Sunday, February 7, 2010

Expectations

Literally snowbound, it's hard to come up with excuses not to blog. And I suppose, on a frigid day, it's not too unusual to be put in mind of old flames.

Mind you, this wasn't someone I actually dated — more an unrequited torch, before th'wife, even before the Before Time. We're talking what cosmologists call Deep Time, back around the point when God was thinking maybe the neutron might be a good idea.

Her name was Kathy Ewing, and she was my Self, Culture, and Society lecturer at Hogwarts [1]. She was young, pretty, and smart as all get-out — and I had a fantastic crush on her, though doubtless she wouldn't remember me at all by now.

At least, I hope to God she wouldn't, the reasons being imminently obvious.

Required background: Like many of the common-core courses [2] at the University of Chicago, Self, Culture, and Society had two hour-and-a-half discussion sections each week, in which Kathy and about a dozen of us would discuss, argue, and hash out the course materials — in this case, a “Great Books” [3] mix of psychology, anthropology, and social science.

A side-note: the U of C is the intellectual equivalent of someone throwing a knife to you and hollering, “Now, come at me!” It's exactly the kind of place that pussy little right wingers like to bitch about being hostile to their ideas — only they're missing the point: it's hostile to everybody's ideas. It's all about forcing you to defend your beliefs, structure your arguments so that they make sense somewhere else besides the addled interior of your head ...

Anyway.

I don't remember whether it was every week, every two weeks, or what, but every once in a while all those little discussion sections met in a big hall for a lecture. The professors and lecturers teaching the study sections [4] would rotate this duty, each taking, I suppose, a topic of particular interest or expertise on which to hold forth.

So this particular upcoming lecture was Kathy's, and I was determined to sit front and center, nodding sagely at all the appropriate points, impressing her with my interest in her topic — and, strange as it seems now, thereby my interest in her. Not that I had it planned out even that clearly.

Only the night before, a good friend who will remain nameless had the latest in a series of fights with her boyfriend, another nameless good friend, and I spent much of the night and next morning sitting in a stairwell offering fantastic advice [5]. Didn't really get any sleep to speak of that night.

So picture me the next morning, bleary eyed, too late to get coffee at the dining hall, stumbling into the lecture hall, only just barely conscious enough for a fogged corner of my mind to remember the plan of the previous day.

Unfortunately, all it remembered was, “Sit front and center.”

I'm sitting there, then, about 10 minutes into Kathy's talk when the realization comes over me that there's no way I'm staying awake. Just can't keep the eyes open. I did my best to hide it — the old, cover-your-forehead-with-your-hand-while-you-look-down-at-your-notebook thing, the hunch-down-over-the-table thing, every trick I could think of. Needless to say, even five minutes afterward it didn't seem likely that I'd had much luck at hiding from Kathy that I was sleeping through her lecture.

Dashed expectations are the subject of today's entry, a bit from C. Eisenegger and pals at the Institute for Empirical Research in Economics in Zurich and other environs, about testosterone.

Now, in the words of Robert Sapolsky — of whom I'm an admirer — in his wonderful essay The Trouble with Testosterone, here's the gist of the results when you inject testosterone into a submissive male monkey:

“Take that third-ranking monkey and give him some testosterone. None of this within-the-normal-range stuff. Inject a ton of it, way higher than what you normally see in rhesus monkeys, give him enough testosterone to grow antlers and beard on every neuron in his brain. And, no surprise, when you check the behavioral data, he will probably be participating in more aggressive interactions than before.

So even though small fluctuations in the levels of the hormone don't seem to matter much, testosterone still causes aggression, right? Wrong. Check out number 3 more closely. Is he raining aggressive terror on everyone in the group, frothing with indiscriminate violence? Not at all. He's still judiciously kowtowing to numbers 1 and 2 but has become a total bastard to numbers 4 and 5. Testosterone isn't causing aggression, it's exaggerating the aggression that's already there.”

So you get the background here: testosterone as a vehicle of aggression, of conflict, if not social dominance.

