall we?

A pollster doesn't have to be dishonest to suggest (as some have done today) that 94% of Registered Voters will turn out on Election Day and actually vote ("Likely Voters"). But there is a huge difference between honesty and their data model being accurate.

Let's take a poll released today in Nevada, showing Obama leading by 4 with a turnout advantage to Democrats of D+10. But in 2010, the actual turnout advantage was D+2. In 2008, D+8.

That same poll shows Romney leading Independents by 7.

Now, if you really truly believe that Democrats are more excited about going out and voting than they were in 2008, and Republicans much less excited, then you can believe that poll. But you have to ask yourself why Romney wins Independents by 7 and it's still Obama who has Democrats excited. The numbers don't add upl

If you, like most people, don't believe that Democrats are more excited this year than they were two years ago, then you look at that poll and see that Romney will win Nevada.

But in this case, the pollster wanted to show Obama winning, so, knowing that there are more registered Democrats than Republicans, found that by saying 94% of RVs are LVs, could get the desired result.
 
Let's try this again, shall we?

A pollster doesn't have to be dishonest to suggest (as some have done today) that 94% of Registered Voters will turn out on Election Day and actually vote ("Likely Voters"). But there is a huge difference between honesty and their data model being accurate.

Let's take a poll released today in Nevada, showing Obama leading by 4 with a turnout advantage to Democrats of D+10. But in 2010, the actual turnout advantage was D+2. In 2008, D+8.

That same poll shows Romney leading Independents by 7.

Now, if you really truly believe that Democrats are more excited about going out and voting than they were in 2008, and Republicans much less excited, then you can believe that poll. But you have to ask yourself why Romney wins Independents by 7 and it's still Obama who has Democrats excited. The numbers don't add upl

If you, like most people, don't believe that Democrats are more excited this year than they were two years ago, then you look at that poll and see that Romney will win Nevada.

But in this case, the pollster wanted to show Obama winning, so, knowing that there are more registered Democrats than Republicans, found that by saying 94% of RVs are LVs, could get the desired result.

If you would like us to discuss a specific poll you should at least link to it. Sources are useful.
 
You're funny. Incorrect, but at least you're funny.

If you don't see the details of the polls, you're probably not looking in the right places.

So you really believe all the pollsters are biased or dishonest and only you and a few key republicans have figured them out and analyse the data better than the professionals.

I really do need to get an invite to that cabal meeting.

And does this mean that Romney is about to win in a landslide? Because that is the conclusion you are advocating.
 
Let's try this again, shall we?

A pollster doesn't have to be dishonest to suggest (as some have done today) that 94% of Registered Voters will turn out on Election Day and actually vote ("Likely Voters"). But there is a huge difference between honesty and their data model being accurate.

Let's take a poll released today in Nevada, showing Obama leading by 4 with a turnout advantage to Democrats of D+10. But in 2010, the actual turnout advantage was D+2. In 2008, D+8.

That same poll shows Romney leading Independents by 7.

Now, if you really truly believe that Democrats are more excited about going out and voting than they were in 2008, and Republicans much less excited, then you can believe that poll. But you have to ask yourself why Romney wins Independents by 7 and it's still Obama who has Democrats excited. The numbers don't add upl

If you, like most people, don't believe that Democrats are more excited this year than they were two years ago, then you look at that poll and see that Romney will win Nevada.

But in this case, the pollster wanted to show Obama winning, so, knowing that there are more registered Democrats than Republicans, found that by saying 94% of RVs are LVs, could get the desired result.

Don't bother posting the poll I found it. And you are wrong.

Here is the raw data from the poll that shows Obama +4%, the numbers add up perfectly and there is no place where they make an adjustments based on prior turn out rates. None.

http://www.lvrj.com/hottopics/politics/polls/oct_2012_polls_president_110112.html

And you obviously didn't read much about the poll or the data because they whole issue is that Obama has a big lead amongst those who have already voted and Romney has to swing way back to win Nevada. There can obviously be no turnout amongst people who have already voted!
 
Don't bother posting the poll I found it. And you are wrong.

Here is the raw data from the poll that shows Obama +4%, the numbers add up perfectly and there is no place where they make an adjustments based on prior turn out rates. None.

http://www.lvrj.com/hottopics/politics/polls/oct_2012_polls_president_110112.html

And you obviously didn't read much about the poll or the data because they whole issue is that Obama has a big lead amongst those who have already voted and Romney has to swing way back to win Nevada. There can obviously be no turnout amongst people who have already voted!

You really are hilarious. Do some basic research. You've demonstrated that you know how.
 
You really are hilarious. Do some basic research. You've demonstrated that you know how.

All the raw data is there, and a list of what they adjust for. The turnout adjustment that you say is being done is not.

You are simply wrong.

Feel free to show us where the adjustment was done in this poll. Here I will make it easy for you:

If the election for President were today, would you vote for (choices rotated) Mitt Romney, the Republican? Barack Obama, the Democrat? Gary Johnson, the Libertarian? Or Virgil Goode, the Independent American Party candidate?

All Gender Age <50 / 50+ Race
Male Female 18-34 35-49 50-64 65+ 18-49 50+ White Black Hispanic Asian/
Other
Mitt Romney (R) 46% 49% 42% 37% 44% 50% 53% 41% 51% 55% 8% 36% 36%
Barack Obama (D) 50% 46% 53% 53% 54% 46% 43% 54% 45% 42% 88% 58% 52%
Gary Johnson (L) 1% 1% 1% 1% 0% 2% 2% 1% 2% 1% 0% 2% 4%
Virgil Goode (IAP) 1% 0% 1% 1% 0% 1% 1% 1% 1% 1% 0% 1% 1%
None Of These 1% 1% 1% 2% 1% 0% 1% 1% 0% 0% 1% 0% 4%
Undecided 2% 2% 2% 6% 1% 1% 0% 3% 1% 2% 3% 3% 2%
Likely & Actual Voters 100% 50% 50% 26% 27% 28% 18% 53% 47% 61% 8% 21% 11%

And here is the list of adjustments done by the pollster:

Where necessary, respondents were weighted using the most recent U.S. Census estimates for age, gender, ethnic origin and region, to align the sample to the population. In theory, one can say with 95% certainty that the results would not vary by more than the stated margin of sampling error, in one direction or the other, had the entire universe of respondents with home telephones been interviewed with complete accuracy.

And then you can try and explain how people who have already voted are subject to your turnout factor.