Clinton v. Obama in GE State Polls
by silver spring, Thu Nov 29, 2007 at 04:28:32 PM EST
A large number of general election polls for various states have come out recently and I wanted to analyze them in some way, whereby the standing of the leading Democrats can be compared to that of the Republicans.
I thought that doing this via map format may be the easiest and most understandable way.
In order not to overwhelm with a large number of maps, I decided to narrow down their number by averaging the Republican results from the individual polls. For example, looking at results for Virginia (where SurveyUSA poll is used) the data included:
Clinton 45 - Giuliani 45
Clinton 48 - Romney 41
Clinton 50 - Huckabee 40
Clinton 42 - McCain 51
The Clinton "number" is derived by averaging the four Clinton numbers above (46.25), while the Republican "number" is the average of the four Republican candidates (44.25).
We can then derive the number by which Clinton leads the "average" Republican (2.0), and then compare this to John Kerry's 2004 percentage in Virginia (Clinton doing 10.2 points better than Kerry). The same process is repeated for Obama.
"Averaging" the Republicans also makes sense, as the Republican field is now in such disarray that it's hard to tell who their nominee will be. It should be noted that, depending on the pollster, different Republicans were matched. For example, SUSA last used the four Republicans mentioned above, while Rasmussen has recently tended to match against Giuliani, Romney, McCain, and Thompson.
I tried to make this analysis as objective as possible. I included each and every state poll I found from the month of November. If a state was polled by more than one pollster, the polls are averaged. I also included data only from polls where at least two Republicans were polled, in order that the data not be skewed by, for example, matching a Democrat only against Giuliani and getting a more favorable Republican number, as Giuliani has often been their strongest candidate. In some cases, you will see that a state has data only for Clinton, but not Obama. Again, this is because some polls have included a lot more matchups for Clinton than for Obama (a case in point is the recent Quinnipiac poll from Connecticut, where Obama ties Giuliani, and no other Republicans are matched against Obama; meanwhile, Clinton is matched against four different Republicans). Therefore, a vast majority of the state polling data here includes matchups of Clinton or Obama versus an "average" of three or four different Republicans. The Clinton and Obama numbers are then compared to each other, as well as to John Kerry's numbers from 2004. The numbers were rounded to the nearest percentage point in map format.
I wanted to look only at recent polls, and decided to limit these only to polls from the month of November. However, a significant number of states did not have polls from this month, so I used numbers from previous months for those states. All in all, approximately 70% of the data here is from November; 15% is from October; and 15% from September and August (bottom of diary lists all polls, dates, and links). Again, only Clinton and Obama are compared. There has been a virtual dearth of polls which include Edwards, and one has to go back all the way to mid-September to get a good feel for Edwards's standing in state polls (yes, approx. 15% of the data goes back that far, as I've said before, but the older data is there mostly to complement the newer November and October data, ie., I tried to include swing states which were not polled recently. Nevertheless, to include Edwards here would be like comparing apples to oranges as the timeframes are so way off.)
[I have double checked my numbers, btw, but if anyone finds any discrepancies, please let me know and I will correct any errors and update the diary.]
State Poll Results:
The first thing that jumped out at me from these general election state poll matchups is how well Clinton and Obama are doing. Compared to John Kerry's 2004 numbers, the Democrats are wiping the floor with the Republicans.
Clinton appears to have large leads on the Republicans not only in the northeast, west coast and upper midwest, but also across parts of the upper south. She beats the "average" Republican by 10-11 points in Michigan, Wisconsin and Minnesota -- all three states where Kerry beat Bush by only 1-3 points. Hillary is also ahead by 10 points in Kentucky and 22 points in Arkansas (although that's one of the older polls included). She rather comfortably carries Ohio and Missouri (by around 5 points, which is comfortable when it comes to those two states), and puts Virginia and North Carolina into play for us. Compared to Kerry, Hillary does very well almost everywhere (including doing better than Kerry in Massachusetts by approximately 4 points). She does very well, in relative (and absolute) terms, in places ranging from California, to much of the Midwest, and especially well in parts of the south, running an incredible 30 points better than Kerry in three southern states. As you can see, she does only slightly better than Kerry only in parts of New England (where the Democratic performance levels may be maxed out) and in parts of the southwest (could it be the immigration issue ?). The only states where she polls worse than Kerry are: Illinois (the Rasmussen poll there probably indicates a bias for native son Obama and an accompanying bias against Hillary, his leading competitor), Oregon (which often seems contradictory), and my homestate, Maryland (for reasons beyond me ?, but perhaps the Democratic percentage here has become so high where it's becoming somehow maxed out also).
If the election were held today, and these numbers proved valid, Hillary would likely win the election with a "guesstimate" of 342-352 electoral votes (with, ironically, Florida going very narrowly to the Republican).
