Public Policy Polling, a North Carolina-based Democratic firm, has earned a reputation as an accurate polling operation. Nonetheless, a generous sampling of Democrats and the nascent stage of the race casts doubt on the veracity of its latest poll showing Rep. Bruce Braley with 9 to 13 point leads over his GOP opponents.
PPP released a poll Wednesday showing Braley, the likely Democratic nominee for U.S. Senate, leading Republican candidates Matt Whitaker (43-34), Joni Ernst (44-32), Mark Jacobs (43-31), Sam Clovis (43-31) and David Young (45-32).
First, the good news for Republicans: Braley might have a +10 favorability rating (34-24), but 43 percent of registered voters don’t know enough about the four-term Congressman to form an opinion. That gives an opening to Republicans—especially independent groups—to define Braley while interest and energy is focused on the GOP primary.
With the obvious caveat that polls 16 months before a general election have limited utility, PPP’s methodology seems suspiciously skewed toward Democrats. Before a chorus of Democratic ranting about supposedly skewed polls that didn’t give President Mitt Romney his due, let’s examine the data (as the two issues are not the same).
PPP’s poll surveyed 668 registered voters, and, like many polling firms, PPP did not weight the results according to expected partisan turnout (academic debate still rages on this issue). Their sample included 38 percent Democrats, 37 percent Republicans and 25 percent independents.
“We are measuring party identification, not party registration,” PPP director Tom Jensen told TheIowaRepublican.com. “Our polls last year consistently showed a much more Democratic electorate than the registration numbers and were correct to do so. The week before the election we put out a D+5 poll that had Obama correctly winning the state by 5.”
Indeed, partisan registration figures can fluctuate and don’t necessarily provide the most accurate barometer. But statistics on who actually turns out—and their party as a percentage of the electorate—are illuminating. An examination of the last two mid-term elections and the last presidential election shows that the PPP model is not a likely scenario for 2014.
In the 2012 election, Democrats represented 33.4 percent of Iowa voters, Republicans had 34.8 percent and independents represented 31.6 percent of voters, according to data from the Iowa Secretary of State. In essence, PPP’s polling methodology assumes that turnout in 2014 will be more favorable to Democrats than in 2012, when President Obama was on the ballot. That’s ridiculous.
In 2010, the Democratic share of the electorate was 35.1 percent, surging Republicans grabbed a 40 percent share and independents dipped to 25 percent of the vote.
PPP’s latest poll is similar to the mid-term cycle of 2006, a high-water mark for Democrats when they retook the House of Representatives. In 2006, Democrats and Republicans both held a 37 percent share of the electorate, while independents represented 26 percent of voters.
PPP’s model may be slightly off, but their top-line take on the Senate race—that an unknown and underfunded field of Republicans trails Braley—is widely accepted by Republican strategists and political observers. What would be really interesting is if PPP provided the data in such a way that journalists could run different scenarios based on partisan turnout.
Perhaps it’s possible that Democrats will have a banner year in 2014 in Iowa that surpasses even their 2006 thumping of Republicans. A lot can happen in 16 months.
But with Terry Branstad heavily favored to win re-election, Sen. Tom Harkin absent from the ballot and President Obama’s -4 approval rating in PPP’s own poll, that’s highly unlikely. Does anyone seriously think PPP should be using a partisan sample for 2014 that is even more favorable to Democrats than the 2006 electorate?
At this stage in the race, PPP’s polls are useful for Democratic candidates seeking to convince donors of the strength of their campaigns. But if PPP really wants an accurate snapshot of the race in Iowa, they should consider using a serious, plausible partisan weight for their Iowa polls.
blog comments powered by Disqus