President Barack Obama arrives in NYC for tonight's debate, which is kind of important.
President Barack Obama arrives in NYC for tonight's debate, which is kind of important.
Mitt Romney is extending his lead on President Barack Obama to 50-46, say both our own
PPP poll and
Gallup. PPP's had a overly heavy conservative sample—outlier or intensity gap? While Gallup has a tied registered voter screen (just like before the conventions) and a likely voter screen that was hilariously off the last two cycles. Unskew them! Then again, did you see Obama's improvement in Gallup's favorability numbers? Given those are on a three-day schedule as opposed to seven days for the tracker, it hints at future Obama gains. Maybe.
On the other hand, the Investor's Business Daily/TIPP poll says that Obama stretched out his lead by a point overnight, and now leads 47.3-45.7. Of course, TIPP was horrible in 2008, and in any case, you should never trust a poll that includes decimal points in the results. And did you see that Rasmussen continues to show Romney with a 49-47 lead? But we can safely ignore that because 1) it's Rasmussen, and 2) he has Obama leading 49-47 in his swing-state sample for the first time since Oct. 8.
But what about Reuters/Ipsos? It now has Obama up 46-43! Five days ago, it had Romney winning 47-45. It's an internet poll, so we don't trust it. But trends are trends. Good! Maybe. So what about this RAND thing? It's not a traditional poll. They took a bunch of people, and have followed them continuously, online, the entire cycle. They have Obama leading 49.66-44.51, a three-point increase since 10/10. Of course, it's a weird "poll," and if one decimal point is bad, two is ...
Then there's ABC/WaPo's poll yesterday. Its 49-46 Obama lead is a point better than what he notched before the debates. Woo hoo! As a super credible pollster, those numbers carry extra weight, right? Sure. Maybe. The poll's nine-point Democratic advantage will sure drive the professional unskewers mad.
So to recap:
Polls that show Obama gaining: ABC, IBD/TIPP, RAND, Rasmussen (swing states), Reuters/Ipsos,
Polls that show Romney gaining: Gallup, PPP, Rasmussen (all states)
Polls that can be methodologically dissembled if someone doesn't like the results: All of them.
National Composites
Huffington Post: Obama +0.5
TPM: Romney +1.1
RCP: Romney +0.4
So what does all this tell you? That regardless the toplines, no poll is methodologically pure and perfect. And they disagree with each other. A lot. And substantively at times.
I still resort to the old standby—average them all together, which right now, gives us the slightest Romney lead, unless you're HuffPo.
But I've always preferred the state-level polling, and I still look at Ohio and feel good. Remember, the same PPP methodology that found Romney +4 nationally also found Obama +5 in Ohio over the same days. The big lead Obama built over the summer helped buffer the hit he took after his first debate. And color me skeptical of the likely voter screens which historically have been less accurate than registered voter screens.
So yeah, I'm zen. I'm confident in where things stand. The bleeding has been mostly stemmed, and Obama continues to lead in enough states to get reelected. Of course, he needs to deliver tonight, but I'm not going to freak out about it. There's nothing we can do except assume he did his homework this time and wait for the underlying trends of the race to keep this one in our column.
But for those of you who screamed all summer that you'd be more motivated if you pretended Obama was 10 points down—do you feel more motivated? I see a lot of hysteria instead. Chill the fuck out.
Comments are closed on this story.