Due to popular demand, I have created a swing state chart that is slightly modified from my normal methodology. Most importantly, I removed all of the Zogby Internet polls. Also, I removed three states, Indiana, New York and West Virginia, whose status as swing states was dubious at best, and based off of single polls that were either old or outliers.
With those tweaks, and looking only at polls conducted entirely after the Republican convention, here is the swing state chart:
Swing State Overview (270 to win, 269 to tie)
(Swing States are defined as states closer than 5.0%)
| State |
EV’s |
Obama % |
McCain % |
Margin |
Obama Total |
| Obama Base |
190 |
|
|
|
190 |
| Minnesota |
10 |
49.0% |
44.3% |
+4.7% |
200 |
| Washington |
11 |
47.0% |
43.0% |
+4.0% |
211 |
| Wisconsin |
10 |
46.0% |
43.0% |
+3.0% |
221 |
| Michigan |
17 |
47.2% |
45.2% |
+2.0% |
238 |
| Pennsylvania |
21 |
47.3% |
45.7% |
+1.6% |
259 |
| Colorado |
9 |
47.3% |
46.7% |
+0.6% |
268 |
| Virginia |
13 |
48.0% |
48.0% |
Even |
281 |
| Nevada |
5 |
45.5% |
47.5% |
-2.0% |
286 |
| New Mexico |
5 |
47.0% |
49.0% |
-2.0% |
291 |
| Ohio |
20 |
45.1% |
47.3% |
-2.3% |
311 |
| McCain Base |
227 |
|
|
|
227 |
This chart conceptualizes the campaign as essentially a seven state affair right now. With Obama starting at 221 electoral votes, and McCain starting at 227 electoral votes, the seven closest states are Colorado, Michigan, Nevada, New Mexico, Pennsylvania, Ohio and Virginia. Such a close forecast is given further credence by yesterday’s tracking poll average, which suggested a 46.0%-46.0% tie.
While these numbers add plausibility to the notion that the campaign is exactly tied right now, I have to admit that my gut does not matchup with that prognosis. Something just feels wrong with these numbers. Maybe it is that the two smaller, newer tracking polls are the better ones for Obama. Maybe it is that the swing state chart only shows a tie because Obama is tied in Virginia, and it seems strange that he would drop nationwide but not lose any ground–or even gain–in Virginia. Maybe it is that I dropped some polls from this average (Zogby tracking, plus the most recent polls from Indiana, New York and West Virginia), just because they “felt” wrong. Maybe I am just a numbers worshiper, but it “feels” wrong to remove polls that “feel” wrong–event he Zogby Internet polls that are, in many cases, obviously wrong (like when they showed Obama up eight in North Carolina).
Pollster.com currently shows McCain ahead by 1.7% nationally, and by 278-260 in the electoral college if every state is pushed to make a choice. Forecasting a slight McCain lead just feels more accurate to me right now than an exact tie. However, since analyzing polls shouldn’t mix with talking about your feelings, I’ll just admit that both scenarios–a slight McCain lead and an exact tie–are real possibilities right now.
Update: Obama leads by 4% in today’s Daily Kos tracking poll. That is a hopeful sign.
Update 2: McCain’s lead in Rasmussen is down to one point. Also, accoding to Haggai, in all likelihood Obama led in yesterday’s polling. I think tremayne is really onto something with his tracking of the Palin crash. |
