debating the effect of polls on the outcome of elections

[onenote]

There seems to be considerable handwring about polls showing Democratic candidates trailing in particular races or overall (and euphoria about certain polls showing Democratic candidates leading or closing the gap).

But what really is the effect of polls on the outcome of an election? The assumption of some seems to be that a poll showing a candidate trailing will demoralize that candidate's supporters, discouraging them from voting and thus becoming a self-fulfilling prophecy. On the other hand, there is a school of thought that a poll showing a candidate losing can energize the candidate's supporters to work harder and may cause the supporters of the candidate in the lead to become lackadasical, figuring that the election is in the bag and they don't have to vote. (Truman's victory over Dewey is sometimes cited as an example of this phenomenon).

My guess is that in the long run, polls in and of themselves have little impact on the ourcome of an election. Good old fashioned GOTV efforts involving cars, shoe leather, phone calls etc. are far more important.

[ananda]

1. Disagree.

When voting machines are rigged, and tactics are used to keep the poor
and the minorities from voting, polls are useful in making people think
that the election outcome actually corresponds.

[elleng]

2. 'The Truman victory

was also an embarrassment for the emerging public opinion polling community. Truman's 4.4 percentage point election margin contrasted with the pre-election polls predicting a Dewey victory ranging between 5 to 15 percentage points. After the election, analysts attributed the polls' failure largely to completing their surveys too early, thus missing a late swing in voter sentiment in favor of the President. Ironically, the polls themselves may have helped Truman's late surge to overcome Dewey when press reports of their surveys showing Dewey ahead energized the Democrats to mount late efforts to increase turnout, and made the Republicans over-confident of any need to get their own voters to the polls.'

http://www.eagleton.rutgers.edu/programs/egov/ap_trumandewey.php

[Ozymanithrax]

3. What is the effect of several snapshots of a horse race on the outcome...

It may be predictive of the winner. But, the horse may get spooked by the flash and stumble.

Polls can be good, bad, or a wash. But what polls are used for is creating a narrative in the media. The prevailing media narrative, supported by the polls, is that the Democrats will not do well and that Republican voters who call themselves a "likely voter" are in a higher percentage of the polled electorate.

So, it isn't the poll but the narrative that affects the election. A narrative of excited Republican voters backed up by selected footage of dancing Republicans holding misspelled signs inspires other Republicans to dance, misspell signs, and vote. That same narrative depresses Democratic voters who become despondent because their vote is pointless and why should they vote.

[onenote]

4. if that is undeniablhy and clearly the case, to what can one attribute polls showing

Democrats closing (or increasing) their margins in some races?

[Ozymanithrax]

6. Yes, a snapshot of a moment in time.

Several snapshots (Polls) will indicate a trend.

How good the poll is, however, affects the outcome. Rassmussen has a bias towards Republicans.

[onenote]

7. you are right about rasmussen

It is skewed. Of course, candidates on both sides also use internal polls that sometimes are skewed to try impact the narrative. I"m not saying polling can't be a factor. I'm just saying its not the only factor or even, in many cases, the most significant factor impacting the outcome of an election.

[alcibiades_mystery]

5. There are only about 8,000 articles and academic studies on this question

dating back to the late 1960's...

[liberal N proud]

8. There is probably a segment of voters who vote base on what they have seen in the polls

but for the most, I think it is either more simple than that or the care and vote for who they support regardless of polls.

[JoePhilly]

9. Wrong.

People are easily swayed.

People LIKE to think that they are RIGHT.

There is plenty of research on this.

Let's imagine that you want to get a room full of fair minded people to agree with you. One of your BEST tactics is to position 5 people. One near each corner of the room, and one or two near the middle of the room. If these people are LOUD enough, the opinion in the room will MOVE.

This is why the right wing keeps claiming that the US is a "center-right" country. They are trying to get the middle to MOVE.

People like to have their views AFFIRMED. When they don't have a strong view, they PICK the view that appearrs to be the MAJORITY view.

Which is why the GOP has been pushing polls for months PRIOR to elections that any one is thinking about.

They are "positioning" their folks in the corners, and in the center, to make it seem as if "everyone" agrees with them.

My Ph.D is in experimental Psych ... and I recognize this manipulation technique.