7 min read.Updated: 05 Jun 2014, 05:41 PM ISTG. Sampath
The most interesting bit in Silver's piece--where he basically damns Piketty with faint praise before ending with some generalizations on data analysis--are his confessions
It is a truth universally acknowledged that more people have opinions on Thomas Piketty’s tome, Capital in the Twenty-First Century, than have actually read it. Nate Silver is not most people. He is the golden boy of data journalism. But that hasn’t stopped him from pronouncing on Piketty’s book without having absorbed all the data, as it were—the words printed on its 696 pages.
In a recent column titled Be skeptical of both Piketty and his skeptics, Silver wades into the squabble between the Financial Times’s Chris Giles and Piketty over the latter’s choice of data sets. A quick summary of the exchange can be found here.
The most interesting bit in Silver’s piece—where he basically damns Piketty with faint praise before ending with some generalizations on data analysis—are his confessions. “A series of disclosures: First, my economic priors and preferences are closer to The Economist’s than to Piketty’s. Second, I haven’t finished Piketty’s book, although I spent some time exploring his data. Third, I’m no expert on macroeconomic policy or macroeconomic data."
The American economic analyst Doug Henwood has parsed this passage thus: “So, I have a bias on a field in which I’m a naïf, and I’m going to comment on a book I haven’t finished reading yet, but listen to me because I’m Nate Silver." Silver’s disclosures, which prompted Henwood to wonder if the statistician-turned-journalist was “an idiot savant", are revealing in the way they slyly smuggle in a quantitative term (priors) to refer to what is essentially a qualitative concept (ideology).
“Priors" is a term from Bayesian statistics. But as the political scientist Corey Robin has pointed out, Silver uses it here as a “stand-in for ideology". Unless we take “priors" as an attempt at metaphor, Silver’s statement that his “economic priors" are closer to The Economist’s than Piketty’s make no sense.
That “priors" has been deployed as a metaphor for ideology, by someone who happens to be an iconic figure in the world of what has come to be called “data journalism"—a field that likes to imagine itself as transcending all ideology and all metaphorical thinking—suggests that it is less a one-off stylistic quirk and more a symptom of the larger project of de-politicization that has come to dominate political, economic, and cultural discourse.
The politics of data
This column is not about Silver or data journalism as such, but about what the viral spread of data-centric discourse says about the world, and how, by reframing qualitative questions in quantitative terms, it is changing the way we construct our truths about it. One way it brings about this change is by, as sociologist and blogger Peter Frase puts it, “pretending that political and moral questions can be reduced to empirical ones".
So it might be worthwhile to spend a little time on the kind of data journalism advocated by Silver, best exemplified in his new media start-up, FiveThirtyEight.com.
The Silver school of journalism has come in for heavy criticism before. He has been faulted for favouring outcomes over processes, for obscuring ideas and ideologies, his neutrality has been questioned, and his hidden biases exposed.
Paul Krugman has even taken issue with the way he uses data. As Krugman puts it, “You use data to inform your analysis, you let it tell you that your pet hypothesis is wrong, but data are never a substitute for hard thinking. If you think the data are speaking for themselves, what you’re really doing is implicit theorizing, which is a really bad idea." Well, a more common word for “implicit theorizing" is ideology.
Another critic, Malcom Harris, editor of The New Inquiry, has labelled Silver’s approach “Actually Journalism"—something that is “not quite reporting and not quite opinion writing".
Silver himself has described his journalistic method as consisting of four steps: collection, organization, explanation, generalization. As Harris was quick to point out, “It is no coincidence that interpretation isn’t on the list." Rather, in Actually Journalism, “news and opinion aren’t needed; to understand something, all the audience needs is the fact, this piece of data, this answer".
In India, we saw this phenomenon manifest itself most clearly in the media coverage of the 2014 general elections. In-depth news reports about the election itself were sacrificed in favour of shouting matches organized around assorted data—most often to do with exit polls and voting patterns. An exit poll, just to clear the air, is not news—it is a data collection process. It becomes news when it is disseminated as such. However, large sections of the Indian news industry, instead of concentrating on collecting and disseminating news, chose to spend their scarce resources on collecting and disseminating data.
