POWIP Piece of Work In Progress

Cognitive Bias – Even Scientists Suffer It

No, this post is not about AGW. Some people are using this Wired article: "Accept Defeat: The Neuroscience of Screwing Up" as one of a myriad of explanations as to why scientists would screw things up so monumentally [like not noting heat transfer of hot air to supposedly hotter ground in their models], but considering the cherry-picking of data all around, and the politics and money at stake, I don't think that's what is being covered here. I think this has more to do with problem-solving, and problems being prematurely "solved".

The main example sandwiching this article is Penzias and Wilson, the guys who discovered the universal background radiation which is some of the evidence for the Big Bang. The thing is that stuff always screws up with experiments, and as an experimenter you're trying to fix the particular problem. In their case, they made a very sensitive radio telescope and they picked up "noise" everywhere that was screwing with what they were trying to observe. They had thoughts as to what was causing the trouble - man-made sources [nope, didn't matter where they pointed the telescope], pigeon shit [scrub scrub scrub... dammit! It won't go away!], fallout from recent nuclear tests [nope, signal not diminishing over time]... they had to get rid of the damn noise so they could look at what they wanted to.

They didn't think that "noise" was a real signal [not initially, though they did muddle through and figure it out eventually].

I've dealt with this with numerical techniques with people who haven't been trained in the theory behind them, and how these things fail. I remember talking with a physics grad student who was working on a model of the sun, and he was getting an oscillatory error term that was growing in a worrying way. I asked him if he checked on whether the set-up was well-conditioned, and to think about the numerical integration algorithm he was using. His reply: Meh, I'm just going to double the precision. Uh, yeah, good luck with that.

The issue is that one has prematurely decided on what the problem is, and what the solution is. In Penzias and Wilson's case, one of the things they tried was basically ignoring the noise and making measurements that didn't require the level of precision at which point the "noise" showed up. [opposite of my "double the precision" example I have above, come to think of it]. If they hadn't gone back to keep testing, and to figure out where that "noise" came from, someone else would likely have been winning that Nobel Prize later.

And note, while the article goes on about scientists being heavily influenced by their initial assumptions, the examples given do end up with the problems eventually being solved. [because we don't hear about the failures...that's a different kind of bias].

This is why other people are so helpful: They shock us out of our cognitive box. “I saw this happen all the time,” Dunbar says. “A scientist would be trying to describe their approach, and they’d be getting a little defensive, and then they’d get this quizzical look on their face. It was like they’d finally understood what was important.”

What turned out to be so important, of course, was the unexpected result, the experimental error that felt like a failure. The answer had been there all along — it was just obscured by the imperfect theory, rendered invisible by our small-minded brain. It’s not until we talk to a colleague or translate our idea into an analogy that we glimpse the meaning in our mistake. Bob Dylan, in other words, was right: There’s no success quite like failure.

And this is why the specific examples of hard science are so important - and keep in mind, the examples being looked at aren't politically charged, they'll probably be bringing in grant money no matter their results - these are people who really want to get it right. And even they screw up in making assumptions, but some put in systems [like these lab meetings, with possible diverse audience] that make it more likely to undo a premature jump to conclusions.

Let's look at the recommendations:

How to Learn from Failure:

1
Check Your Assumptions
Ask yourself why this result feels like a failure. What theory does it contradict? Maybe the hypothesis failed, not the experiment.

2
Seek Out the Ignorant
Talk to people who are unfamiliar with your experiment. Explaining your work in simple terms may help you see it in a new light.

3
Encourage Diversity
If everyone working on a problem speaks the same language, then everyone has the same set of assumptions.

4
Beware of Failure-Blindness
It’s normal to filter out information that contradicts our preconceptions. The only way to avoid that bias is to be aware of it.

Again, this is nice to think from the point of view of hard science [or use this to flog the AGW-proponents or climate change-deniers], but this goes well beyond in terms of problem-solving and decision-making.

This being a blog with a lot of political content [and most of us co-bloggers being conservatives], one can obviously look at how this applies to us and to our own blog-reading. Seeking out the ignorant is very easy to do online, but often the other items are harder to find happening.

I had my own "check your assumptions" process over the last decade. I am in a very different place with regards to a whole bunch of issues now than I did ten years ago for a variety of reasons; the biggest items being someone living in NYC from 1996 - 2007, getting married, and having kids. Also, going from relative youth to middle age. Nothing like a creaky back to make you rethink some stuff.

Only too often, though, neither do we check our assumptions, nor seek out diverse voices [and making excuses for political failures for one's "side" has been around forever]. Lots of people read blogs and pundits only to the extent that they agree with them. Yes, I tend to hang out amongst like-minded people, too, but I like reading blogs and writings where I don't necessarily agree with the author [and there's good discussion amongst commenters besides]. For example, I love reading Camille Paglia's too-infrequent pieces [at Slate, I think?] though I disagree with her on a lot of stuff, she has a very interesting point of view, incisive writing, and spurs a lot of interesting discussion. Mickey Kaus is another.

In any case, go forth and seek diverse voices. You might end up in a place you weren't expecting.

[It's Science!]

Meep

Meep is a member of the Irish Catholic mafia, having a suspiciously high number of green-eyed, red-haired friends. While she doesn’t have red hair herself [except when she goes into the sun (rare for any vampire)], she does have green eyes. She’s a raving Papist and is a life actuary on the side [i.e., she counts dead people]. An amateur pain-in-the-ass [willing to go pro!], she likes covering retirement, mortality, math, and education issues.