This level of data analysis is already happening across a variety of businesses; as political campaigns gather more and more sophisticated and comprehensive voter data, and as technology allows for increased targeting of microniches, expect to see these techniques applied to political campaigns as well. They know where you live, what you drive, how much you make, where you went to school, causes you support, television shows you watch, products you buy and so forth - all that’s missing is a way to use that to finesse your vote or solicit funds from you. The Obama campaign is simply the furthest along this route at the moment, but there’s a long way yet to go.
It seemed that Penn was very slow to cotton to how the 2008 democratic primary worked, namely that it wasn’t winner-take-all, and delegates were being awarded proportionally by small geographic units. (precincts?) In many cases Clinton “won” a state, yet evenly split, or often came up short on delegates. Obama was able to build an early delegate lead while not doing all that much better than, or even falling behind Clinton in popular-vote terms. Recall the Clinton message that while she was behind on delegates, she had “won” more states than Obama?
Meanwhile team Obama was targeting resources toward precincts (?) with an odd number of delegates, where the winner of near-ties could pick up one extra, and letting Clinton win where the even number of delegates would be split equally unless the result were very lopsided…while making sure that it didn’t get that lopsided.
So beyond targeting voters, they were looking at where they would get the most bang for the buck with respect to the nuts and bolts of the election rules. It seems that Penn was doing fairly well at targeting and influencing voters, but not paying attention to where they lived and how many delegates they might deliver.
Nate got his data from the polls, true, and did his prediction on electoral votes, not the popular vote counts that showed it to be close.
Here is a link to the Technology Review story. And a relevant paragraph
They did this kind of thing to a level that no pollster could afford. They also could target voters, experiment on techniques to sway them, and thus affect the outcome.
I remember that article, and that is an excellent example of the analytical methods the Obama campaign used.
I don’t know if anyone does this, but a store with a preferred customer list, or a list of potential customers, could do the same DVR analysis the Obama campaign did and find out where to put ads which would have maximum impact for minimal cost.
The key point here is that presidents are not elected by popular vote, so it really doesn’t matter how close that is. Obama owned the electoral college throughout most of the race, so why do we persist in calling it close?
Yes and no. The margin in the popular vote doesn’t matter, but the result of the popular vote in each state (or district, in Maine and Nebraska) determines how that state’s electoral votes go. And results of the states are interrelated for demographic reasons. So it’s an oversimplification to say it doesn’t matter if the popular vote is close. It’s a little odd to me that this tangent pops up in almost every thread about the election; I think people want to rewrite the history so it looks like Obama never broke a sweat and everybody knew all along that he was going to win. Before the election the state polls were consistent in showing a narrow but clear lead for Obama and that that lead should translate into an electoral college win, and that’s how it worked out. Romney’s campaign was convinced all the polls were wrong (nope) and that things were bound to break their way (nope). The margin in the electoral college always distorts things. Obama won 51 percent of the popular vote, which was a bit more than most people expected, and 62 percent of the electoral vote, which was basically what the pollwatchers expected.
Says who? People pay lip service to that but I’ve read articles here and there that state its also equally important, if not moreso, to motivate your own base to come out.
One reason Republicans lost is that they didn’t expect the Democratic enthusiasm to match 2008. It was obvious, that was the first time a black guy was on the ticket, he had a chance of winning, and the country was absolutely not going to elect a Republican after 8 disastrous years of Bush. But the GOP were wrong this time, as many pollsters will tell you that Democratic turnout in some areas exceeded 2008, and with the increase in minority voters, the Democrats were motivated to beat Romney. Didn’t Obama do worse with independents than Romney? That’s all you need to know about the importance of independents
Because if it looks close, more people will pay attention to it, and thus more people will watch TV news, go to news sites, or buy the paper. Duh.
The politicians want to call it close so that the one ahead can keep the energy and the money up, and the one behind can keep the donations rolling in and stop people from defecting.
I got the impression it was a legit surprise to them and by them I mean Romney and his family as well as Ryan and his family. Supposedly that is why they went to PA a few days before the election, their weighted polls (which supposedly factored in what they perceived to be enthusiasm gaps and lower minority/young turnout than in 2008, probably other things too) showed them neck and neck in PA (Obama won by 5.4%). Plus Romney didn’t write a concession speech because he didn’t think he would lose. If the GOP thought they were neck and neck in PA (which they lost by 5.4%) they could’ve easily assumed that meant they would win OH, FL & VA which were won by Obama about 1-2%. Maybe their internal pollings showed Romney ahead by 2-3% in those states.
I am interesting in reading it too and seeing how it happened. These were supposed to be the organizational and business professional republicans, not the kool aid drinking echo chamber republicans. Yet they acted like the latter, I don’t know how that happened.
This is a classic example of someone clinging to conventional wisdom even when it clearly conflicts with observed reality.
In this case, the problem arises from the non-fixed definition of “independent”. The current crop of “independents” includes a large infusion of people who are more accurately described as “Republicans who would rather be seen at a Nazi pedophile kitten barbeque than declare that party affiliation”. Ergo, reaching out to “independents” is less useful* now than it was twenty years ago, since “independents” are less likely to be actual persuadable voters.
*Less useful for both sides, but more so for the Democrats (who simply aren’t going to get those votes) than for Republicans (who might pick up votes from people who otherwise would stay home or vote Libertarian or something).
They used to have that BBQ every June up in Illinois but they stopped after this one year when a couple of guys drove their two-tone '70s sedan right thru the middle of everything. Miraculously, no kittens (Nazi, pedophile or otherwise) were harmed.
I’m a fairly reliable Democratic voter of the same ilk. Although there are probably more of us “darn it, I’m not going to pledge loyalty to a party” individualists “independents” on the right. It sort of fits the rugged individualist paradigm.
However, data mining shouldn’t make it too hard to figure out how I vote, if you go beyond “white, middle aged, suburban, high income” and into religion, occupation, spending patterns, known associations. Even my car (Prius) is likely to tag me as a probable Obama voter.
Well, to be fair, I also subscribe to Forbes, Fortune and The Economist. We don’t all fit in tidy little boxes - which is why you get “there is an 87% she’s pregnant” and there is still a 13% that she has a vitamin deficiency, doesn’t like strong smells, and needs a blue rug.
In 2008 some one wrote to Mr. Roadshow, the Bay Area traffic guy, that he put a McCain sticker on his Prius so that he could identify it in the parking lot - since all the rest had Obama stickers. So that might be the explanation.
