Third debate: McCain vs Obama

[hiding behind a poster]

If 40% of Americans were prepared to vote Republican even with 25% unemployment and a 50% foreclosure rate, then what makes you think even more won't vote for McCain with just a 6% unemployment rate?

Oh Good Lord. I've no doubt that more than 40% will vote for McCain. But he'll still lose. Your argument is ridiculously simplistic - just like all the other "arguments" you make, like "FG and Labour doubled unemployment" (wrong, but it doesn't stop you), "State A voted for the winner every time since 1952, and polls show McCain leading there - therefore he'll win", etc. You're now reduced to this kind of nonsensical mathematical chicanery, whereas the REAL maths show Obama ahead by 7% nationwide, and winning all the key swing states. Those are the figures that count.

[FutureTaoiseach]

Oh Good Lord. I've no doubt that more than 40% will vote for McCain. But he'll still lose. Your argument is ridiculously simplistic - just like all the other "arguments" you make, like "FG and Labour doubled unemployment" (wrong, but it doesn't stop you), "State A voted for the winner every time since 1952, and polls show McCain leading there - therefore he'll win", etc. You're now reduced to this kind of nonsensical mathematical chicanery, whereas the REAL maths show Obama ahead by 7% nationwide, and winning all the key swing states. Those are the figures that count.

The Electoral College system means that some of the leads in swing-states are effectively statistical-ties, especially in Ohio, Nevada and North Carolina.

[cyberianpan]

The Electoral College system means that some of the leads in swing-states are effectively statistical-ties

What exactly do you mean by a "statistical tie" and what do you think it implies?

cYp

[FutureTaoiseach]

What exactly do you mean by a "statistical tie" and what do you think it implies?

cYp

On or within the margin-of-error. It implies - correctly - that they could go either way.

[Rocky]

Didn't happen in the first two debates.

Can't see any reason why it should happen in the third.

Neither can I and just for the record if I had a vote, I would be voting for McCain.

[cyberianpan]

On or within the margin-of-error. It implies - correctly - that they could go either way.

Do you think it has equal likelihood to go either way ?

cYp

[hiding behind a poster]

The Electoral College system means that some of the leads in swing-states are effectively statistical-ties, especially in Ohio, Nevada and North Carolina.

You see that argument would hold water if they were balanced ties, but they're not. For instance:

Ohio: Obama ahead in 8 of last 9 polls, by between 2% and 6%. McCain ahead in just one poll, by 2%. Average Obama lead - 3.2%.

Virginia: Obama ahead in 6 of last 7 polls, by between 3% and 12%. McCain ahead in just one poll, by 3%. Average Obama lead - 6.5%.

Florida: Obama ahead in 11 of last 12 polls, by between 2% and 8%. McCain ahead in just one poll, by 3%. Average Obama lead - 3.9%.

Missouri (FFS!!): Obama ahead in 4 of last 5 polls, by between 1% and 8%. McCain ahead in just one poll, by 3%. Average Obama lead - 2.2%.

Colorado: Obama ahead in 11 of last 13 polls, by between 1% and 10%, with 1 tied. McCain ahead in just one poll, by 2%. Average Obama lead - 4.6%.

New Mexico: Obama ahead in 8 of last 9 polls, by between 4% and 13%. McCain ahead in just one poll, by 2%. Average Obama lead - 6.7%.

Nevada: Obama ahead in 6 of last 7 polls, by between 1% and 8%. McCain ahead in just one poll, by 2%. Average Obama lead - 2.9%.

Iowa: Obama ahead in 9 of last 10 polls, by between 7% and 16%, with one tied. McCain ahead in none. Average Obama lead - 10.6%.

North Carolina: Obama ahead in 6 of the last 11 polls, by between 1% and 6%, with 2 tied. McCain ahead in 3 polls, by between 2% and 3%. Average Obama lead - 1%.


Now strictly speaking, by the definition of statistical ties, (two candidates separated by less than twice the margin of error), all bar Iowa from that list are technically, just about, "statistical ties". But they're not by any stretch of the imagination balanced ties, in which each candidate leads by a small amount in a few polls, with a few more polls tied. Obama leads in the vast vast majority of those polls, with only the occasional outlier that gives McCain a tiny lead dragging the overall figure back (just) into the definition of "statistical tie".

Of 85 state polls referenced in the analysis above, Obama leads in 69 of them. 4 are tied, and McCain leads in just 12 of 85 polls. Furthermore, Obama has double-digit leads in many of those polls, whereas McCain's biggest lead in the 12 he leads is just 3%.

And finally, you know and I know the significance of the states listed above - Ohio, Virginia, Florida, Missouri, Colorado, New Mexico, Iowa, Nevada, North Carolina. And its this they're all states Bush won in 2004, and at the very very best, McCain can only lose the two smallest states in that list if he's to win. If Obama wins ANY of the rest, McCain is toast. And those polls say that Obama will win them all. Sure, maybe one or two might flip for McCain by a fraction, but there's no way they all will.

They are NOT "statistical ties". They were once, but they're not anymore.

[cyberianpan]

Could I caution people against using the phrases "statistical tie" or "statistical dead heat", they are not particularly meaningful - please see this thread.

cYp

[Trampas]

And those polls say that Obama will win them all. Sure, maybe one or two might flip for McCain by a fraction, but there's no way they all will.

Unless of course a triumphant and triumphalist black America is more than they will be able to stand.

[jdckelly]

Hope McCain puts in a good performance tomorrow otherwise he'll have a difficult time of it. That said winning debates doesn't necessarily mean winning the election but it doesn't hurt.