Like Hillary, Obama does very well in the northeast, west coast and upper midwest; he leads very comfortably in those regions (despite his tie result with Giuliani and only a 6-point lead against McCain in a SUSA New York poll released yesterday, his lead over the "average" Republican is still 14 points in New York due to Obama's relative strength against Romney and Huckabee in the same poll). He is also leading or competitive in a number of recently red states like Nevada, Missouri and Virginia. Obama also does relatively better than Kerry in a number of southern states. His margins across most of the south (except Virginia) nevertheless do not come anywhere close to Hillary's numbers, and he therefore does not appear competitive in much of that region. Ohio looks like one trouble spot for Obama. He is behind the "average" Republican there, and it's the only state, other than New York and Massachusetts, where Obama actually performs worse than John Kerry. Hillary's Ohio numbers are much better in comparison. Interestingly, Obama does better than Hillary in a number of states. In California, Nevada, Kansas, and Missouri, it's almost a statistical tie (Obama doing 0.25, 1.75, 0.25, and 0.875 points better than Hillary in those four states), while in Iowa and Oregon his lead compared to Hillary seems statistically significant. One kind of odd thing I noticed here is how both Hillary and Obama are not doing that well in Florida even though they are doing relatively good in other purple states and even in certain red states. I have a strong hunch that the DNC decision penalizing Florida delegates is at least partly responsible, and can only hope that we can recover in Florida in the near future.
All in all, if the election were held today, and these numbers proved valid, Obama would likely win the election with a "guesstimate" of 288-298 electoral votes.
A number of persons have cited the Iowa numbers as indicative of Hillary somehow doing (progressively) worse and conversely Obama doing (progressively) better in the high-profile, high-intensity environment of Iowa, where the candidates have campaigned over all these past months. But if you look at some of the earliest state polls -- from April, seven months ago before the campaign really got off the ground -- you will see that, when comparing the relative strength of Clinton vs. Obama, the situation was very similar back then, with Iowa, Oregon and Kansas being the only states where Obama was performing better than Hillary against the Republicans; Washington state was tied (compare MAP 7 to MAP 6).
Therefore, the hypothesis that Hillary is doing worse in places where voters have taken a good look at her is false. The Iowa numbers, instead, have something to do with the inherent nature of Iowa voters. Iowa has historically been a very contradictory state in elections - both primary elections [e.g. Iowa Democratic caucus winners: Muskie (1972), "uncommitted" (1976), Gephardt (1988; with Paul Simon in second place), Harkin (1992; with Tsongas in second place); Iowa Republican caucus winners: Bush (1980), Dole (1988; with Pat Robertson in second place)], and general elections [e.g. Nixon (1960), Ford (1976), Dukakis (1988) -- this bizarre general election voting pattern mimicked only by Oregon and Washington state, btw]. This inherently contradictory nature may also say something about the relative effect (or lack thereof) of the Iowa caucus on subsequent state primaries.
I have tried to be as objective as possible in doing this analysis, and I hope that is reflected in the work. I am happy that my candidate, Hillary Clinton, is performing so well in state match-ups (contrary to the BS put out by the MSM), but I am also happy that Obama for the most part stands his ground in these polls as well. I sincerely hope that these numbers remain this way throughout the following year and, like the rest of you, look forward to a Democratic inauguration in January 2009.
Alabama (Rasmussen - November)
Alabama (SUSA - September)
Arizona (Rasmussen - October)
Arkansas (Rasmussen - August)
California (SUSA - November)
Colorado (Rasmussen - August)
Connecticut (Quinnipiac - November)
Connecticut (Rasmussen - November)
Florida (Mason-Dixon - November)
Florida (Palm Beach Post - November)
Florida (Rasmussen - November)
Florida (SUSA - October)
Georgia (Rasmussen - November)
Illinois (Rasmussen - October)
Iowa (SUSA - November)
Kansas (SUSA - November)
Kentucky (SUSA - November)
Maryland (Rasmussen - October)
Massachusetts (SUSA - September)
Michigan (Rasmussen - October)
Minnesota (SUSA - November)
Missouri (Research 2000 - November)
Missouri (SUSA - November)
Nevada (Reno Gazette - November)
New Hampshire (SUSA - September)
New Jersey (Quinnipiac - September)
New Jersey (Rasmussen - October)
New Mexico (SUSA - November)
New York (SUSA - November)
North Carolina (Rasmussen - October)
Ohio (Quinnipiac - November)
Ohio (SUSA - November)
Oklahoma (SUSA - September)
Oregon (SUSA - November)
Pennsylvania (Quinnipiac - November)
Rhode Island (Brown Univ. - September)
Tennessee (Middle TN State Univ. - November)
Tennessee (Rasmussen - November)
Virginia (SUSA - November)
Washington (SUSA - November)
Wisconsin (SUSA - November)
Polls utilized for April analysis:
New Mexico (SUSA)
New York (NY1)
New York (Siena College)
New York (SUSA)
Mason-Dixon: http://www.mason-dixon.com/public/index. cfm
Research 2000 (via St. Louis Today): http://www.stltoday.com/stltoday/news/st ories.nsf/politics/story/F9C2BDEAD0FF525 88625739600212529?OpenDocument
Brown University: http://www.insidepolitics.org/REL907.pdf
Middle TN State Univ.: http://www.mtsusurveygroup.org/mtpoll/f2 007/MTSUPoll_Fall_2007_national%20report .htm
NY1: http://www.ny1.com/ny1/content/index.jsp ?stid=1&aid=68562
Palm Beach Post: http://www.palmbeachpost.com/state/conte nt/state/epaper/2007/11/11/m1a_PREZ_POLL _1111.html
Reno Gazette (via My Silver State): http://mysilverstate.com/showDiary.do?di aryId=372
Siena College: http://lw.siena.edu/sri/results/2007/07_ Apr_NYPoll.htm
Another good link to assorted state polls: http://www.usaelectionpolls.com/