I’m not getting anything from the link you provided so I don’t know what you’re talking about in the granular differences. (Unless you mean the two states that do not have an all or nothing approach to electoral votes. The polls in those states gather information in such a way that calling that granularity is just as possible as calling the likely winner in one of the other states.)

Also, the point is that the popular vote is NOT the important part. We had an election not so long ago in which the winner did NOT get a majority of the popular vote (Gore won the popular vote by 0.5 % and lost the electoral vote by 1%). The way the system is set up, it is the electoral college that is the important part.

@Timm,

You’re right about how it is. In ANY state (except Nebraska and Maine) the loser gets nothing .. even if they come VERY close in the popular vote. California is the largest state, but it’s not the only state that can swing an election. Ohio and Florida have played pretty big roles in recent elections even with VERY small differences in the popular vote within those states. (Also, California has the LEAST number of electoral votes per person since that number is skewed to small states.)

There are arguments on both sides of the electoral college issue .. and we heard a lot of discussion in 2000 when the popular vote winner didn’t win the electoral vote, but (obviously) that discussion died down. (At the time the Republicans weren’t clamoring for change.)

Considering the population of California is over 10% of the population of the whole United States, I really consider this an injustice to the election process.

When in fact nearly half the popular votes in California are going to McCain, they are automatically neglected.

Do I have this wrong, or is this how it is? Is it apporpriate for our modern elections?

-=T=-

I worked that part out in my own question. I think the part that was confusing, as I said, was this image:

http://farm4.static.flickr.com/3191/2967554019c50f583cdfo.png

The GRANULAR differences within a particular state. I don’t see how that’s being called at this point.

But yeah, I get the electoral college fine. It’s just that it seems like a cart-horse thing given the fact that the popular vote is the important part. But I get it now…it’s ALL based on the polling for the POPULAR vote, THEN the electoral results.

Again, I worked that part out during my original post as I was asking the questions.

And thanks for the FAQ pointer. I somehow missed that.

The vast majority of states are winner-takes-all but this is decided by the state. Nebraska and Maine, I believe, are the only two not to award all electors to the candidate who wins a plurality of votes.

http://www.archives.gov/federal-register/electoral-college/faq.html#takeall

I’m stunned to see this question from you. I’ve argued for years that the polls that talk about the nationwide popular vote are meaningless and kept saying that what we needed to look at was the electoral college. With the internet becoming a primary news source, it becomes easier to devote the time and space to doing this right.

One issue is that the national polls are too small to give good predictions within individual states — they don’t poll enough people in each state. Any given day will have SOME new state results but not fresh results for ALL states. So, these sites (fivethirtyeight.com is being very careful about methodology) are gathering those state results, and making predictions for each state based on those local polls. Most of the sites draw a line somewhere (above 50% + the margin of error) and say that if a candidate is above that line then they are likely to win that state. They typically draw two lines so they can say “below the first line = tossup, between the lines = leaning, above both lines = likely victor”. Fivethirtyeight actually goes a bit further. He takes the polling information (size of the lead, margin of error) and uses statistical analysis to turn it into a likelihood of winning the popular vote in the state (and thus the likelihood of winning the electoral votes from that state). He then runs repeated simulations (I think he runs 10000 simulations each time he updates the polling data – which he does daily). Each simulation looks at each state and “tosses a coin” Heads OBAMA, Tails MCCAIN. Of course the probabilities are not 50-50. They are determined by the polling data. So, for example if polling data shows that there is a statistical tie in a state, then Obama will win (about) half of the simulations for that state and McCain will win the rest. If polling data shows that Obama has a 60% chance of winning a state, then he will win (about) 60% of the simulations for that state and McCain will win the rest. If data shows that McCain has a 100% chance of winning a state, then he’ll win all 10000 simulations for that state.

Anyway, he simulates elections in this way for each state and then tallies the electoral votes. He then repeats this 10000 times. In this way he is calculating the likelihood that Obama or McCain will win the election. The numbers in the graphic you included indicate that Obama won an AVERAGE of 354.4 electoral votes in the 10000 simulations he ran most recently. McCain won an AVERAGE of 183.6 electoral votes. Further, Obama won 96.3% of those simulations and McCain won 3.7%.

PS It is quite rare that an electoral college voter doesn’t do what they have pledged to do.

PS fivethirtyeight has a FAQ link (upper left of front page). He answers your questions.