BASED ON EXIT POLLS, THE PROBABILITY KERRY WOULD LOSE OH AND FL = 0.15%

[TruthIsAll]

That's less than 1/6 of one percent!

To put it another way, the chances are 1 out of 667 that Bush would win BOTH states.

Assuming a 2% MoE for the exit polls (they are much more accurate than standard polls), the probabilities are:

Ohio Exit Poll:
Kerry 52 - Bush 48
Prob (Kerry wins OH)= 97.7%

Florida Exit Poll:
Kerry 51 - Bush 48
Prob (Kerry wins FL)= 93.5%

Then
Prob (Kerry loses FL and OH) =0.15% = (1-.977)*(1-.935)

Prob(Kerry wins OH or FL or Both)= 99.85% = 1 -.15%

How did I calculate the probs?
Simple. Feed the data into the Excel Normal Distribution function:

Since the MoE = .02, the Standard Deviation is .02/1.96 = .01

Prob (Kerry wins OH)= NORMDIST(0.52,0.5,0.01,TRUE) = 97.7%

Prob (Kerry wins FL)= NORMDIST(51/99,0.5,0.01,TRUE) = 93.5%

[JohnnyCougar]

1. I wonder what would happen if

we did t-test for independent groups based on the differences between your two groups of states. Or an ANOVA. I wish I could get SPSS on my computer.

[TruthIsAll]

7. Forget SPSS. You have all the functions you need in Excel

tia

[shelley806]

2. It's so good to see probabilities worked out in black and white....

I know it makes sense, but I can't stand the media downplaying the significance of exit polls. Do they really take us for such fools?

[LisaL]

3. Yes. And most people seem to accept it without questions.

[George_S]

4. Maybe you have it there already....

but what are the odds that Bush would win both? I mean, Bush has a .15 chance, I guess, but that would be 1 out of what chance?

[TruthIsAll]

5. Its right there: 1 out of 667

tia

[TruthIsAll]

6. am kick

tia

[TruthIsAll]

8. PM Kick

tia

[Cha]

9. But rove told us to

disregard the exit polls..

That should be our fucking clue right there.

rove or cheney..whatever.

[peacetalksforall]

11. He said to disregard? But he also said voters lied to exit poll takers.

So according to the lying excuse - about 50% of Repubs lied by saying they had voted for Kerry when they actually voted for Bush. Wouldn't that mean that they would have had to orchestrate a request to Repubs to ask them to do that by phone or the pulpit or e-mail or billboards?

If they didn't orchestrated that, would it mean that there is a tradition of lying to pollsters?

If you were inclined to respond to a poll, would you lie?

Or did I misunderstand the figures?

[Thomas Jefferson]

10. kick