Skip to main content

The Science of Good Government

The Obama administration talks a lot about making policy based on evidence rather than politics. A basic question remains unanswered: Which evidence?

As we went to press late in May, President Obama was giving another of the "major addresses" that signal how he wants his administration to be perceived. As with previous speeches on race and abortion, the Obamoration at the National Archives — a justification of his national security approach in general and the closing of the Guantanamo detention facility in particular — was long on nuanced, fact-filled reasoning and replete with implied failures by the preceding occupants of the White House. Clearly, this is an administration that hopes to be seen as deliberative and evidence-based rather than ideological and careless, like the one those cowboys from Texas and Wyoming ran.

As part of its evidence-based theme, the Obama administration has taken pains to telegraph its devotion to science and scientists, often using Kennedy-esque imagery that emphasizes the soaring futurism of the "hard" sciences and high-tech innovation. In the JFK era, the glamour of putting man on the moon within a decade was front and center; for Obama, the touchstone is green energy as a remedy for global warming and a substitute for imported oil.

In the political sense, the current president's science strategy is understandable. If scientists can be made culturally cool (and the longevity of Einstein T-shirts certainly suggests they can), they could work in concert with the political hipness the Obama movement has tried to evince on its march to the White House, helping policy initiatives connected to science gain traction with the public.


Among the administration's impressive list of quality scientific advisers, the poster child for the new, hard-science chic is Steven Chu, a secretary of energy so accomplished — winning the Nobel Prize for physics in 1997 is just a hint of his multidisciplinary brilliance — that his appointment to the Cabinet drained the world supply of superlatives available for media use. Beyond the unimpeachable scientific credentials, he's also telegenic, funny, self-effacing and authoritative — The Jetsons, Tom Swift and the Apollo Program all rolled into one, then updated and fused into a vision of scientific progress that will save the American economy and, simultaneously, put the world's oil-producing countries in their place. (One can almost imagine the president thinking: I bow to you, Saudi princes, but ironically; Steve Chu is in my back pocket.)

Clearly, the technologists who foster and create the green energy and transport solutions to the climate change problem are more marketable in the current culture than, say, researchers of early childhood education. Just the same, the inaccurately labeled "soft" social sciences — sociology, psychology, political science, economics and brethren disciplines that study human interaction — are at least as central to the proper workings of modern governance as the supposedly harder variety. Given the complexity of 21st-century policymaking, the government could plan and accomplish little of importance without the sophisticated data analysis that the social science industry offers. To have confidence a proposed program will accomplish its goals, the government needs detailed empirical knowledge on how people have responded to similar efforts undertaken elsewhere or at other times. To oversimplify, social science tells government whether its plans will work once people get involved.

But there's an odd disconnect in the social science-government partnership: Although the government consumes a huge amount of research into society's workings, no one, it seems, has documented clearly how the officials charged with solving our problems acquire, weigh and then use the research that comes before them.

The rather sophisticated observation I just made — that social science research underpins a lot of what the federal government does, but no one really knows how the research gets chosen or used — is not something I'm smart enough to have devised on my own. I'm borrowing it for the nonce from Ken Prewitt, a vice president for Global Centers and the Carnegie Professor of Public Affairs at Columbia University, a director of the Census Bureau under President Clinton and, for decades now, one of country's leading social scientists, both in and out of the academy.

Among Prewitt's attractive qualities — particularly from a journalistic point of view — is directness. Many academic social scientists believe that their work is largely ignored in policymaking circles, even though there have been many efforts through the years to bring more social science research to the attention of the powers that be and the general public. Prewitt acknowledges that this belief has a certain amount of validity, but he thinks it exaggerated; in fact, he calls the complaint "the long whine," and he has little patience with it.


The government doesn't just use social science research, Prewitt notes, it buys huge amounts of it — billions of dollars worth — from contract research organizations like the Rand Corp. and Mathematica Policy Research Inc. This is not to mention the mountains of think tank and university research that executive agencies, congressional committees, and state and local governments parse every year. As Prewitt notes, the National Academies has hundreds of research committees under contract at any one time, with dozens of them involving social science of some sort. The government employs social scientists from the cabinet down through the bureaucracy, and each year schools of public policy spew out thousands of people who are trained in the social sciences — specifically for the purpose of working in government and the nongovernmental organizations that help set and execute policy.

All the same, Prewitt says, there is a problem in regard to evidence-based policy formation and social science. No one, he says, really knows how and why some research gets used and other research gets ignored. There is a vague theory that once social scientists reach some sort of consensus on an issue, it seeps into the policy process indirectly, as if by osmosis. He calls the osmosis theory not necessarily wrong but "woefully inadequate."

Prewitt is chairing a committee of the National Research Council that's charged with improving the use of social science evidence in the policy process. In an overview of its mission, the committee says it hopes to answer a series of basic but vital questions, including: "What are the best ways to measure quality and promote improvement of research? How much evidence of various sorts is necessary (or sufficient) to support programs and policies? What happens if the evidentiary standard is set too high or too low? How can research evidence contribute to improved policy and practice decisions? What types of policies are most amenable to various types of research and how are research results best communicated to decision makers?"

