They should not be released until the polls are closed. For that matter the media shouldn't discuss the election or at least draw conclusions until the polls are closed.

and does not release it when they get it then they are not a news organization.

And that goes for all "embargoed" news.

The news operations conspiring with each other over how to best protect us from what they know but we do not is not, in my view, a good model of journalism.

there's no useful exit polling data.

Ideally you want a random sample. In practice, that's too much work and too expensive. So you do the best you can and then reweight your data to make it match that ideal random sample.

That means a poll requires a model of the cohort being polled. For exit polls, you pick a few precincts that are fairly representative of what you think the electorate is going to be and you adjust it to make the sample match the electorate. Perhaps there's a preliminary adjustment to make the sample match the supposed electorate, but at the end you take the numbers you have and reweight them to match what actually turned out.

So if you think that the turnout in WI is going to be 60% D, 40% R; 13% black, 12% Latino, 4% Asian, 1% Native American, and the rest white; if you think the turnout is going to be 65% urban and 35% suburban/rural; 20% > 55 years old, 5% < 25, the rest in between. ... Then you pick your precincts based on the past voting history.

So let's assume that the black turnout plummets. That 13% "black vote" becomes 2%. You've just given the black respondents far too much importance. Let's say that the youth vote is 33% of the total, yet you say they were supposed to be 5%--and you've underweighted their vote.

We understand this quite well when we look at a poll and see that 65% of the respondents were (R) and the numbers weren't reweighted. We know that it's biased becaue we assume that in an actual election it won't be the case that 65% of the voters are (R). We don't like the numbers, we can accept that they're flawed. But we go stupid as soon as we like the unadjusted or misadjusted numbers.

With the very high turnout, all the voting histories for precincts are meaningless and the model the poll was built on is trashed. The electorate won't look like it was supposed to. The raw or early-adjusted poll numbers, if there are any, would be trash.

Exit polls can be wrong, if say at 5pm they say Walker wins. Someone on their way home from work may feel oh well nice try I will just go home we lost.

Exit polls serve their purpose but that purpose should not be to predict the election, but to how groups voted.

Mistakes in polls will always have whatever influence they have.

It is not the media's job to not influence votes. If that were the case they would all simply close up shop because all news will always run the risk of influencing votes.

The polls that most influence voters are pre-election polls. Should they all also be embargoed until after the election?

Releasing an exit poll is the same as not releasing an exit poll in that either action may or may not influence voters.

Voters are influenced by campaign ads, which are often full of "mistakes" in the form of flat lies.

If people are led to believe that their candidate has won they may not bother showing up to vote. Flawed and/or falsified exit polling can have that effect.


Not releasing a poll does not dissuade voter from showing up at the polls.

I honestly don't think any election night results should be released until all polls are closed. That means that we don't get to hear about what happened on the east coast until Guam is done voting.

People can wait to hear the results. It won't change anything. Unlike releasing information which just might change things.

The integrity of the voting process is far more important than a few hours wait.

CNN at least is only discussing aspects which doesn't tell how people say they are voting, but they are giving a support for restricting collective bargaining for public jobs poll. If one assumes those who strongly favors or favor restricting collective bargaining will vote Walker, then they in essence are telling how they predict the election will go.

At the same time, they are predicting a late results and have even discussed recounts.

So, it is going to be close.

Exit Polling data is NOT news and should not be reported in real-time.

However, the data should be utilized AFTER the election to check for disparities in pre and post election results and analysis.