|Here are just some of the highlights: (May 14-18, 608 registered Democrats likely to vote in the primary, MoE 4.0 for the entire group, larger for smaller subsets)
The polling memo also features complete crosstabs, and numerous different messaging tests. Check it out.
- Specter leads generic Democrat 50-37: In addition to the trial heat against Sestak, the poll also tested Arlen Specter against a generic Democrat, showing Specter with a 50%-37% advantage. However, a closer look shows that Specter's support is soft:
Specter vs. Generic Democrat (page 5)
Definite Specter: 22%
Probably Specter: 23%
Lean Specter: 5%
Lean Generic: 3%
Probably Generic: 14%
Definite Generic: 20%
Notably, "Definite Specter" and "Definite Generic" begin statistically even. There is a lot of room for persuasion on both sides, and no clear advantage in hard-core support.
- Sestak already leads among voters who know both candidates: Perhaps the most remarkable number of all in this poll is that Joe Sestak is already leading Arlen Specter among the 30% of voters who know both candidates (p. 4-5):
Among voters that know Sestak (mostly in the Philadelphia inner suburbs) he enjoys an 18 - 4 favorable-unfavorable ratio, and among voters who already identify both candidates, Sestak actually leads Specter in the initial head to head 52 - 44 percent.
With numbers like these, claims from Pennsylvania Governor Ed Rendell that Sestak has no chance are unmasked as either absurd or desperate. Sestak is already winning among Democrats who know both candidates. Rendell might have even seen this poll before he made those remarks, since it was completed ten days beforehand.
- Belief in why Specter switched is top determining factor in vote: Specter leads by 56% among voters who think he changed parties because he agrees with Democrats, but trails by 32% among voters who think he switched parties to save his job. As such, the primary campaign will likely turn on perceptions of why Specter switched (page 4):
Nearly half of voters agree with the statement that Specter switched parties and became a Democrat "mostly because he agrees with the Democrats more on issues," while the other half says he became a Democrat "mostly because he couldn't win election as a Republican."
This appraisal strongly impacts voting preferences. Among those inclined to believe Specter switched because agrees with Democrats more on the issues, he trumps a potential Democratic challenger by 56 points, 74 - 18 percent. However, among those who believe Specter's primary rational was political expediency, he trails a generic Democratic candidate by 32 points, 28 - 60 percent, suggesting that should more primary voters come to believe that Specter won't be a Democrat when it counts, the race could dramatically shift away from him.
Most voters already view Specter as doing whatever it takes to save his own job, but, as the memo notes, his numbers in this area are "par for the course with politicians." (page 3) Defeating Specter will require making him appear exceptional, even for a politician, in doing whatever it takes to save his job.
- Sestak's message tests higher than Specter's: The messages GQR tested for both candidates give the edge to Sestak (page 6-7):
In the exchange of positives, voters give Specter's independent message a mean rating of 6.9 on a 1-10 scale, while Sestak's message garners a 7.4 rating. Importantly, undecided voters rate Sestak's message higher than Specter's message by more than 3 to 1. In the follow-up vote after primary voters hear the messages about the candidates, undecideds break for Sestak by 3 to 1, which helps close the gap between the two. Sestak also makes up ground among what could be considered Obama's base in Pennsylvania-voters in the Philadelphia market, African Americans, and younger voters.(...)
Three specific charges resonate most powerfully. Specter's past votes in support of the Republican agenda of tax breaks for companies that move jobs overseas (39 percent very serious doubts), Medicare cuts and Social Security privatization (41 percent), and Bush's tax policies (35 percent) put him directly at odds with the Democratic electorate and raise serious doubts for large majorities of primary voters-especially among core blocs of the Democratic electorate such as women, seniors, union households, and American Americans.(...)
Conversely, despite the fact that this is a Democratic primary audience, the least troubling negatives for voters are the so-called inside baseball attacks, such as his quote that he switched parties because he saw a poll, or placing special emphasis on the fact that he took votes against Democratic initiatives after he switched parties. Similarly, his record on Supreme Court nominations in the past is not of particularly high importance. The strongest issues on which to frame the situation are the economic ones that directly affect people's lives.
This all leads to a core conclusion that Specter's performance in supporting the Democratic agenda over the next year might be the determining factor in the campaign. However, I wonder if even that will be enough to save him, as newfound support for those priorities can also be portrayed as merely an attempt to save himself, and does not change the underlying perception that you can't trust Specter to stand with Democrats over the long-term.
Finally, a quick note on why this poll is showing a closer campaign than the other polls on the race. First, discard all polls taken in the week immediately after Specter's switch. Not unlike a convention bounce, a campaign launch, or winning an early presidential primary state, the huge amount of favorable, free media Specter received that week was certain to temporarily skew public opinion in his favor.
Second, compared to the Quinnipiac poll in late May, which showed Specter ahead 50%-21%, this poll was of likely voters instead of registered voters. It also had fewer undecideds, breaking roughly 3-1 in favor of Sestak compared to Quinnipiac. Notably, that is exactly the break predicted by the polling memo quoted here.
Sign up to volunteer for Joe Sestak today!