The expert exception to consensus reliance

I've really enjoyed Chris Mooney's Point of Inquiry podcasts, so it's too bad the latest one with pro-GMO activist Mark Lynas failed to wrestle significantly with real arguments about GMOs (Chris, you talk about GMOs too much to say you don't want to delve into technical issues).

One interesting issue did come out of the podcast - at one point Lynas says the Union of Concerned Scientists' rejection of the National Academy of Sciences position on GMOs is a contradiction of UCS arguments that we should rely on the consensus opinion on climate change.  I've argued something similar before, that the expert consensus should basically just plug in as fact in policy discussion, but there's one exception - if you yourself are an expert in that field, I think you could conclude that you're right and the consensus is wrong.  Presumably UCS could make this conclusion.

It takes a lot of hubris to conclude you're right and the other 90% of the experts are wrong. You might carefully examine whether you're missing something little or something big, and withhold judgment for a while. In the end though if you feel like you really understand the issue, you don't have a choice but to conclude what you conclude.

So what should happen next depends on who you are.  If you're the dissenting expert, you should work to move the consensus in your direction.  If you're someone else, the dissenter is just a dissenter, and only if there's enough of them to disrupt the consensus do we need to worry about them.

So no change to the big picture, but I disagree with Lynas that experts are required to toe the line of an expert consensus when they disagree with that consensus.

Precautionary principle possibilities

The options:

1. Prove a negative, or no dice.  No technology can be used unless it's proven in advance to never have risks.  That can't be done, so we can't use any technology, new or existing.

2. Risky, smishky.  Defined risks are no reason to stop or alter use of a technology unless the risk has moved from possibly harmful to scientifically-certain harm.  With no certainties in science, this also can't be done.

3. Squishy real life choice.  evidence rising to a level of x for harms whose severity rises to a level of y compared to the technology's benefit of less than z is the point at which you stop or alter the technology's use unless and until additional research changes the value of those variables.  The only minuscule problem with this is it doesn't provide much of a guideline.  But it's right.

Discussion below in this blog and at Stoat are relevant.  Anti- and pro-GMO forces argue for positions 1 and 2 although they're usually vague about it, as a clear description of their perspective doesn't help them win.  Switching over to variables x, y, and z, pro-GMOers like Keith Kloor obfuscate the non-zero value of x, while anti-GMOers exaggerate the other way.

So what to do?  Caution, I guess.

To try to move from saying something squishy to something interesting, how about this:  prior to the 1970s, the correct policy conclusion was that CFCs have no broad-based environmental harms and should be used liberally.  That's not right, it turns out, but that's the information we had at the time.  The key is moving quickly as new evidence for variables x and y came out.  From a political science/international coordination perspective, the world moved amazingly quickly on ozone-depleting chemicals.  From an environmental perspective, my sense is that we barely dodged a bullet (and we still have another 40 years or so of slowly declining CFC concentrations where something new could go wrong).  We'll see how well we'll do on climate change with already, well-established x and a range of bad to horrible values for y.


Moderately relevant:  I've had a sci-fi story in my head for a while, an alternative history where CFCs were invented and widely-used a century earlier than in real life and long before the science could anticipate their problems.  Society would be finding nature melting and cracking around them, along with peoples' skin, and have no idea of the reason.  Not a particularly happy story.  So, caution.

Four out of five dentists say it's the scientific consensus, stupid

I'm late to the game of the Cox/Ince editorial on science and policy, but two points:

1. It's not "science" that matters most at the intersection of science and policy but scientific consensus that does.  If there's a well-established consensus with very few expert dissenters, then you've got factual conclusions as far as policymakers are concerned.  The consensus could be wrong of course, but that's not really relevant to policymakers - they don't have a choice of waiting for a perfect consensus because that won't happen.  What's missing from the commentary I've seen is that policymakers also don't have the choice of second-guessing the consensus by becoming their own Galileos.

Let's forget climate change for a while and take my constrained policymaking field instead.  Should I direct my water district to add fluoride to our water supply?  The vast majority of dentists say it helps teeth, but some disagree.  Does adding fluoride create non-dental health risks for the general population?  The vast majority of oncologists, endocrinologists, and neurologists reflected in the scientific consensus say somewhere between "no" and "not proven," but some disagree.  It's ridiculous to think I could get sufficient expertise in those three fields as well as dentistry to let me judge between experts.

And fluoride is just one issue.  What about water supply decontamination - should we use chlorine or chloramine?  What is the maximum horizontal acceleration that a worst-case earthquake will exert on each of our eleven dams?  Is the stream gauge appropriately sited to be provide accurate data on stream flow during storm events?  How much chromium 6 is tolerable in our stored groundwater?  Can migrating steelhead trout make it up the proposed fish ladder? I'm not second-guessing these things if they have an expert consensus behind it, regardless of a few dissenters.

Incidentally, I don't distinguish between scientific consensus and other expert consensus.  Seismologists tell us the maximum acceleration during an earthquake, and engineers tell us what dams can handle.  I don't see a difference.

The intersection with policy gets complicated if the consensus has an important dissenting faction or if there's no consensus at all, but that's not what we're facing on some issues, like global climate change.

2.  Contrary to what scientists often say, the science can all-but-decide policy because some policy questions are easy.  Scientific predictions of climate change in the next century under business as usual emission scenarios run from "bad" to "potentially disastrous".  That's pretty much all we need to know from a policy perspective to conclude that we have to deal with it.  The question of how we deal with it isn't easy or exclusively a science question, but whether to deal with it was answered by science.