Slashdot Mirror
Can Data Mining Win a Presidential Campaign?
Nerval's Lobster writes "According to the Associated Press, Mitt Romney's campaign has contracted consumer-analytics firm Buxton Co. to drill deep into consumer data, with the aim of digging up 'wealthy and previously untapped' donors. (Romney digital director Zac Moffatt told political Website Politico as far back as June that the Romney campaign would 'outsource' its data analytics rather than develop the necessary infrastructure in-house.) In addition to hooking the digital side of their campaign to the Facebook data hose, Obama's election managers have hired a mix of digital directors, software engineers and statistics experts. 'Obama for America is looking for Quantitative Media Analysts, Analytics Engineers, Battleground States Elections Analysts and Modeling Analysts,' reads a want ad on the campaign's Website. The goal: to create data processing pipelines, integrate new data into models, build tools, and generate reports. In an election this close, with a rapidly shrinking number of undecided voters and contested states, a razor-thin advantage created by data analytics could mean the difference between success and failure."
I don't think the polls are accurate. Is it the mainstream media that is pushing this notion that it's a close race?
It might actually be nice to have an election won on the back of how bright your geeks are, rather than just how much money you have.
I would impose the condition that you are only allowed to use geeks that support your politics though. Rather than permitting outsourcing, I want to see this work being done by card-carrying members of your political party.
So Romney is outsourcing Republican party functions too?
Ibid.
Romney has already lost this election. When he chose Paul Ryan as his running mate he sent the middle-of-the-road independent voters running away as they want nothing to do with the extreme conservatism that he represents. He gained only the far-right voters of his own party, but they would have eventually voted for him anyways because they hate Obama. He could have chosen Bill Clinton as his running mate and the GOP far-right still would have voted for him just because they believe Obama to be the devil in the flesh.
The only way Romney can win this is if the GOP makes an even more epic voter suppression effort than they did in Ohio in 2004, coupled with crooked balloting (and counting) like they did in Florida in 2000, and who knows what else.
The real puzzler here is why the GOP even let Ryan accept the nomination to be VP on a can't-win ticket. That really doesn't look that great for his future and the GOP loves Ryan. Not many people who were VP nominee on a losing run have come back to make a significant career in national politics (and some polls are already suggesting his congressional seat is now in play, too).
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
Anyone with money and a low profile wants it to stay that way. The sign at the gate will now say, no trades men, no door to door and no begging politicians.
Mitt Romney's campaign has contracted consumer-analytics firm Buxton Co. to drill deep into consumer data, with the aim of digging up 'wealthy and previously untapped' donors.
Mis-read it as "wealth of previously untapped policies", im obviously not cynical enough.
Reform what you are currently calling "Land of the Free" and your entire culture around elections.
If you can consider to change the polictics with data mining,something is horribly, horribly broken on a very fundamental level.
The answer is no. Not when you have fellow party members legitimately raping your campaign.
If you look at the numbers, the general election is usually decided by a few percent.
Those few percent aren't really worth reaching. A lot of them decide at the booth, making saturation advertising a desperate attempt to shove your name into their heads so it bubbles up to the top in a moment of indecision.
But, if you look at the numbers another way, the real key to winning the election is getting voters who already like you to vote. The party that wins is the party who's voters show up.
Will data mining help get people out and vote? Doubtful. Buying all the prime time slots and using them for nagging would probably be more useful...though data mining could identify and drive small donors to donate. Again, though, undecided voters probably don't donate to campaigns a whole lot. Why donate to a campaign if you're undecided?
when data mining and shaping the public opinion are the means to winning.
Redistricting. Let's redraw the lines folks!
While we're at it, let's change the voting laws across the country.
Your finance system for elections is in desperate need of reform. I live in Canada where companies can not contribute st all. For people the max donation I think is $1500. So the party has to get wide appeal and convince enough people to give them money.
ron paul just says logical arguments and listens to what people have to say.
And how is that working out for him?
I'd argue that data mining for donations could well increase the number of people who vote. People who can be persuaded to give, even a token sum, are more likely to show up and vote because now they feel a closer stake in the election. In addition, to support a candidate financially and then not to show up to vote for that candidate would create an uncomfortable level of cognitive dissonance.
