Rules Of The Game #18: The Social Butterfly Effect

[koganbot]

Latest column, in which fame is shown to create greater fame:

The Rules Of The Game #18: The Social Butterfly Effect

The point I make at the end in regard to the Dolls and the Stooges is clear to me but I'm not sure it's clear on the page - that the Dolls' and Stooges' subsequent canonization is somewhat self-perpetuating in just the way that popularity is self-perpetuating (not that the Dolls and the Stooges don't deserve it).

Also, I call "inherent appeal" and "underlying appeal" imaginary; this was a shortcut for saying that the appeal is relative and contingent, which actually is a very different and much better point that I didn't have space to go into. That something is social doesn't mean that it's imaginary, and setting "inherent" in opposition to the social simply removes a perfectly good word - "inherent" - from the language. So I wish I hadn't taken the shortcut or used the word "imaginary," given that I tend to chide other people who say such things. (Note to self: regain self-esteem by finding someone to chide for this.)

A question I'd have asked if there'd been more space: Do you suppose the results would have been as extreme if the tested population had been adults rather than teenagers? If it had been children under ten rather than teenagers? I'm sure Watts and co. would have found the same general syndrome, but my guess is that it's strongest among teenagers, especially males, who are self-conscious and self-doubting and self-dramatizing in their responses to anything, even when ensconced in the privacy of their own minds.

As I recall, when the Watts experiment was first reported in Science in early 2006, poptimists linked to some dipshit (in the Guardian?) who claimed that it explained Ashlee Simpson's reaching number one while Sir Paul McCartney languished lower in the charts.

In any event, what use would you put to Watts et al.'s findings? One thing they underscore for me is that received ideas tend to stay received, but my guess is that this conservativism is mitigated by the fact that ideas don't always reinforce each other (e.g., the idea that Beethoven is unquestionably great is a popular idea, but so is the idea that we should question something's being called unquestionably great). And the findings also tell me that there must be other people of the quality of Shakespeare and Timbaland but who didn't make it, who didn't benefit from the cascading popularity and canonization but who nonetheless produced equally good work (though maybe not in the same quantity, if they lacked the fame to support themselves), so maybe we could go out and find them.

EDIT: Here are links to all but three of my other Rules Of The Game columns (LVW's search results for "Rules of the Game"). Links for the other three (which for some reason didn't get "Rules Of The Game" in their titles), are here: #4, #5, and #8.

UPDATE: I've got all the links here now:

http://koganbot.livejournal.com/179531.html

Comments

Links

[koganbot]

the original report of the study in Science

A good, clear article about it that Watts published this year in the New York Times

Watts' book, Six Degrees (and in the U.K.)

The NY Times and Science may make you register, but it's free.

50,000,000 Elvis Fans Can't Be Wrong

[freakytigger.livejournal.com]

If I've understood them correctly, as a marketer what the findings suggest to me is that the key factor is not whether or not something else has a level of popularity, but how visible that level of popularity is, i.e. a song which appeared on the lists as having 20 downloads but which had recieved none would benefit from the Watts effect as much as one which had actually received 20. So what an idea needs to do is create the impression that other people have thought it's a good idea.

[freakytigger.livejournal.com]

Not having read the full article yet (sorry) - what was Watts actually measuring when you say "did better" - got more downloads, or improved their overall scores? Were there limits on how many songs *could* be downloaded?

[koganbot]

They were measuring downloads.

[freakytigger.livejournal.com]

Another interesting implication is that the effect must take hold at a pretty early level (assuming all universes start with a level playing field) - picking a song based on it having 1 rather than 0 downloads, or 2 rather than 1 - so visible actions have real significance well below the level of 'statistical significance'.

[koganbot]

Yes. This is how I understand what Watts is saying, and this is what guarantees unpredictability. For instance, if - as sometimes happens - I've only got time to listen to some of the songs being listed on a weekly poptimists poll, say of new songs in the British charts, I might only listen to the songs getting the most votes, so I'm more likely to listen to a song getting seven votes rather than one getting three. But seven out of ten flips of a coin, for instance, is not a surprising result, and two songs that poptimists overall would rate about equally could easily have a seven to three split in one of your polls. In any event, either if one song has a slight superiority (in the poptimist mind) or the two are equal and the higher score is due to random fluctuation, I'm still more likely to listen to the higher-scored song and therefore have the potential to vote for it. So the advantage locks in early and builds and builds. Watts calls this the "rich get richer" effect.

[koganbot]

And then of course there's the fact that sometimes - many times - on Kat's yearly polls and canon polls I don't have time to listen to the tracks I've forgotten or never heard, so the only ones I'm voting for are ones I remember well, which are ones that hit (albeit in the U.S.). Again, the rich get richer.

[koganbot]

Also, in an interesting example of the Watts effect, you and I are calling it "the Watts effect" even though he had two co-authors of the Science article - in fact Salganik is listed first in the study, which in the literature therefore would therefore be cited as "Salganik et al." - owing to my emphasizing Watts in the piece, which is because Watts writes well, hence has a book and wrote the New York Times piece. (And Science co-author Peter Sheridan Dodds is the odd man out here.)

[koganbot]

You might want to do a poptimists poll in which people said how they go about voting: before or after looking at other votes, listening to everything they hadn't yet heard before voting for anything, listening to some things but not all things they hadn't heard (this would be most interesting, where the "rich get richer effect" is likely to kick in), not listening to the songs they hadn't heard (ditto), etc.

