Tribetime
(1000+ posts)
Send PM |
Profile |
Ignore
|
Thu Nov-13-08 08:57 PM
Original message |
question on Ohio being called at around 15% in |
|
I assume the early exit polls must have showed Obama with a comfortable lead well over the 4% he won by. The original was too early to call when the polls close. Does anyone know the margin they must have to characterize it as such. It still seems fishy to me that the final numbers must have been much closer than the original exit polling......Don't get me wrong I'd take any margin even one vote and be happier than shit.
|
TheKentuckian
(1000+ posts)
Send PM |
Profile |
Ignore
|
Thu Nov-13-08 09:03 PM
Response to Original message |
1. The results probably included Republiscum strongholds and when compared to the polling data |
|
they knew McCain was toast.
|
dsc
(1000+ posts)
Send PM |
Profile |
Ignore
|
Thu Nov-13-08 09:04 PM
Response to Original message |
2. They use a variety of methods |
|
but one big thing is that there has to be a 95% CI showing Obama winning vs. McCain to call on only polls. That would have to be about a 6 or 7% lead to call on only polls. If returns tend to confirm the polls then the call can be made with less of a lead.
|
ShadowLiberal
(1000+ posts)
Send PM |
Profile |
Ignore
|
Thu Nov-13-08 09:07 PM
Response to Original message |
3. Not sure how they do it, but they try to be within 3 standard deviations of being right |
|
Somehow with the exit polling data they get they figure out if they pass the three standard deviation mark, and call it if they meet that requirement.
When the results start coming in though, I think they start to more look at where crucial swing counties went, and what areas have and haven't reported yet and how they're likely to vote based on the past.
|
alcibiades_mystery
(1000+ posts)
Send PM |
Profile |
Ignore
|
Thu Nov-13-08 09:15 PM
Response to Original message |
4. Experts can tell by looking at partial results from two or three counties |
|
4% is huge.
These population effects are strata. They don't shift substantially. They looked at turnout and results in a strong McCain county and results in a middle Obama county and probably could have called it earlier.
It's not that complicated.
|
EconomicLiberal
(554 posts)
Send PM |
Profile |
Ignore
|
Thu Nov-13-08 10:01 PM
Response to Original message |
5. Even with 15%, you could tell McCain was underperforming in the GOP strongholds |
|
that a GOP candidate needs to dominate to win the state. They looked at W's performance there as a baseline. McCain underperformed Bush by a wide margin.
|
Tribetime
(1000+ posts)
Send PM |
Profile |
Ignore
|
Thu Nov-13-08 10:16 PM
Response to Reply #5 |
6. I just had to watch election night over again tonight on tape |
|
and I noticed Obama had about a 200,000 vote lead with 1% in from early voting I'm guessing. A lead which remained constant the whole night, would this help call it earlier also? Thanks
|
EconomicLiberal
(554 posts)
Send PM |
Profile |
Ignore
|
Thu Nov-13-08 10:33 PM
Response to Reply #6 |
|
Early voting reporting is pretty erratic. You really don't know if the initial 1% was early voting, or if it was just results from election day voting from quick reporting precincts.
When they saw enough data from Cincinatti, Dayton, Cleveland and Western Ohio, they had enough data points to realize that McCain was underperforming Bush by a large margin - and thus they called the state.
|
Tribetime
(1000+ posts)
Send PM |
Profile |
Ignore
|
Thu Nov-13-08 10:18 PM
Response to Original message |
7. Much thanks to all who responded, been thinking about this all week |
DU
AdBot (1000+ posts) |
Wed Apr 24th 2024, 06:16 PM
Response to Original message |