Even if we don't bring out the Betteridge's Law of Headlines ...
With current candidates I'd say - no.
Of course data mining can win an election. Why do you think Gerrymandering works? To Gerrymander a district they need to know the likely voting pattern.
When the Republicans cut voting hours in Ohio, they needed to know which districts were Democrat and thus which districts should be given shorter voting hours to swing the vote:
http://thinkprogress.org/justice/2012/08/10/670441/ohio-limits-early-voting-hours-in-democratic-counties-expands-in-republican-counties/?mobile=nc
Then there's the GOP's Pennsylvania law, requiring ids their data mining tells them have a strong Republican weighting:
http://www.democracynow.org/2012/8/17/judge_upholds_penn_voter_id_law
To choose the id form, they cross minded the different id databases with Republican voters to choose the maximum benefit for Romney.
I don't think it's enough to let Romney win, he's a 1% candidate presented as mainstream and I don't think he's fooled enough people, but certainly in closer races it absolutely IS enough to win an election.
So now even the race for the office is a Big Data problem. In the future, the winner always has the biggest clout on clouds of computing.
ron paul just says logical arguments and listens to what people have to say.
That is why he doesn't do well with the American electorate.
In a PEW study, some 93% of elections (the big ones, like president and house) are won by the candidate who spends the most money. Of the remainder, most fail because of political scandal.
Romney won't be the center of a scandal. Being Mormon, it's unlikely that he'll be accused of having sex with an aide, and it's unlikely that any financial shenanigans will come to light.
Romney has been "out funding" Obama by a wide margin, and very little of this money has been spent. There's also the likelihood of pro-Romney advertizing by various groups independent of his campaign - anyone, any PAC or business can take out an ad.
After the nomination, check the proportion of pro-Romney advertizing versus pro-Obama.
I think you'll find that a fire hose of predominately Romney advertizing will change a good many opinions.
Obama has already won, Romney just doesn't understand it yet.
The reason for it is that the dialog now is not about being fiscally conservative and pro-individual freedom vs big government, it is now about who is the bigger Democrat: Obama or Romney?
Obviously Obama IS the bigger Democrat than Romney, thus Romney has already lost.
The moment Romney said for the first time that Obama is 'hurting Medicare by cutting it', that was the moment Romney lost. Why? Who believes Romney that he is going to out-Obama Obama?
--
But it doesn't matter who wins, Romney or Obama, the end result will be exactly the same, there is no difference. Actually it doesn't matter who wins right now, even if it could be Ron Paul or Garry Johnson, the coming collapse of the US dollar and bonds is unavoidable. The difference would have been what would happen AFTER the collapse, and with Obama or Romney (doesn't matter) after the collapse either of them would attempt to grow the government even more by completely finishing off the private sector and free market economy, whatever is left of it in USA.
For reference you can see a couple of comments that give a little more perspective.
MY OTHER COMMENTS
I disagree, it doesn't need to be a sex scandal, anything bad can kill a candidates chances. Romney for example seems to stumble from one war-on-woman disaster to a pays-no-taxes to, well today, he's facing the 'bain-shipping-jobs-to-china' scandal:
http://www.democracynow.org/2012/8/28/workers_at_bain_owned_illinois_factory
People don't donate money unless they think someone has a chance to win, or if they desperately need him to win.
For the 1% that's backing Romney, they're throwing money at him in hope, simply because it's cheaper than paying the same tax rate as the rest of us.
If Obama wins the rich will pay taxes like the rest of us, if Romney wins they keep their tax breaks. So it's not a donation to a candidate, it's an investment in a tax reduction. Not the same as before.
Data mining is very useful in advertising, which is what gets people to the polls, or, more often, convinces them to stay home.
I don't see how adding more donors to your rolls at this point is going to help either candidate with the undecided in the practically 2 months that are left. Will making one more commercial spend actually change someone's mind? Don't they spend enough as it is? At some point there has to be diminishing returns, if not negative, on a PR election campaign. I know at least for me, if I'm constantly bombarded by political ads, it begins to turn me against the candidate (just like car commercials).
