Friday, November 7, 2014

Post Election Hangover

Liberals and Democrats are crying in their milk, Conservative and Republicans are celebrating what seems like a clean sweep of the 2014 mid-terms, and we at the Power-Elite blog are busy celebrating this article in the Berkeley Journal of Sociology on why Big Data and the "Polling Industrial Complex," where predicting statistically who is going to win and/or prevail, has become just as meaningless and vapid in politics as the standard attack ad.

While the 2012 re-election of America’s first African-American President signaled resurgence in support for social liberalism, it also announced a sea change in our media culture: the triumph of polling statisticians over the “go with your gut” prognostications of traditional punditry. On Election Day, Nate Silver finally burst the bubble of such notable blowhards as Joe Scarborough, Peggy Noonan, and most dramatically Karl Rove, whose on-air denial of the election results and subsequent dismissal by Fox News’s election team quickly became YouTube spectacle. Meanwhile, Silver published quantitative analyses of how minorities’ limited access to voting booths would affect Obama’s chances of victory; the Obama campaign itself received plaudits for its use of statistical algorithms to identify likely voters and encourage turnout. It seemed the era of Big Data had finally come not just to political coverage, but to political activism.

It wasn’t supposed to be this way. If 2012 was the tipping point for media-savvy statisticians, 2014 is the first cycle where their reign is undisputed. Silver launched the new under ESPN’s auspices with a self-described manifesto ending with the line “It’s time for us to start making the news a little nerdier.” Other bloggers such as Ezra Klein proudly crow their “wonky” credentials. Wasn’t our media shocked back into shape by Silver’s data-centric journalism?

Apparently not. Data journalism has failed to mitigate the feedback loop governing Americans’ distaste for mass media, and has become a manifestation of the very social phenomenon it was meant to dissect: the bifurcation in media culture between fear mongering and colorless prognostication. The real problem with our media wasn’t that it was bad at predicting elections (although it was)—it’s that it spends so much time on predicting elections at all, as opposed to moderating and shaping a national debate on what is at stake at the ballot box. Statisticians like Silver have helped eliminate bias when it comes to election prognostication, but there hasn’t been a similar commitment to eliminating the bias of spurious political narratives peddled by major media outlets. This leaves data journalism in the unfortunate position of helping to predict our electoral choices without evaluating their significance and pointing to alternatives.
Exactly. Over the past two years I've been surprised by the lack of analysis over the quality or point of such predictions (and the general hagiography around Nate Silver). Coming from academia, predictive analysis is great and a gold standard to pursue, but without context, your research data is meaningless.

It's ironic that Silver's is under ESPN's banner, because when you watch predictive analysis on SportsCenter or MLB's Quick Pitch, not only do you get quantitative predictions, but you get in-depth evaluation of the significance behind a team or pitcher, and alternatives to quarterbacks, draft picks, etc.

Put it this way: the press coverage and analytics behind the NFL and MLB drafts is far superior to any political reporting that was done in the run up to the mid-term election this year. 
This isn’t to say there isn’t value in the technocratic skill and rigor behind data journalism. There is no question that a refined quantitative methodology for predicting election results is leagues beyond the horserace neuroticism of sites like Politico. But if, as Silver has said, he will not “do advocacy” and “won’t do a ton of public policy coverage,” then sites like FiveThirtyEight are really just a more skillful extension of the media circus Silver made a career out of criticizing. This is because eliminating bias when predicting events, in the absence of preventing bias when interpreting them, leaves intact the dysfunctional trajectory mass media has taken: its propensity for navel-gazing and sensationalism over actual journalism. As a result, data journalism runs the risk of statistically aggregating the U.S. political electorate before it can even express itself—and thereby downplaying its potential for transforming the political realities we face.
It's like being told your team preference before having a chance to figure it out yourself. Not only is that "advocacy" of the worst kind, it circumvents any kind of thoughtful input from the voters themselves. In that sense, Fantasy Football or Fantasy Baseball is far preferable to the media coverage of partisan politics in the U.S. right now.

But I think proof of how meaningless this Big Data dump on people is, without context, is seen in the results on Tuesday. Once you get past the simplistic binary battles between candidates and parties, you find that Congress still, as a whole, has an approval rating of 11% (only pedophiles poll slightly lower), and despite that, the incumbency return rate on Tuesday was 95%. Meaning, 95% of incumbents were voted back into office.

So bravo on being the "most accurate" in calling the election. The predictions and results are simply meaningless.

Getting past partisanship, the real surprise for this blog was in the social issues, which struck a surprisingly liberal/libertarian theme that contradicts the conservative electoral sweep. In every state where the issues were on the ballot, marijuana was legalized, the minimum wage raised, and sentencing reform enacted.
On at least six high-profile and often contentious issues — minimum wage, marijuana legalization, criminal justice reform, abortion rights, gun control and environmental protection — voters approved ballot measures, in some cases overwhelmingly, that were directly at odds with the positions of many of the Republican winners.

MINIMUM WAGE Initiatives to raise the minimum wage appeared on the ballots in four deep-red states — Alaska, Arkansas, Nebraska and South Dakota — and passed in all of them. The new hourly minimums range from $8.50 in Arkansas by 2017 to $9.75 in Alaska by 2016. Minimum-wage increases were also approved in San Francisco (to $15 an hour by 2018) and Oakland (to $12.25 an hour by 2015). In all, an estimated 609,000 low-wage workers will see raises from these approved increases.

MARIJUANA Oregon and Alaska became the third and fourth states to legalize marijuana for recreational purposes (Colorado and Washington were the first two), while the District of Columbia repealed all criminal and civil penalties for possession and allowed limited, private cultivation of the drug.

CRIMINAL JUSTICE REFORM For the second time in three years, Californians voted to shorten the sentences of people serving time in prison. The state — which created the notorious three-strikes law — remains under federal court order to reduce prison overcrowding.

On Tuesday, the measure, which enjoyed broad bipartisan support, passed with more than 58 percent of the vote. Many politicians are still afraid of looking soft on crime, but California’s experience shows that voters can lead the way.

ABORTION RIGHTS The overwhelming rejection of “personhood” measures in Colorado and North Dakota dealt another well-deserved blow to the effort by some opponents of reproductive rights to ban all abortions (and some common forms of contraception) by passing laws giving fertilized eggs legal rights and protections that apply to individuals.

GUN CONTROL In the aftermath of the school massacre in 2012 in Newtown, Conn., Congress — caving to the National Rifle Association — did nothing to protect the public from gun violence. In Washington State, a campaign started by outraged church and community leaders fared much better. Initiative 594, which will require criminal and mental-health checks on gun buyers, drew an impressive 60 percent voter support on Tuesday.

CONSERVATION Environmentalists who may be singing the blues over the election results can take heart from approval of a record $13 billion in land conservation measures in states and cities across the country.
Again, further proof that Big Data and "predictive" quantitative politics without context is pointless.

When you get out of the cesspool and circle jerk of partisan politics in this country, what you see is how the people really are moving on a variety of social and economic issues. And it has nothing to do with partisanship, Washington, Republicans or Democrats, or the next "big victory."

Cross posted to: The Cranky Sociologists

No comments: