Also in this playlist...
This transcript is automatically generated
Now more election myths president Obama's director -- opinion research says polls don't matter.
Reporter about that because I've been fooled by polls in 1980 every pollster miss the Reagan landslide.
-- -- -- polls accurately predict who is likely to win.
Let's take that criticism to someone who runs a big polling company Scott Rasmussen.
So polls get trashed.
All the time.
Of course they get trashed all the time do you usually wrong what -- usually get trashed by the team that's losing they don't like the results of the Sallie can't trust the polls -- -- we got some secret information somewhere.
The truth is we're shocked wanna poll really gets -- wrong in 2008.
The entire industry -- Hillary Clinton's win in the New Hampshire primary and the reason people were shocked is becomes they've come to expect.
Those things don't happen.
And I must say analysts looking for -- mistakes they hit -- with hand -- -- and that's what really named the -- one that I was about -- -- was the other was 1948 but I wasn't around then when -- won't really be true -- so we've improved since that.
So how do you sample.
You know the the theory is you just call thousand phone numbers at random and it all works out at what you do now because first of all you have to make sure you place a lot more calls into a place like California than Wyoming because there's a whole lot more Californians.
We also know that women answer the phone more than men and older people more than younger people one -- more than non whites.
So you have to call and areas to entry trying -- pretty representative sample and Italy do all that you have to do what we call -- the sample to make sure you have the right mix of men and women young and old.
-- way to sample.
Well I think that's total statistics course is that a lot of people hated.
But basically the concept is if we know that we're supposed to 52% of women live in a particular state we want to make sure that our sample matches -- 52% when.
All right chris'.
I'm told you favor Republicans time magazine -- your conservative.
Leaning in the New York Times 538 blog says.
Rasmussen polls were biased.
There's always criticism of -- And it is all the pollsters -- senate is always to get it right I mean we have nothing to gain by gaming the system.
But if you have a reputation for getting it right you make more money people trust betrayed if you have a reputation for getting it right -- get more people -- actions to.
Going on his criticisms you speak with the first person who answers the phone rather than a random selection.
You know there's an old academic theory about how to do poll what -- back in the 1960s and seventies when Gallup -- in the just a few other pollsters are worn anywhere near as many pollsters -- there -- today.
They would say I'd like to speak to the oldest male or the youngest female that would rotate within a household that was a very good technique.
Everything has changed since 1970s and the -- you may have heard we have Twitter we have cable news we have.
-- few other things homeless and have to do with all it has everything to do with -- when people don't communicate the way they used to.
One of the big issues in polling today -- -- cell phones it's not just a question he says you don't call cellphone actually we don't we reach people who abandon land lines through an online approach I -- I have two sons.
There even -- you can talk on a cell -- they think it's for tweeting and -- We need to find new ways to reach people where they are communicating five years from now there will be no more phone poll.
How do you get the information -- -- you say the Internet but.
Where will -- -- very select group that might respond Sharon.
And I don't respond to of these people -- -- -- -- the answers we don't know the answer -- -- -- the way things were in the 1930s when you're trying to experiment with different -- our firm this year is going to be doing some polls -- -- these these are internal experiments.
Using entirely different techniques so we can measure them up against the election results to see how well we do.
He also says in the 2000.
Bush win you were off.
Way off pro Republican 5% higher than -- you know we did we missed in the bush year we showed the president winning by or that then governor bush winning by more than he did.
We also then went back in this is the way -- -- work she would compare what -- actually found.
-- of the actual results and you learn from it.
So -- 2004.
We were within a half a percentage point of both pushing Carey's total.
We did miss with the Obama we predicted he would win 52 to 46 -- and that winning 53 to 46 but that's pretty small.
And on his behalf I have to say I -- -- was looking for stuff to confront him with and the bulls were the only examples I could find.
-- -- look we make mistakes every pollster makes mistakes.
It is part of the process even polling fear Reese's.
One result -- every -- will be off by more than the margin of error so that is part of the process and that's why you wanna look at a variety of polls look for the common ground.
Or even better given the pollsters can make mistakes I would say the -- with all you guys and look at the people who bet.
And the wisdom of crowds and where people bet on these things like in trade dot com.
What I'm here that's even that's the most accurate thing better than bolster hey listen I trust the wisdom of crowds I would trust the American people more than their politicians.
Any day of the week.
In trade though is driven a lot by conventional wisdom and by polling data it's a fairly thinly traded markets.
Solid -- and much for the presidency that's not as much was still relatively fit and in trade -- the end is very very good.
I don't but I don't mean -- it but it is not a replacement for -- We'll find out -- next and we sure well let's move on -- one last minute I assume.
Most people assume that the vice president and the president.
Are going to be from the same party and I thought that may be and that.
Well if this year is election ends up in the electoral apology 269.
-- so long shot maybe 200 to one against.
But if that happens the house picks the president which -- likely be Mitt Romney and the senate picks the vice president which will be Joseph Biden.
And it's unlikely but possible I was surprised looking at the totals how close they've been.
-- -- for -- one by one electoral vote and 1876.
More recently bush gore.
Five electoral votes so this could happen this time it it could I think we're headed to a close election were headed to a close -- in the electoral college and you've got to remember we have a fragile system.
The government today really struggling to earn the legitimacy the American people.
And to the way the election process is playing out is not helping that at all.
Thank you Scott Rasmussen not coming up.
Filter by section