Daily Kos has given us a new R2K poll of the now-vaunted Georgia Senate race, and I wanted to write a post going through the data provided in the cross-tabs. The results from the last poll two weeks ago will appear in parenthesis next to the current number(s).
Head-to-head
Saxby Chambliss 47 (45)
Jim Martin 45 (44)
Allen Buckley 5 (4)
The data has moved a little as more undecided voters are making a decision on the race. Martin gained one point, as did the libertarian Buckley, while Chambliss gained two points. This leaves just three percent of voters undecided. Chambliss should be pleased that he has not fallen behind, but he is still under 50 percent.
Let's move to the personal splits for the two major-party candidates.
Saxby Chambliss favorables
57 favorable/40 unfavorable (53/42) -- Change: +11 to +17
Among men: 62/37 (58/39) +19 to +25
Among women: 52/43 (50/45) +5 to +9
Republicans: 76/22 (71/24) +47 t0 +54
Democrats: 39/58 (39/58) -19 to -19
Independents: 54/43 (51/45) +6 to +11
18-29 year olds: 50/47 (49/47) +2 to +3
30-44: 59/38 (55/42) +13 to +21
45-59: 59/37 (56/39) +17 to +22
60+: 60/37 (49/38) +11 to +23
What we see here is that Chambliss has gained support across the spectrum, but most notably among Republicans, where he went from a personal +47 to +54. With indies, he went from +6 to +11. This is critical for him to win. The most dramatic move seems to be among seniors 60 years and older. This is called solidifying one's base.
His overall number is at +17, up from +11 two weeks ago. Chambliss seems to have gotten past the economic crisis which was badly hampering his personal numbers.
Jim Martin favorables
56 favorable/37 unfavorable (51/34) -- Change: +17 to +19
Among men: 53/41 (48/38) +10 to +12
Among women: 59/33 (54/30) +14 to +26
Republicans: 36/58 (27/54) -27 to -22
Democrats: 77/16 (75/15) +60 to +61
Independents: 56/35 (53/29) +24 to +21
18-29 year olds: 59/34 (57/28) +29 to +25
30-44: 57/36 (51/33) +18 to +21
45-59: 56/39 (49/37) +12 to +17
60+: 52/39 (47/39) +8 to +13
Martin's overall split is +19, slightly up from +17 two weeks ago. He has improved two points with men, and two points with women, so it is evenly split. What should be concerning is that his level of improvement with many groups of voters was not as great as Chambliss' over the same period. Yes, Martin's personal ratings rose among just about everyone, but not drastically.
What is also potentially worrisome is that whereas Chambliss improved seven points among his base, Martin's numbers with his base remained basically unchanged. To be fair, Martin was already at +60 with Democrats two weeks ago, while Chambliss was at +47. However, Chambliss is moving fast to shore up Republicans. Perhaps more troubling is that Martin went from +24 to +21 with independents. This is likely the product of Chambliss' attack ads which are defining the still-mostly-unknown Martin for moderates and voters on the fence. At this time, however, Chambliss only went from +6 to +11 with indies, so he is behind there, but catching up. Interestingly, Martin did improve five points with Republicans, with seems counter-intuitive since Chambliss has been hammering Martin with some new tv spots. This may be the most disenchanted Republicans who cannot stomach voting for Chambliss. I would still guess that their ranks are not great.
The reason Martin is still right on Chambliss's heels despite him not improving as much as the incumbent can be traced to one group: women. Women went from a good +14 to an even better +26 personally approving of Martin. While his personal approval numbers with men remained around the same, he is doing a lot better with the fairer sex, so much so that it washed over his mediocre gains with age groups and independent voters.
Let's now look at the head-to-head splits.
Overall
Chambliss 47 (45)
Martin 45 (44)
Undecided 3 (7)
Men
Chambliss 51 (48)
Martin 41 (41)
Undecided 1 (6)
Women
Chambliss 43 (42)
Martin 49 (47)
Undecided 5 (8)
Looking first at the male/female split, we see the usual men-go-to-the-GOP, women end up with the Democrat dynamic. As the approval numbers foreshadowed, Chambliss has improved with men, making a seven-point lead an even 10-point one. Conversely, Martin has made less progress with women, going from +5 to +6 in two weeks.
