| Quinnipiac (9/16-20, likely voters, 8/23-29 in parens):
Andrew Cuomo (D): 49 (60)
Carl Paladino (R): 43 (23)
Other: 1 (1)
Undecided: 7 (14)
The entire political world's been abuzz about this poll, showing an astounding tightening of this race in under a month, going from Cuomo +37 to Cuomo +6. In that timeframe, Cuomo's unfavorables spiked, going from 54-21 to 51-34. Still good numbers, but Paladino's name rec has soared, taking him from 16-13 to a not-exactly-great 36-31. The biggest change of all, though, is structural: Quinnipiac finally switched to a likely voter model, whereas all of its previous polling on this race relied on responses from registered voters.
A potentially major issue with this poll is Quinnipiac's failure to include Rick Lazio, who when we last checked still had the Conservative Party line. While Lazio hasn't made up his mind about how vigorously he plans to contest the election, his name will remain on the ballot unless some shenanigoats are pulled on his behalf. (After the Jon Powers debacle last cycle, in which he unsuccessfully tried to have himself taken off the Working Families line by moving out of the state, pretty much the only reliable way off the ballot now is to get nominated for a judgeship - or croak.) Quinnipiac's reasons for not including Lazio don't seem compelling, and as Nate Silver suggests, they could have tested with and without him.
One other detail: Q's poll was in the field for an unusual five days... and I notice their last survey was in the field for seven days. That's definitely on the long side. But in any event, the real question is, is the race really this close? At least one other pollster seems to think so.
SurveyUSA (9/20-21, likely voters, no trendlines):
Andrew Cuomo (D): 49
Carl Paladino (R): 40
Of course, it's SUSA, whose internals have left a lot to be desired this cycle. Once again, 18-34 is the strongest GOP segment, preferring Paladino by a kind of amazing 56-33 margin. But they may yet be right about some things, such as the fact that it's apparently a close race among independents (41-39 Cuomo) - which Quinnipiac sees as well (Paladino is 49-43 with indies). Thankfully Democrats still dominate the New York electorate, but this is still not where you want to be. And if SUSA's 44D-35R-20I sample is correct, then we've had an amazing comedown from 2008's 50D-26R-25I distribution.
But another pollster is seeing things very differently - though with two major caveats.
Siena (PDF) (9/16-17 & 9/19-21, registered voters, 8/9-12 in parens):
Andrew Cuomo (D): 57 (56)
Carl Paladino (R): 24 (14)
Rick Lazio (C): 8 (16)
Undecided: 10 (14)
The analysis here is basically the inverse of what we said regarding the Q poll. As you can see, of course, Siena tested Lazio, and they don't appear to have done a straight Cuomo-Paladino head-to-head. If the Lazio numbers are at all accurate (and he does stay in the race), then this further butresses the argument that Quinnipiac erred by not including him, since the two right-wingers are clearly sharing the same part of the pie.
But the bigger issue here is that Siena is still clinging to a registered voter model, rather than using a likely voter screen, despite election day being just six weeks away. However, Siena may have offered us a further means for comparison. Here, Paladino narrowed the race by ten points, while Quinnipiac sees him tightening things by thirty. If Siena is right, then the change in Quinnipiac's model would account for something like a twenty-point shift. That would be more or less double the largest shifts we've seen when other pollsters move from RVs to LVs. (Of course, we're comparing apples and oranges here for a variety of reasons; this would be a lot simpler if we had just one outfit that tested both types of voters.)
So what's the answer? Has Paladino turned this into a real race? Or is Cuomo still cruising? I think it's impossible to say without more polling, and I'm sure that the stir these numbers have generated will inspire more companies to go into the field. (And for what it's worth, Harry Enten claims on Twitter that Marist will have a new poll out on Friday showing better results for Cuomo.)
If you're in the mood to be depressed, my good buddy Jake (aka the artist formerly known as Trapper John) thinks Cuomo's in real trouble, and he may well be right. Voters are ticked off and, especially in New York State, truly disgusted at politicians. Cuomo's spent his whole life acting like an entitled twit, and he's seemed to view this election as a coronation. Meanwhile, Paladino, revolting meatbucket that he is, is a true outsider with a ton of money and a lot of anger on his side. If this race comes down to barfbag versus d-bag, voters may just wind up holding their noses. And as a New Yorker, I can tell you that the thought of Gov. Paladino chills me to my soul. Let's just pray Siena is right.