But the hegemony of data is not restricted to politics alone, even though the media is central to the enforcement of its hegemony. In the Indian context, perhaps no single phenomenon epitomizes the triumph of data over reality as much as Sachin Tendulkar, a sportsman whose sporting legacy career would perhaps have been far greater had his career and outlook not been usurped by the statistical imperative of aggregates and averages, records and milestones.
But the biggest ideological function of data has been in the realm of economics—not so much the classical economics of Adam Smith or David Ricardo, but its 20th century neo-classical avatar. It is data and statistics that have converted the social problem of poor people into the economic problem of poverty. Quantification is the magic wand that transforms a reality tainted by the dirt of perception and the dross of ideology into the gold of data.
If the early 21st century has come to be known as the age of Big Data, there is good reason for it. Data is ‘big’ not only because it is now generated, solicited, collected, extracted, hoarded, bought, sold, and bargained over on a scale never seen before. Data is big today because it signifies a merger of power and ubiquity. The word ‘big’ in Big Data means essentially the same thing it does in Big Brother. Far from being innocuous units of academic study or disinterested analysis, data is becoming a potent weapon against two elements that have always been at the heart of human resistance to tyranny: ideas and imagination.
Data as an antidote to ideas
If defeating ideas and imagination is one part of the ideological battle where data-centrism has a role, another is the re-creation of reality along the lines that are acceptable to the ideological consensus of the power elite.
A datum is a unit of fact. A fact is a particle of what we consider knowledge—knowledge, that is, of reality. So, the power to manufacture, control, and disseminate data also gives you the power to manufacture reality. In the past, data-based ideological operations passing themselves off as science or journalism have claimed—with successful impact on public policy—that global warming is a myth, that cigarettes don’t cause cancer, or that economic policies that pauperize a vast chunk of the population are the best way to reduce poverty.
Silver writes that the aim of his media project is to present “the truth beyond our perceptions". What better way to do so than by burying perceptions under data? It should therefore come as no surprise that superior truth claims would be made for statistical analyses of quantitative data than for the intellectual processing of perceptions and ideas through the application of reason and imagination (more popularly known as thinking).
In this battle, the dice is loaded in favour of whoever controls the discourse. At present, it is the data fundamentalists—by this term, I mean those who believe that the path to truth and enlightenment goes through data—who seem to have the upper hand. In a world dominated by data fundamentalism, the only way to make a truth claim is in terms of data. And the only permissible way to verify/dispute those truth claims would also be in terms of data. In other words, non-aggregated, non-abstracted reality never need come into the picture.
That is why it is possible for Silver—despite all that he reveals in his “disclosures"—to pass judgment on Piketty. For Piketty, data is one of the crucial raw materials for thought. For Silver, data “speaks for itself", and speaks more eloquently, than any thinking person can. But the belief that data can “speak for itself" is just that—a belief, and it is this belief that Silver felt compelled to mask under the statistical metaphor of “prior". There is a word for beliefs that are “prior" to data: ideology.
Aggregated statistical data are easier to manipulate than the individual parcels of sensory data imbibed directly via lived experience by live human beings. It is, therefore, ideally suited to serve as the dressing that enables the contraband of propaganda to pass through the customs of truth and circulate among the public as common sense.
All this does not mean, of course, that statistical data are all bad or that they do not have their uses. You cannot fly an airplane or predict the weather or run an economy without accurate data and proper analysis. But data works best when it is in its place—as a means to an end, not when it becomes an end in itself. In fact, it is at its most dangerous when it hijacks public discourse and begins to dictate or limit what the ends should be or could be. Why? Because the ends are always determined by political and moral questions (whether acknowledged as such or not), and those can only be answered in political and moral terms, not empirical ones.
Subscribe to Mint Newsletters
* Enter a valid email
* Thank you for subscribing to our newsletter.
Never miss a story! Stay connected and informed with Mint.
our App Now!!