This is interesting because right now, Martin's personal number with women is +26, while the incumbent's is +9. What I think this means is that a small, but nonetheless potentially decisive bloc of female voters like Martin more than Chambliss, but they are not fully ready to commit to him. This is a more likely explanation given that there are five percent undecided, versus just one percent undecided for Chambliss.
This is good news for Martin, as his base is with women, and there are simply more of them out there to be gotten, while Chambliss' male base is basically completely decided. Martin has more potential to pluck votes up. My sense from these numbers, and Chambliss' behavior already, is that the incumbent will keep hammering Martin to drive up his negatives with women in order to deprive him of enough of their support to win.
Democrats
Chambliss 12 (13)
Martin 80 (78)
Undecided 4 (6)
Republicans
Chambliss 81 (77)
Martin 11 (11)
Undecided 3 (8)
Independents
Chambliss 45 (42)
Martin 47 (46)
Undecided 2 (7)
From these numbers, we see confirmation of two key themes mentioned above. First, there is only marginal movement of support for Martin among his base, while Chambliss upped his GOP number a bit more, though not hugely here. More formally-undecided Republicans (five percent) than Democrats (two percent) have made a decision in the past two weeks.
Second, Chambliss has made key strides among independent voters. A five percent slice of indies made their decision in the last two weeks, and many of them went into Chambliss' camp. Martin actually improved one percent with this group, which may be insignificant, but at least he seemingly did not drop as the result of Chambliss' attack ads.
My personal sense about Chambliss' GOP numbers is this: I think that as the economic crisis has passed, and Chambliss' unpopular support of the bailout faded from memory a bit, many of his base supporters came back into the fold for him. It was not that many of these Republicans were in the Martin camp, but that they just had not committed, many of them upset both at Chambliss and the GOP in general over the economic crisis. Obviously, these issues still persist, but right now they are not as fresh in voter' memories to the point of denying Chambliss more GOP support.
If I was advising Martin, I guess I would note that chipping away GOP support from Chambliss over the bailout may be worthwhile, but nonetheless trickier. Chambliss is softer among independents, voters who are not as rigidly ideological as stalwart supporters of either party. Whether this would imply a difference in specific tactical approach, I do not know, as I have not deeply considered the issue. But it goes without saying, and it is no brilliant thing to note that Martin needs to bring home those undecided women and work hardest on independents.
Whites
Chambliss 66 (63)
Martin 26 (26)
Undecided 1 (5)
This should be heartening for Jim Martin. He did not lose any support among white voters. Chambliss has gained three percent here, but these are those 60+ voters and others that make up his base. Either Chambliss' attack ads have not impacted Martin, or he has simply hit his hard ceiling. I am inclined the believe the second explanation. There is a base of white Democrats in states like Georgia and Mississippi, but it is not high. In Mississippi it is much lower than in Georgia, but here we see it isn't that high in the Peach State either. I wonder if Martin can improve much here, especially with just one percent left undecided. I don't think so. That doesn't mean he shouldn't try to improve here, but he may have hit his ceiling. This is actually fine as Chambliss might have hit his limit too. To improve, he will have to go after Martin, which he likely will. Martin's objective should be to simply preserve his one-quarter and look elsewhere, say to blacks.
Blacks
Chambliss 8 (8)
Martin 86 (83)
Undecided 6 (9)
This is the whole ball game for Jim Martin, and so far, so good. It looks like all the formally-undecided black voters who made their decision in the last two weeks have decided to go with Martin. Chambliss has not lost support, but he has apparently gained zero backing. As black voters have learned more about Martin, they have come to support him, and this is salient since they make up the most reliable voting bloc for Democrats.
Importantly, there is six percent undecided. This is high, with just three percent overall undecided. Here's another brilliant insight: Jim Martin needs basically all of these voters. If he gets them, he likely wins, especially if black turnout is high. As a Democrat, he should be well-positioned to do this.
It is important to note that 27 percent of the poll's sample is black. This is a bit below the statewide black population which sits at 29 percent. There is an argument to be made that this poll slightly undersamples black voters. This is entirely fair, and I imagine that slightly more black voters in the poll sample would close the two-point gap either some or entirely.
