wpostServer: http://css.washingtonpost.com/wpost2

Most Read: Politics

Read In

Now Viewing: People from around the country looking at Post Politics section

See what's being read across the country ›

Social Surface: Politics

Behind the Numbers
Posted at 02:05 PM ET, 01/03/2012

Don’t get burned on the exit (entrance) poll

The Tuesday evening kick-off for the Republican nomination contest brings not only the first real votes of 2012, but also the first analyses (and bogus analyses) of exit polls.

In Iowa, it’s actually an “entrance poll,” a survey of randomly selected voters as they enter randomly selected caucus precincts across the state. But the outcome is the same: The data quickly become the main material for subsequent scrutiny of who showed up to vote and why they voted the way they did.

Numbers from the poll – conducted by Edison Media Research for the National Election Pool – will be blithely tossed around, so some background is important to poll watchers.

As subscribers to the poll, we will, when appropriate, report on Behind the Numbers early numbers that provide clues about the Iowa electorate and information about the issues and candidate qualities that are resonating. However, we’ll stay far away from the initial “horse-race” estimates. This avoidance is not only for ethical and contractual reasons, but also because the poll numbers aren’t magically predictive.

One confounding issue with the entrance poll – as with any exit poll – is that the data bandied throughout caucus night will shift around, and what’s used for final analysis Wednesday may be quite different from Tuesday’s early numbers.

The numbers aren’t bad or wrong, just preliminary.

All entrance and exit polls include multiple waves of data, so if one candidate’s supporters show up earlier than others, there may be a skew to the first results. For example, on the Democratic side in 2008, then-senator Hillary Clinton had the edge in the first wave of interviews, with now-president Barack Obama’s tally increasing with each of the subsequent two waves.

Early results are also adjusted to expectations gleaned from pre-election polling and historical vote patterns. These may or may not be on-the-mark.

Adding to the uncertainty in Iowa – and other caucus states – is that voters are interviewed on their way into the caucus – before, not after, selecting a candidate. At the caucuses, candidate representatives have the option of speaking briefly before the start of voting. So persuasion is possible, and some voters may change their minds.

As the actual votes are tallied, the entrance poll will be adjusted to the actual results, firming them up. When it’s all said and done, the numbers will rightly be the lingua franca of post-election analysis, but it’s best to be cautious with the early numbers.

By  |  02:05 PM ET, 01/03/2012

 
Read what others are saying
     

    © 2011 The Washington Post Company