The Keystone State of Play

Last week, we highlighted a flood of surveys showing Obama slashing Clinton's once formidable lead in Keystone State from 16 points to about six—a development that signaled trouble for the former First Lady.  "Pennsylvania is as much about who wins the vote as who wins the expectations game," we wrote. "The problem for Clinton is that, with April 22 long established in Hillaryland as a must-win (if not a will-win), expectations are astronomical for Clinton and abysmal for Obama... In other words, Clinton wins only by winning big; a close finish may look like a loss." Since then, an additional four surveys have hit the wires--and one in particular has Clintonistas chattering. In a poll of 597 likely voters, SurveyUSA found Clinton (56 percent) enjoying an 18-point lead over Obama (38 percent), her largest since the middle of March. Because the gap matches Clinton's margins from before Obama started stumping and advertising in the state, her supporters are citing it as a sign that the Illinois senator's massive money advantage--he's currently outspending his New York rival five-to-one on the ground--has done little to endear him to Pennsylvanians.

Are they right? I suspect not. The devil, as always, is in the details. At the same time SurveyUSA released its results, a new poll from Quinnipiac also came out--and it showed Clinton (50) ahead of Obama (44) by only six, down three since its survey of last week. Which to trust? I'm leaning toward Quinnipiac. For starters, Quinnipiac quizzed 1,340 likely voters, more than twice SurveyUSA's tally--a typical sign of greater accuracy. Secondly, the startling difference between the two polls can be mostly attributed to the results from Southeastern Pennsylvania, an area that includes Philadelphia. In the SurveyUSA crosstabs, Obama is losing the region 52-43, while Quinnipiac has him winning 54-40. Both totals seem exaggerated, but Quinnipiac's is clearly more plausible--according to experts, the combination of the city's large black population and the suburbs' upscale Dems should make the southeast Obama's strongest swath of the state. Finally, the two other polls released in April mirror Quinnipiac, not SurveyUSA. The latest Rasmussen survey puts Clinton up by five (the same as last week), while American Research Group has her down 12 since its March 27 survey--for a 45-45 tie.

Regardless of which outlet is right, it's better strategy for Clintonistas to treat SurveyUSA's pro-Hillary poll as an outlier rather than a lone beacon of truth. Blame the expectations game. SurveyUSA excluded, Clinton's average lead in the April polls is now a meager 5.7 percent; if the range is restricted to this week alone, it's down to 3.7. Any stats that suggest the senator still retains her double-digit advantage--even if, in reality, she does--only serve to keep expectations sky-high. And who knows? Maybe a few more results like Quinnipiac's will help level the playing field--and make a narrow Clinton win in a very friendly state look a little bit less like a loss.

UPDATE, 11:50 a.m.: More nettlesome stats for Clinton (and more proof that this week's SurveyUSA numbers are a little skewed): the latest PPP poll gives her a mere 3 percent lead over Obama. Factoring PPP into our no-Survey-USA, this-week-only polling average, Clinton's edge dips to 3.5 percent.