AuH2O
YaBB God
Posts: 4,239
|
|
« on: November 02, 2006, 09:21:01 PM » |
|
I know, different environment, different candidates, etc. I just wanted to take a look at how close the final polls in that year were to the actual results, which could tell us something about what kind of poll deficit can be overcome. If anything polling is probably harder now than ever, well since the spread of the telephone. I'm posting this as I go so I don't know what the results will be.
I might look at 2004 also, but I thought the last off-year was a good place to start. I have to think that a lot of local 2004 races were heavily influenced by turnout in the Presidential race that produces a much different outcome than in an off-year. Bush was actually popular in 2002 which is a major difference too, but you would think polls in the final week would have a handle on that.
*indicates a poll taken more than a week or so prior to election day. Partisan polls excluded. Final results rounded to nearest integer.
Arkansas Opinion Research: D +8* Zogby: Dem +12 Gallup: Dem +8
Actual: D +8
Colorado Talmey-Drake: Dem +4 Zogby: Dem +5 Gallup: GOP +2 Ciruli: Dem +1
Actual: GOP +5
Georgia Mason-Dixon: Dem +6* Zogby: Dem +2 Atl Journ-Const: Dem +5
Actual: GOP +7
Iowa Research 2000: Dem +9 Selzer: Dem +9 Ced Rap Gaz: Dem +22
Actual: Dem +10
Louisana Excluded due to annoying runoff system.
Minnesota Zogby: Dem +5 StarTrib: Dem +5 Mason-Dixon: GOP +6
Actual: GOP +2
Missouri KC Star: GOP +5* Research 2000: GOP +4 Zogby: GOP +7 Gallup: GOP +4
Actual: GOP +1
New Hampshire ARG: GOP +4 UNH: GOP +1 Research 2000: Dem +1 Zogby: Dem +7
Actual: GOP +4
New Jersey NYT/CBS: Dem +13 Mason-Dixon: Dem +6 Rutgers: Dem +12 Zogby: Dem +11 Research 2000: Dem +9
Actual: Dem +10
North Carolina Mason-Dixon: GOP +6 Zogby: GOP +7
Actual: GOP +9
South Carolina Mason-Dixon: GOP +17 Zogby: GOP +11
Actual: GOP +10
South Dakota Mason-Dixon: Dem +2* Zogby: GOP +4 Gallup: GOP +3
Actual: Dem +fraction/1
Tennessee Mason-Dixon: GOP +10 Zogby: GOP +13* UT: GOP +12*
Actual: GOP +10
Texas Scripps: GOP +9* Zogby: GOP +4 UH/Rice: GOP +6
Actual: GOP +12
A few random thoughts:
- Mason-Dixon is usually right but not always!
- The only total miss by polls was Georgia. And even then, it was just outside the MoE and M-D didn't take a poll right before the election, so the late movement in that race wasn't really picked up due to a lack of polling. Colorado was pretty bad too
- Beware the MoE! Minnesota and Missouri were races with a lot of +5s one way or the other, but those wound up being pretty close. There's a big difference between 5 points, which could easily mean 1-2, and 10, which even with Zogby usually means a real lead.
|