Thursday, January 20, 2011

Examples of the Consensus

Robin Hanson reminds us that the scientific consensus is often wrong. Ron Bailey did a Nexis search of the phrase 'scientific consensus over the past 25 years, and found the following:
  • saccharin causes cancer in humans
  • dietary fiber appeared to reduced the incidence of colon cancer.
  • agents found to cause cancer in animals should be considered suspect human carcinogens
  • fusion energy reactors would produce more energy than it consumed within five years
  • acid rain is destroying lakes and forests

These are no longer consensus findings. He did find the phrase 'scientific consensus' in regards to uncertainty about when life starts, which probably still stands. Yet in all, that's a pretty weak record for the consensus.


Erik said...

I've never quite understood your "beware of science" posts. No reasonable scientist believes that scientific consensus is always correct. We just consider it more likely correct than the alternatives (assertions supported by tradition, revelation, anecdote, etc.). Good scientists are willing to take new evidence into account, and thus science is eventually self-correcting and functioning as it should. If one wants to envision "science" that is immune to correction, then take a look at chiropractic, homeopathy and astrology. The alternative to science that is never wrong is no science at all.

Why not list some ideas where the core scientific consensus was eventually established and has then survived decades/centuries of peer review with only depth and further refinements being added (think evolution, DNA, germ theory, optics, vaccination, etc.).

If your main point is that there are bad scientists out there that cling to their biased beliefs, despite new evidence, I'm all with you, but your point seems much broader - attacking the scientific process and peer review/consensus system itself.

Dave said...

I'm not sure what the point of that last example (acid rain) was. The consensus was that it was a problem. And then new regulations and technologies were developed to ameliorate it, and now it's not so much of a problem. That doesn't mean it wasn't a problem 25 years ago; it just makes it a problem that was successfully dealt with.

TVOAT said...

I don't think he's attacking science. He's probably pointing out, as he usually does, that scientists (and economists) understand the world less than many (espcially science reporters or politicians) would claim.

Tying the results of the above are important make the news. Reporting that the uncertainties remain great, and that taking actions against them would be premature, rarely does.

Anonymous said...

I assume Eric's point is that when the "consensus" language is used, it paradoxically signals a high degree of uncertainty. In low uncertainty science, there is also a consensus, but it goes unremarked.

Foolish Jordan said...

Presumably this means the scientific consensus now is that saccharin doesn't cause cancer in humans, and therefore, since the scientific consensus is often wrong, we smart people will cleverly avoid saccharin at all cost so as to avoid getting yucky old cancer.

Or perhaps I am overstating things - it's not that the scientific consensus is exactly wrong but that the scientific consensus is actually completely uncorrelated with the truth and so we should rely on the obvious alternative to scientific consensus to guide our actions.

Anonymous said...

There is no doubt scientific consensus is often wrong.

The question remains: so what?

What should be the alternative?

The opinion of the mob?
PR-propaganda of big corps?
Politically infested speech of politcians?

AT least the scientific consensus is self-corrective, although with an unfortunate lag (re: Planck on new scientific truth advancing grave by grave).

However, what one should take away from this is that there is no substitute for personal critical study, esp. if you can do it scientifically and learn to love the truth, instead of trying to prove yourself right.

-- anttik

Eric Falkenstein said...

I'm an analytical person, so I mainly use logic and data to figure things out. But scientists, like economists, are often subject to fads because of a big idea (most macroeconomists were socialists in the 1940s), or methodology (the idea of perceptrons in AI). A lot of these ideas, like the CAPM, 'should work', in the sense one has some model of how things work, and if x then y. The map is not the territory. I think Popper's idea of science is a good normative guide, but Feyerabend's description of science is more how it really is.

Plus, I see 'the consensus is overwhelming' argument used quite a bit, and I find it quite lame.

The Pensum said...

There was a good article on this topic published in the New Yorker a couple of weeks back.

"The Truth Wears Off (Is there something wrong with the scientific method?)"

Unknown said...

Citing consensus is a substitute for doing science. A really crappy substitute.

Just keep in mind that scientists aren't being scientists when they talk about their beliefs, convictions, guesses, or impressions.

Anonymous said...

Eric, I know you're really fond of this train of thought but it seems really barren to me. Saying "the scientific consensus is often wrong" is plain inaccurate, you don't want to be tied to a statement like that. The only way it's true is if you define "scientific consensus" in some perverse way. I know you think economics is a science but it really isn't.

Anonymous said...

Reminding us that consensus can be wrong is a good thing. But consensus is usually right, probably overwhelmingly so, and Eric's post sounds, even if this is not the case, like he doesn't believe that.

Anonymous said...

Great article in the New Yorker that deals with this:

Contemplationist said...

What are you all bickering about? Consensus is important but Eric's highlighting this because of the hysterical cries of internet bandits of OMG CONCENSUS!! whenever anything is challenged regarding certain topics. This is a very poor substitute for actual thinking.

If the activists and bureaucrats masquerading as scientists stopped shrieking like harpies about consensus and "debate is over," then there would be no need to point out that consensus is often wrong.