I think that black turnout in states like Georgia and Mississippi will be high this year. How much, I have no idea. That is up to anyone's amateur guess. So far, early voting statistics in Georgia seem to bolster this view. As of today, there have been 585,398 Georgians who have voted early. Of that number, 212,853 are black, or 36.36 percent. This is a big jump of over seven percent of what the black population is relative to the overall state population.
Admittedly, the number stood at 40 percent early in the early voting process, but it has been slowly dropping over the last few days. It may go even lower as more eager voters taper off. Still, as it stands now, there is evidence that black turnout will rise to at least 29 percent of the eventual total vote, if not higher. I also concede that if black turnout is much higher in the early voting stage, it may be diluted come November 4. This is obvious, if the most gung-ho voters headed to the polls right away. That should be a fear of Democrats, but one we won't know the full extent of for a while.
The whole point is that the higher the black turnout, the better it is for Martin. As things are right now, he is winning the share of the decided black vote he needs to win. As long as the undecided bloc comes into his camp in the end, he may have enough to win, even if black voter turnout is average or at 29 percent.
Finally, among Hispanics, who make up around five percent of the state, Martin leads 68-17, down from 69-11. I will only say that because Hispanics made up such a small part of the polling sample, it is likely that Chambliss did not rise a big six percent among this group. I think that is just a reflection of movement in the tiny sample. Indeed, the first poll found 20 percent undecided, and now it is 15 percent. There is no way the undecided percentage is that high for any group at this late stage. Look for Martin to ultimately win this group substantially, which will be helpful in a tight race.
Also, you may notice I am not including the libertarian Allen Buckley in this analysis. Needless to say that his level of support overall, and across the different cross-sections remained stable and the same. He is polling at about five percent, which is likely votes that would go to Chambliss otherwise. Any movement at all in his direction would be a great movement for Jim Martin.
Conclusion
After spending the entire campaign way behind Saxby Chambliss, Jim Martin has clawed his way up, and now finds himself just barely behind the incumbent, really in a statistical dead head according to every recent poll. However, while he has gotten this far, he has not yet taken the final step and assumed a lead of any kind in the contest. Finally getting over the hump might be the hardest thing to do.
Nevertheless, I think that this result should be received as fabulous news for the Martin campaign for a simple, yet not-so-obvious reason. When the first poll -- SUSA, I believe -- came out in late September showing the race a dead heat, everyone was skeptical. Then, when the handful of other polls including R2K, Insider Advantage, et al. said the same thing, people finally believed the race was tight. But one big question was: for how long? I had a sinking suspicion that once the economic crisis passed from the upper fold of the frontpage somewhat, Chambliss would regain a healthy lead. In other words, that Martin's rise was fueled solely by the economy, and that over a short period of time, he would fall a bit back to earth.
This poll confirms that that hypothesis was probably wrong. Yes, the economy remains the top issue, albeit not at an end-of-the-world, crisis level. But Martin's numbers have remained stable. He is thus in contention right now with just 18 days remaining. This is a real race.
In terms of winning the whole thing, I think Martin is well-positioned. He is already basically tied, and that is assuming average to slightly below average black turnout. As turnout among African Americans rises, so too does Martin's number. And right now, if Martin remains strong enough with women, he has a solid base.
A big key will obviously be protecting himself with that group and independents, something that will no doubt be a bit tougher in the face of an onslaught of ads from Chambliss' campaign and the NRSC which will likely spend to save the incumbent. Last quarter, Martin actually out-raised Chambliss, driven in large part by a huge push among Democratic loyalists and the netroots upon the discovery that the race was tight. Still, Chambliss has a big cash-on-hand advantage. The DSCC recently dropped $500,000 in attack ads on Chambliss. Chuck Schumer will need to spend even more here to even the playing field. I bet that he does.
This is a fun race, and one Jim Martin may just be in the right position to steal. Such a win would be even more unlikely than a Ronnie Musgrove victory, which has at least always been seen as a distinct possibility. We will know in just 18 days.
Subscribe to:
Post Comments (Atom)
Posts
No comments:
Post a Comment