I started getting mail from Mitt Romney addressed to the former occupants at my address from TWENTY YEARS AGO. If data mining is the answer, they're going to lose. Big time.
I'd rather talk about Ron Paul than either of these clowns. I can't tell them apart anyway.
The skin tone should give 'em away. And I hear one of them is kind of a nice guy.
“He’s not deformed, he’s just drunk!”
ron paul just says logical arguments and listens to what people have to say.
And how is that working out for him?
We'll let you know once he gets started.
Dark Reflection
How is exploiting razor-thin majorities different from all those ebils of "mob rule" that supposedly make "democracy" worse than "republicanism"?
I say you need more parties and less incestuous "stances" between the two, and less of this both the same in all but a name two party system. Government-by-the-people should not be a stupidity contest.
The current concentration of wealth and political power in the hands of the Board and Executives of a corporation is a throwback to the European feudal system from which the U.S. originally revolted.
The recent Supreme Court decision in the Citizens United vs. FEC is The MOST undemocratic thing to happen in the U.S. since the apportionment clause in the Constitution failed to count women, servants indentured for greater than 2 years or Indians. Oh, then there's the failure to adhere to the Constitutional requirement of at least 1 House Representative per 30,000 of those represented, but we've only been in breach of that Constitutional requirement since the census of 1790 was released.
The State of the (Republican) Union may be strong, but it's respect for a (Democratic) Constitution is weak.
Data mining for dollars from "previously untapped" sources may be futile, but it's certainly not futile. Just review Toqueville's record on Frenchman buying hereditary title and the nontaxable status that went with it.
Romney's career and personal life as they relate to his politics are part of a pattern that revolve around wealth and privilege for the monied class. It's not pretty. But whole system is corrupted by the lopsidedness of wealth and political power. The Bush league tax cuts and repeal of the inheritance taxes are only symptomatic, and so is this this datamining drive.
This is America. No one gives a crap if Romney didn't pay income taxes for ten years. They will just invent alternate "truths" where he did pay and his opponents are lying, or where it was heroic of him not to pay, or where Obama is much worse. If Fox, Limbaugh, and Beck repeat it often enough, it becomes true!
I agree with you (and them) completely, and had come to that same conclusion independently.
I don't mean to say that the elections can be bought in this manner - I'm neutral on the issue (or at least, undecided). I think of it more as an example of crowd-sourcing a question with an easy way to determine the result.
Correlation and causation really don't matter in this instance. The observation is that the two correlate; consequently, we can predict the future from past observations.
The 1% figure is interesting. If (as I suspect) Romney has around 10x the campaign money as Obama, that accounts for a little over 3% of the popular vote (three doublings of spending). Would that be enough to give Romney the election?
(I don't care for either of them - this as just an intellectual exercise.)
I mean, Yes WE can. How do you think Obama got the last election, by data-mining facebook for key phrases. This article is 4 years too late.
Instead of spending tons of money on analysis, focus on the issues. Make all state and federal elections publicly financed and eliminate all insider trading by elected and appointed officials. Those 2 things alone will eliminate most of the evils in our current system. The rest will be easy.
I disagree. On election day there's only so many human resources you've got available to you. You can't call everyone. You can't send rides to everyone's house. Rather, you've got to use some system for determining who is likely to vote for you, but is also not likely to vote (and is also registered). In the past, that was just smart precinct captains sitting in a room and checking off people on a list because they knew them all personally. Now a days, you can use huge amounts of data and past voting history to give everyone a score based on the criteria I just outlined. If you can get clusters of high scoring people, you send a van to the cluster and call everyone to say "Hey, the voting van is here. Hop Aboard!". Without data, you're sort of shooting in the dark.
This. This is where data and turnout collide. Except it begins much earlier. The computers spit out lists of likely undecided voters and supporters likely to stay home and we call them and knock on their door. I've been doing it for over a month already, and will be at it twice a week until the election. It's the results of all that canvassing that gets put to use on election day. The data gets put to use much earlier. At least, that's how the Obama campaign is doing it. I guess the Republicans probably aren't spending a lot of labor on increasing turnout,.