I took a break from doing my taxes today, took at the NYT Magazine, and had a rush of the warm feelings of love for social science that drove me into social psychology in the first place, back in the day. So my taxes can wait a while. Every so often I read a social science article that I want everyone I know to read because it will lead to our future discussions taking place on a higher intellectual plane.
Don't be fooled by the stupid headline about Justin Timberlake: this is one of those groundbreaking articles. It discrubes a series of experiments conducted by Duncan Watts, a sociology professor at Columbia, and his colleagues on the development of cultural preferences. The study involved music preferences, but the ideas apply as well to choices among political positions, candidates, blogs, diaries, and comments. All meta discussion of such topics should rely on this concept as a launching point.
And, being in the NYT, it will only be up for free for a week, so you'd better read it now. More below.
The theories being discussed here aren't all that new, but the points are nicely made and the experiment in question is elegant and lovely. (It was evidently reported on in Science last year, but this is apparently its launch into the broader public sphere.) The authors have set up a website called musiclab where they invite people to listen to and rate songs. Some people on the site saw only the names of the songs and the bands; others also saw information about how often that song had been downloaded by other registered users.
The latter sort of information -- feedback on others' preferences -- is part of what is broadly called "social influence": how our beliefs, actions, intentions, and behaviors are affected by those of others. This nicely allows the authors to see how others' behavior changes our own behavior, compared to a condition where that feedback is absent.
The authors argue, I think fairly, that the absence of feedback provides a more pure measure of the intrinsic quality of each song: the user is downloading based on information about the song (and the band name, I suppose) alone, without knowing how other people feel about it. And what they find, unsurprisingly, is that the presence of others affects choices.
(You may note, by the way, that this is not in itself definitive: the fact that the two groups' preferences are different doesn't mean that they are different because of the experimentally manipulated factor. It would be shocking indeed if two large groups of viewers came up with substantially identical preferences, given random variation alone. But as will be described below, they can tell why they are different; that reason is a surprise I'm saving for later in the diary.)
What I truly love about the study is that they did not stop there. Each person in the "treatment group" (the ones receiving the feedback) was assigned to one of eight separate little feedback worlds. That is, they did not receive feedback from the entire treatment group, but from only a small portion of it. This let them, in essence, run eight separate evolutionary experiments to see how people's preferences grew over time.
What they found is that while the "intrinsic worth" of the song (as measured by the people receiving no feedback) did matter to some degree, it was a much smaller effect than social influence. To quote the author:
This setup let us test the possibility of prediction in two very direct ways. First, if people know what they like regardless of what they think other people like, the most successful songs should draw about the same amount of the total market share in both the independent and social-influence conditions — that is, hits shouldn’t be any bigger just because the people downloading them know what other people downloaded. And second, the very same songs — the "best" ones — should become hits in all social-influence worlds.
What we found, however, was exactly the opposite. In all the social-influence worlds, the most popular songs were much more popular (and the least popular songs were less popular) than in the independent condition. At the same time, however, the particular songs that became hits were different in different worlds, just as cumulative-advantage theory would predict. Introducing social influence into human decision making, in other words, didn’t just make the hits bigger; it also made them more unpredictable.
To some extent, this is just a recapitulation of "network effect" theory, which explains why you need a critical mass before your game platform or interactive technology or blog or diet product takes off, tied to the Stephen Jay Gould notion of evolution (notably from his book "Wonderful Life") as contingent. But I also see this theory as highly relevant to issues -- including many meta issues -- taking place on this blog and in the blogiverse generally. Let's go back to the article for one last paragraph:
In our artificial market, therefore, social influence played as large a role in determining the market share of successful songs as differences in quality. It’s a simple result to state, but it has a surprisingly deep consequence. Because the long-run success of a song depends so sensitively on the decisions of a few early-arriving individuals, whose choices are subsequently amplified and eventually locked in by the cumulative-advantage process, and because the particular individuals who play this important role are chosen randomly and may make different decisions from one moment to the next, the resulting unpredictably is inherent to the nature of the market. It cannot be eliminated either by accumulating more information — about people or songs — or by developing fancier prediction algorithms, any more than you can repeatedly roll sixes no matter how carefully you try to throw the die. [Emphasis mine]
"A few early-arriving individuals ... chosen randomly ...." Great article, Dr. Watts, but there's a stumble there at the end! There's no reason to believe that in the real world the influential early-arrivers are "chosen randomly," or that they even necessary exist at all. That's what the payola scandal is about, after all: music companies giving viewers the impression (or, more precisely, inducing them to make the inference) that other earlier-arriving people must like a given song, or else it wouldn't be played so much. Does this have any relevance to political blogging? Well, consider the discussion topics in which you'd want to keep this in mind:
- what blogs are popular?
- what diaries are recommended?
- what comments get mojo?
- what candidates create buzz?
- what news stories gain traction?
- what pundits retain credibility?
There are some specific applications to DKos governance issues -- this explains why I don't like candidate-flogged diaries and would like to see them segregated out onto a separate Rec List, because the coordinated efforts of campaigns to get diaries Rec'd up uses a process similar to, while not corrupt like, payola -- but the greater import is to use it in politics.
What is DKos and other blogs like it? We are the early arrivers. We identify blogs, diaries, comments, candidates, news stories, and pundits that are worth other people seeing, and we give them some buzz. We are doing this honestly and based on our own preferences -- hell, we don't have enough money to pay people to corrupt the system. And, to be fair, some of the people on the right act out of sincere (if misguided) convinction as well. But what you have on the right is much more in the "payola" vein -- the early arrivers supporting a candidate, issue, or story may not even be real, they may be ginned up, fabricated, alluded to, etc. without actual proof. Talk about "what the American people will or won't support" is intending to invoke just such early arriving ghosts. But we're catching it, and we're calling them on it.
We are the real early arrivers -- combing the web, stories, document dumps, etc. -- who are countering the mostly fake early arrivers arrayed against us. We're there way ahead of the public, but by expessing our interest and spreading the word, we're showing people what information is worth their "downloading."
Some people wonder what we're doing at DKos. This is it:
the decisions of a few early-arriving individuals, whose choices are subsequently amplified and eventually locked in by the cumulative-advantage process
So the time you're spending here is a way of having a small effect at the outset the implications of which will redound loudly down the line. What you read, what you recommend, what you write -- this is one big reason why it does matter, why we're not just wasting our efforts here. So enjoy your time here and use it well. It matters.