Imagine Herr Eisenegger & Co.'s collective surprise, then, when they gave sublingual testosterone to a bunch of women [6] and then had them play one of those social-strategy games: In this case, they gave one woman $10 [7] and told her that she had to make an offer to another — to give that other woman $5, $3, $2, or nothing. If the second woman refused the offer, nobody got anything. If she accepted, she got what was offered and woman number one kept the rest. So there are two separate motivations: the less woman one offers, the more she gets to keep; but if she doesn't offer enough, woman two can play the spoiler.

The expectation, of course, is that testosterone will bring out woman one's inner total bastard, to use Sapolsky's words, and offer less, even at the risk of losing it all.

You've probably guessed that it didn't come down that way: in fact, testosterone made woman one offer not less to woman two, but more. On the average, she offered about fifty cents more than when she hadn't been given the testosterone.

This isn't necessarily a surprise; some experts the investigators had polled beforehand had actually predicted this result, on the grounds that testosterone would enhance the women's desire to take leadership: and one way to establish leadership is to make a more generous offer that makes the other gal more likely to trust and follow you.

But that's not the interesting bit. Eigenegger etc. then took their data, and sorted it differently. Instead of separating the “women one” who'd gotten testosterone vs. those who got a placebo, they separated by which of the two the women thought they'd gotten. And guess what? The women who thought they'd gotten testosterone acted, if not like total bastards, then at least like bastards: on average they offered a dollar less than the women who thought they'd gotten placebo.

That's right, the expectation of testosterone was more potent in terms of both size of effect and statistical significance than the real thing was. And it had the opposite effect of the real thing. For all the world, it looks like testosterone's bad rap in popular culture carries more weight than its biological effects in our little brainbones.

Deep waters here: did expecting testosterone make gals one a bit on the bitchy side, or did the bitchier girls expect they'd be given testosterone? And since the anti-placebo effect was stronger than the real thing, what thence?

Our authors did a bunch of work to control for various complications; you can read more about it in their paper. As always, a single experiment isn't going to be gospel. But it does give us a wallop of a lesson in being careful about our expectations.

I don't know what expectations Kathy had of me, if any — for obvious reasons I never had the guts to ask her about that day. But I got a disturbing window on the question a few years later, as a senior, when I ran into a fellow student from Kathy's study section.

When we figured out that we'd been in the same class, his eyes first narrowed, a bit angrily. Then the light bulb went on, and he said, “Oh, I remember you — you always asked the stupidest questions!”

Oh well; expectations be damned.

[1] A prospective student I was interviewing for the college put this metaphor into my head. Geeks surrounded by stone: God, yes. At Harvard, Heather may have been a resident tutor at Slytherin, but we had Lord Voldemort — I mean, Milton Friedman.
[2] U of C is one of those liberal-arts-and-a-bit-angry-about-it places — everybody shares the same common core of science, social science, and humanities courses before they get to take their major requirements and electives.
[3] Mostly dead white European men, though there were a few chicks in there —
Mary Douglas and the like. And I understand they've been broadening it since.
[4] And I think most of them actually were lecturers or professors of varying type, and not grad students — that's another thing U of C is serious about to the point of it being a “thing.”
[5] I don't know what's harder to believe — that we actually did this, or that anybody would want to hear my advice.
[6] This, of course, is a loose end to the study: would the result have been different had you done the experiment with men? It's a quandary, though, since in men there'll be a normal variation between individuals and from day to day, and so, like the monkey, you have to give a snoot-load of it to make sure you're significantly changing what's already there — and then you have to worry about non-natural effects of what amounts to an OD of testosterone. And they won't let you do that with humans anyway. Women, at least, are more of a blank canvas, though you do have the possibility that their brains won't react in the same way that mens' will.
[7] Actually, it was “monetary units” — maybe euros — but you get the idea.

2 comments:

BPL said...

Don't be too hard on Milton Friedman. He was a libertarian, but he was no anarcho-capitalist--his negative income tax was just the kind of thing hard-code libertarians oppose on principle. And he was kind enough to correspond extensively with me before his death, and praised a paper I sent him, amateurish though it was. So I can't help liking him a little, even if we were far apart on politics.

And the money supply really does affect the inflation rate.

BPL said...

I meant "hard core," of course, though if they write software they are undoubtedly hard-code as well.