Science can drive policy

I'd like to pick up a point that Dan Sarewitz made about the role of science. If I take his point about Bush v. Gore not being a matter of "science," in the sense that deciding who was going to become President was better handled by a "political" or "legal" framework, which is set up to make decisions by process. Leaving it to science would take a long time and would entail lots of dispute and might not reach resolution, and in the meantime, we would need a President.

OK. I think I agree, but it depends how far to take it, and over what time scale. It is surely a matter of empirical fact that one or the other candidates got more votes in Florida in Nov 2000, as measured by the intention of voters. The problem was capturing those votes accurately. And since that had been hopelessly botched, something other than a pure correspondence method was needed to decide who won. But in fact there are much better ways to ensure that votes are counted the way people voted, and we've got serious problems counting (see this weeks Rolling Stone about voting in Ohio in 2004). But I don't think the story ends there, and it's scary that it seems to have ended just there, without the credible process for doing a better job of counting votes. It is indeed a matter of technology to come up with reliable voting, and it is a matter of science to count well. And it is a matter of science to ensure that correspondence, and not just process and coherence, matters in the long run. So using the courts short-term is fine, but that solution will produce questionable elections in the long run. That's a matter for science to help solve.

Another example of a similar outcome is the seemingly bizarre decision of the Supreme Court that the Constitution mandates a rather unreliable and inaccurate means of counting heads for the census, because that must have been what the framers meant by enumeration. Never mind that statistical methods would give an answer closer to how many people actually live in various parts of our country (which an NRC committee laid out in a report). Apparently the thought is that Scalia, et al., can know that folks like Benjamin Franklin and Thomas Jefferson would have wanted us to persist in inaccurate methods that seem more intuitively obvious, and would have rejected advances in discrete mathematics and statistics. If I were to guess, that churning we heard the day that decision was announced was some framers of our Constitution who had serious scientific talent turning in their graves. Strict constructionist indeed.

Seems to me science plays from strength when it comes to correspondence criteria of truth. Apparently the Court is happier dissecting words and relying on process, but that's pretty perilous when the facts matter. Recall that one of our best justices ever, Oliver Wendell Holmes, went on and on about process in Buck v. Bell--the problem was that the facts and the underlying science were a crock. And that was a Supreme Court that seemed to have a lot less penchant for ideological zeal than the one we've got now. I worry.

So Dan may be right that sometimes we have to fall back on process and coherence as our criteria for making decisions. But that does not mean that science should not trump those criteria if it can get its act together. It is surely a technically achievable feat to accurately count people in the census and votes in elections much more accurately than we do now.

Here we have two cases where our highest court has basically reached a decision that almost surely means that a shoddy process is used to make an important decision, where better methods of counting would give different answers. And those answers matter--for representation in Congress and allocation of federal dollars in the census, and for who got to lead our country 2000-2004. Dan argues that the Bush v. Gore decision was better to leave to the courts than to hope we could sort out the count. OK, but that's far from the end of it. I'm *really* not happy with leaving the solution there. Bad counting matters, big time. And it appears that all three branches of government are perfectly happy with bad counting. If one were deeply cynical, or perhaps just awake, one might think it is because those in power are happy being there and would like to twist the rules to stay there. Plenty could be done to give us a better census count and more accurate voting. Science, math, and technology have a lot to contribute. So while Dan may be right about some decisions not being best left to science in the short run, leaving crummy methods in place--or mandating their permanence from the highest court--is very dangerous in the long run. And that's a matter for science.

OTA RIP: sad reunion

A month ago, the American Association for the Advancement of Science hosted a bittersweet event in its New York Avenue Headquarters in downtown DC, a reunion of former staff and “friends” of the congressional Office of Technology Assessment (OTA). OTA was a victim of the 1994 congressional revolution. The newly elected Congress eliminated the congressional agency with the weakest constituency. The favored joke among OTA staff was that Congress started by cutting off its head. No kidding.

We got together to commemorate OTA’s demise after a decade. Looking back, it feels like OTA’s death was a harbinger, a foretaste of continual partisan combat over S&T policy, a political domain once relatively nonpartisan (with some conspicuous exceptions, such as Star Wars). The new Republican majority distrusted OTA, suspecting it was a nest of covert liberal Democrats disguised as analysts. So kill the enemy.

The Republicans were correct that most OTA staff were more liberal than the new Members of Congress. But OTA was also scrupulously bipartisan. And that’s why its absence is most significant. Institutions seeking common ground find themselves in a killing zone.

The very first lesson I learned at OTA was to listen to all stakeholders, across the ideological spectrum, and to be rigorously disciplined about meeting with and paying attention to both parties. The process was systematically bipartisan. Bipartisan is not the same as nonpartisan or objective, but it’s a damn sight better than what we’ve got now.

Without an OTA equivalent, S&T advice is channeled through external constituencies and the executive branch, which is inherently administration-dependent. Congress has lost most of its S&T analytical capacity, and the executive branch has lost its credibility. Have we given up on bipartisanship, resigned to polarized S&T war rhetoric, without even the option of consensual, incremental building?

Mr. Gingrich and Senator Clinton, tear down that wall! Then build sturdy structures that restore a bipartisan ethos. A good place to start is science and technology policy, where partisanship is particularly stupid and destructive.