Friday, October 18, 2013
It's A LONG Way From 2010: Number Of Self-Identified Democratic Voters Skyrocket In Kentucky
In what has become typical of the controversial Democratic polling firm, PPP, (see their surveys post Newtown shooting, post failed-immigration-reform, post government shutdown) they're out with new numbers illustrating how Democrats have been able to capitalize on the latest drama being played out in D.C.
According to pollster Tom Jensen, Kentucky likely voters are furious over the government shutdown, and "taking it out on Mitch McConnell," the Senate minority leader. Though looking at the numbers themselves, it's still not entirely apparent what evidence he is basing his conclusion on.
It can't be the topline result, which shows Democrat Alison Lundergan Grimes leading Mitch McConnell, 45-43%. That's because their previous poll, taken months before the shutdown, found nearly the same thing. To Jensen's credit, Kentucky voters disapprove strongly of the government shutdown (60-32%), and also claim they'd be less willing (48%), rather than more willing (34%), to support a candidate who supported the government shutdown. But outside of those findings, McConnell doesn't appear to have been harmed by recent events in Washington - not even when PPP asks respondents who they'd vote for, knowing Mitch McConnell supported the shutdown (the result is unchanged, with Grimes maintaining 47-45% lead).
Besides the misleading memo, the PPP survey looks sound in terms of demographic findings, at least compared to the last midterm election. Whites make up the overwhelming proportion of the electorate, with older voters being over-represented in comparison to the general population.
But one data point stands out for diverging remarkably from past Senate results: self-identified partisan identification.
At the time the poll was conducted, PPP found 53% of likely 2014 Senate voters calling themselves Democrats, with only 37% identifying as Republican (11% identifed as Independent). Put another way, Democrats hold a 16-pt party I.D. advantage over Republicans (matching the party's advantage in deep blue states like Vermont, California, New Jersey, and Illinois). It's that advantage that explains how McConnell trails Grimes overall, despite winning more of his own base, more crossover support from the opposing party, AND more Independents.
Such a high Democratic partisan identification advantage in Kentucky wouldn't be all that surprising in a survey of REGISTERED voters, as Lake Research Pollster Matt McDermott notes. Democrats have held a large, long-term advantage in partisan registration for years in Kentucky. But in actual elections, that registration advantage has evaporated, sometimes significantly. Consider the chart below:
Of the six most recently exit polled Kentucky federal elections, the Democrat's party ID advantage has never exceeded nine points. And that was in a wave election year, in which Democrats won the White House with 53%, defeated Republicans in the House of Representatives 53-42%, and the Senate 52-45%.
And in stark contrast to PPP's finding, the most recent exit poll to be conducted in Kentucky found 40% of voters identified themselves to exit pollsters as Republican, 38% said they were Democrats, and 22% said they were Independent.
In other words, PPP is finding an electorate that's 15-pts more Democratic than it was in 2010, and 3-pts less Republican. All total, partisan identification has shifted a net 18 points in 3 years.
In short, as far as recent Kentucky federal elections go, there is nothing even approaching the partisan advantage among likely voters for Democrats found by PPP. Even Wenzel Strategies, a GOP polling firm, found Democrats with a smaller 12 pt advantage over Republican party affiliation.
Just for fun, what would the PPP survey have looked like if Jensen had found 2014 voters partisan identifying in a manner similar to the 2010 Senate race between Rand Paul and Jack Conway, the 2008 Senate race between Dan Lunsford and Mitch McConnell, the 2004 Senate race between Jim Bunning and Daniel Mongiardo, or the 2014 Wenzel Strategies party ID finding?
As you can see, depending on your pick of partisan turnout scenario, the PPP poll could flip anywhere from 1 point to 9 net-points in Senator McConnell's favor.
A quick note for poll watchers who shriek at even the mention of a poll's partisan identification finding. Yes, it is fluid. Yes, it has been known to shift rapidly, and is more a reflection of voters' overall feelings on a contest at a particular moment in time, rather than a reflection of how they're actually registered to vote. And certainly, in the wake of the government shutdown, with GOP popularity scraping the bottom of the barrel, it's conceivable that considerably more Kentuckians consider themselves Democrats than they did on election night 2010, 2008, or 2004. But if the November 4, 2014 electorate is anything like the last several elections in Kentucky, Mitch McConnell wouldn't be trailing Alison Grimes in the PPP survey.