Prewitt says it's become clear that the science base doesn't yet exist to answer those questions definitively and to put forward a solid report about how research is acquired and used by policymakers. The lack of authoritative findings in this regard is impressive; Prewitt says he can't now document, for example, whether policymakers are more likely to use social science the government pays for than science that comes in "over the transom" from other sources. "Therefore, I think we have to sort of clear out the underbrush ... and start fresh by doing some serious social science about what is actually happening in the interaction between getting knowledge and using knowledge," Prewitt said.

To say that evidence doesn't always carry the day in Congress or the executive branch is to take understatement to an extreme. As Prewitt notes, "Whatever explanatory models we come up with, they will have to take politics into account, because you don't have policy without politics."

Still, the evidence that science provides can leaven and even radically change the politics of an issue. In just a few decades, it can turn the use of tobacco from a governmentally subsidized business — even, as any episode of the cable TV hit Mad Men shows, a cultural icon — into a filthy public health threat being excluded from our public spaces as it is gradually taxed and sued out of existence.

The Obama administration's focus on high-tech solutions to the climate change problem is appropriate and politically smart; making green energy a cultural imperative could do for global warming and the Saudi Kingdom what multibillion-dollar lawsuits and pictures of cancerous lungs have done for big tobacco. By contrast, a call to clarify the process for government's use of social science can seem, I am utterly aware, a process-minded exercise in esoterics.

But if it really intends to improve the overall performance of government — to actually make it less ideological and more evidence-based — the administration needs to focus some energy on setting standards for the acquisition, assessment and use of research, particularly research produced by social scientists. The president's ambitious goals in energy, health care and education policy can't even be approached, much less accomplished, without the kind of careful analysis of contradictory research into human behavior that those standards will encourage. Once standards are set, policymakers can be more certain they are acting on reliable evidence that will work in the real world, researchers can have a better idea of how to get their work in front of government officials who can act on it, and persuadable minds in both political parties will be able to say "yes, we can" more often than "no, we just won't risk it."


When I named this column in our launch issue last year, I took a journalistic oath not to drone on about the brilliance of the articles you can read in the rest of the magazine, as do the editor's letter columns that front so many American magazines. This time out, though, I will suggest that you read "Keystone Cops at the Police Lab," Steve Weinberg's opinion essay on a scandal afflicting forensic laboratories across the country. There are reasons.

First, the essay cites just the type of authoritative, solution-oriented research we need more of, a National Academies/National Research Council report on the sorry state of the nation's crime labs in particular, and, more generally, on the plague of "nonscience science" being presented as evidence in court. Second, Weinberg — a premier investigative reporter with long and deep experience in criminal-justice matters — brings his own knowledge of bad crime labs and corrupt criminalists to bear, providing a concrete, scary picture of the way law enforcement misuses forensic testing and evidence, too often convicting the innocent and letting the guilty murder, rape and rob again.

And third, Weinberg's piece mentions a piece of the national forensics problem — the shameful failure of crime labs across the country to test rape kits in a timely fashion — that the nation's journalists need to attend to, immediately, regularly and until it is corrected. In Los Angeles County, it has long been known that police departments have backlogs of thousands of untested rape kits that contain body-swabbings and other evidence that could provide DNA samples that would help convict sexual predators.

The Los Angeles Times and other media have done a pretty good job of reporting on the rape kit problem in Southern California, which has been the focus of a Human Rights Watch investigation and an audit by the city controller. One cannot say that law enforcement agencies and other responsible officials have covered themselves in glory with their responses. "I think it was surprising how slow they were to get to this," says Sarah Tofte, a Human Rights Watch researcher who has pushed Los Angeles agencies to count their rape kit backlog. Now, though, the agencies seem to be making at least some effort to whittle down the backlog, Tofte says.


Because of its need to identify bodies in the wake of 9/11, New York is far ahead of most local governments in terms of its DNA testing capabilities, and it's making a real attempt to test all rape kits as they are submitted. Meanwhile, New York Times columnist Nicholas Kristof recently noted, the arrest rate for reported rapes has risen by 30 percent.

But last summer, Tofte wrote that National Institute of Justice put the number of untested rape kits at 400,000 nationwide. Newspapers and television stations have reported occasionally on the rape kit problem in a few other cities, but the issue has — for reasons that escape me entirely — largely remained a minor one on the national scene.

When the evidence that could help convict the perpetrators of 400,000 rapes is put on shelves in police departments and forensic labs and ignored, it's a national disgrace — and a major story that America's media organizations are failing to tell effectively. Police, prosecutors, mayors and city councils across the country need to be pushed to provide an accounting of and plans for eliminating their rape kit backlogs.

Rape is not a minor crime. Police who don't investigate it properly are failing to meet their solemn duty to serve and protect. Metropolitan daily newspapers that don't respond to this failure with teams of dogged reporters and sustained editorial-page outrage are, in a very real sense, enablers of sexual abuse.

Miller-McCune welcomes letters to the editor, sent via e-mail to; via the comment sections of our Web site,; or by standard mail to The Editor, 804 Anacapa St., Santa Barbara, CA 93101.

Sign up for our free e-newsletter.

Are you on Facebook? Become our fan.

Follow us on Twitter.

Add our news to your site.