Sunday, September 12, 2010

How your beliefs shape your perceptions: If a tree falls down in Smolensk, how thick is it?

Quick: how thick is the tree in the picture? What I mean is, what is its diameter at the point where it broke? Take a guess and make a mental note of it before you read on. You can click on the picture to enlarge it. Don't overthink it though; just pick a number that seems reasonable to you and move on.

Republicans are much more likely than Democrats to think that Obama is a Muslim who was born in Kenya. This correlation doesn't make any sense to me, and to a whole lot of others. People become Democrats or Republicans because of their values or policy preferences; Obama's faith and place of birth are not a matter of values or preferences. They are a matter of fact. There is no a priori reason why people with different values should have different judgments in factual matters. It really doesn't make any sense whatsoever. The only reason why it doesn't baffle people as it should is because it's something that happens all the time, so we just get used to it and forget how utterly insane it really is.

Someone might object here that it's perfectly reasonable for people with different perceptions of facts to have different values, so the observed correlation may not be so strange after all. If you think Obama is a foreigner and a Muslim, perhaps you're more likely to become a Republican. To this I would object that, in the real world, causality seems to run the other way much more often. Most of the people who have strong beliefs about Obama's spirituality and birthplace chose to be Republicans or Democrats long before they even heard of Barack Obama. It really is very striking: instead of choosing policies that match the facts as they see them, people choose to believe the facts that back up their values-driven policies.

What does all this have to do with a picture of a broken tree? A lot, actually, because it's not just any old tree. This tree has tremendous political importance. According to the Russian Interstate Aviation Committee, which is investigating the recent crash of the Polish presidential airplane near Smolensk, Russia, the tree in the picture is broken because, in the final seconds of its flight, the presidential aircraft has hit it with its left wing. It's hypothesized that the impact caused part of the left wing to snap off, which in turn caused asymmetry of lift, forcing the aircraft into a complete roll to the left, after which it hit the ground inverted, killing everyone on board.

OK, but again, what does all this have to do with facts and values and Obama being born in Kenya? Well, lots of Poles, and some Polish media are convinced that the Smolensk crash was in fact an assassination organized by Russian special ops forces. The "accident vs. assassination" divide cuts deep and seems to be driven mostly by values and preferences just like political ideology cleavages are. At any rate, if you believe in a deliberate act of destruction, you must think that the official story of the airplane losing part of its wing on a tree, rolling to the left and hitting the ground upside down cannot be correct. Most assassination-believers claim that it's impossible for the aircraft to have lost a wing on this particular tree because it isn't thick enough.

Which brings me to my point. As I read discussions between accident- and assassination-believers that go on in Polish blogosphere about whether it is or is not likely that this tree has indeed destroyed the airplane's wing, it's very hard to believe that all those people are talking about the same damn tree. Most assassination-believers claim that the trunk is "at most 40 centimeters [about 16 inches--p] in diameter" where it broke. (I've seen claims going as far as 5.5 inches.) Most accident-believers claim that the diameter is "at least 50 centimeters" [about 20 inches]. All of them are basing their opinions on the same pictures.

So there you have it: values shaping perceptions, and very basic perceptions at that. Here's an idea for a fun experiment. Show the picture above to a large group of people who have never heard of the Smolensk crash, and ask them what they think its diameter is. Then show the same picture to a large group of people half of which are accident-believers and half of which are assassination believers, and ask them the same question. How would the answers differ? I don't know if the mean of the guesses in both groups would be different, quite possibly not. But I think the distributions would be nothing alike. The guesses of the first group would be distributed normally, whereas in the second group they would follow some sort of a bimodal distribution. However, one simple additional twist in the experimental design would make that difference disappear. Go to Smolensk and measure the actual diameter of the tree. (It can be done; the tree is still there.) Once you've obtained the correct answer, gather your two experimental groups and ask them to guess the diameter, but this time tell them that everyone who gets within an inch of the correct answer will be paid $500. You will see bimodality in the second group's guesses disappear as if by magic. You will see both distributions assume a normal shape with roughly the same mean, which will be roughly equal to the true value. Such is the difference between situations in which people form their perceptions of facts in order to signal their preferences, and situations in which people form their perceptions of facts in order to find out what those facts actually are.

To sum things up, and in case you were wondering, my own guess is 50 centimeters, plus or minus 10 centimeters (which is about 20 inches, plus or minus 4 inches). My wife's guess was "a foot an a half," or about 47 centimeters. (She has heard of the Smolensk crash, but didn't know what tree she was looking at.)

No comments:

Post a Comment