POLITICS

How well did the pollsters do predicting party turnout in 2010?

|
Politics,Beltway Confidential,Conn Carroll

The Weekly Standard’s Jay Cost does a fine job addressing conservative claims that recent polls showing President Obama with strong leads over Mitt Romney here. Cost’s bottom line:

If we instead assume that the GOP will have a net of +84 among its own partisans, and Obama can recreate the +79 he managed in 2008, then we are talking about a 2-point shift in these polls in favor of Romney. In other words, the above national polls give Obama a 3.6 percent edge over Romney; if the two bases fall back into historical alignment, then that lead would be cut to about 1.5 percent.

But Cost’s post only looks at this year’s polls and then compares them to recent exit poll data. But how well did polls from this stage in the campaign do at predicting eventual turnout in the past?

Let’s take the most recent NBC News/Wall Street Journal poll which currently shows Obama beating Romney 50 percent to 45 percent. The poll has a +5 Dem advantage, 42 percent to 37 percent.

Now let’s look at the most comparable 2010 NBC News/Wall Street Journal poll, conducted in late August of that year. That poll had a very comparable +4 Dem sample, 39 percent to 35 percent, and had Dems beating Republicans by two in a generic congressional ballot, 34 percent to 32 percent.

So how did the NBC/WSJ prediction of a +4 Dem advantage at the polls hold up on election in 2010? Not well.

According to the official exit polls party turnout was even in 2010, 35 percent to 35 percent. Additionally, Independents broke strong for Republicans giving them an overall six point win on House ballots nationally, 51 percent to 45 percent.

View article comments Leave a comment