[katstevens.livejournal.com]

An interesting experiment! I vaguely remember reading something that mentioned it (almost certainly on poptimists).

A possible extension: Social Group X (let's say 'emo kids' for an easy label) sees a (possibly fictional) breakdown of downloads from the other groups, with each group being assigned a particular label? Would the fact that RnB kids (or Swedish/well-off/one-legged/sporty kids, whoever) apparently liked the track deter the emos from listening, even if the track was called I Hate My Life by a band called My Biological Misanthropy? Not sure how practical this experiment would be to carry out.

[koganbot]

I'm sure you're right, that people would be deterred by the "wrong" people liking something and be inticed by the "right" people. In his book, after giving the restaurant example, Watts goes on to say, "Even where we explicitly eschew the majority, we are rarely behaving as atomistic individuals or as pure contrarians. Rather, we typically have in mind some minority group of actors whom we do wish to emulate."

[koganbot]

inticed = enticed

[katstevens.livejournal.com]

I know I would avoid the restaurant with the racous hen party inside.

[katstevens.livejournal.com]

*raucous* arrgh it's been a long day....

[koganbot]

Hen w/ racous is considered a delicacy in Indonesia.

[freakytigger.livejournal.com]

Yeah restaurants is a bad example in many ways because I think for a lot of people the choice would be on a bell curve, i.e. you might look for somewhere which was QUITE busy, not empty or packed. Partly because you're also factoring in how long it might take you to get served, how overwhelming the atmosphere is, even assuming no hen parties.

Whereas you might choose the busy bar over the quite busy one, or the big party over the medium one, depending on what kind of person you were or what you were looking for from a bar or party. (I used to work with people who would ALWAYS choose a rammed pub over one with the chance of a table, because of the 'atmosphere' - but they wouldn't do this in restaurants I'm sure.)

So with music is there an imagined 'crowd' involved - it would be really interesting to see if for some kinds of music there was a limiter effect, a level of downloads at which point something is simply too popular.

Further restaurant metaphor untangling:

[katstevens.livejournal.com]

The hen party will directly affect your enjoyment of the restaurant, but downloading a track is a very individual experience so technically it shouldn't matter who else is listening. Things change when the consumption of music becomes social, i.e. going to gigs or clubs - my enjoyment of the lovely bosh at X Central was utterly spoiled by the horrid crowd who have seemingly latched onto minimal techno :( but my enjoyment of hitherto unfathomable Scooter was massively enhanced by the awesome crowd at that Club FT. This seems like an obvious point, but I wonder if this experiment's methodology (viz isolated downloads) is focusing on the admittedly increasing personal consumption of music (mp3 players) and missing a big chunk of the social environment (gigs/clubs/going over to your mates' houses/back seats of the bus) which is by no means on the wane?

Re: Further restaurant metaphor untangling:

[koganbot]

Well, one point they make in the Science piece is that they believe that they are actually underestimating the effect of social influence:

Our experiment is clearly unlike real cultural markets in a number of respects. For example, we expect that social influence in the real world - where marketing, product placement, critical acclaim, and media attention all play important roles - is far stronger than in our experiment. We also suspect that the effects of social influence were further diminished by the relatively small number of songs, and by our requirements (which aided control) that subjects could participate only once and could not share opinions. Although these differences limit the immediate relevance of our experiment to real-world cultural markets, our findings nevertheless suggest that social influence exerts an important but counterintuitive effect on cultural market formation, generating collective behavior that is reminiscent of (but not identical to) "information cascades" in sequences of individuals making binary choices.

So what they're showing is that even when social influence is limited, it has an enormous effect, leading towards extreme results and limited predictability. Any of the other social effects we factor in will just increase the unpredictability.

One potential lesson from this experiment might be, "It's best to go with the tried and true even though you know it won't always work," and this is where some of the conditions you're bringing up might act as a counterweight. The point being that the untried have a shot, but which of the untried do well is also unpredictable, owing to social influence, but when something does do well it will seem to explode out of nowhere. (He mentions Harry Potter in his NY Times piece; it had been turned down by eight publishers.)

Possible further experiment.

[freakytigger.livejournal.com]

Take a band, preferably one that's not putting out new material and look at their last.fm stats for most listened-to- tracks. Then go back in 6 months and see if the gap is widening between the leading tracks and the other ones.

The Bobby Kogan Effect

[koganbot]

My brother, Richard, writes:

1) I consciously use the restaurant technique on movies. There are some that I know I want to see for one reason or other (I loved the trailer, the Times recommended it and the Post dissed it, it's playing at the Shirlingotn but not playing at Cineplex 16, I liked previous work by the same team - all good reasons, I assure you). But if I am ambivalent, as I often am, I'll also check the box office and go with the crowd, especially if it holds on for a few weeks. I am taking advantage of the aggregate positive correlations that were found.

2) Of course time and chance play a huge role (as Ecclesiastes pointed out a few thousand years ago). Of course there is unmerited fame and unmerited obscurity. (Bach was not exactly obscured after his death, but the work of his sons was considered far superior and even at his death he was viewed primarily as a great organist and clavier performer, not as a composer. Only 100 years later did Mendelssohn proselytize on his behalf, so that his great fame as a composer reached its height in the 19th and 20th century, not during his own times.)

3) "Things happen for a reason." Well, if quantum theory teaches us
anything of philosophical import, it is that the reason things happen is
the law of randomness. Or to put it differently, "Yes: good reasons, bad reasons, and random reasons. The reason I go hit in the head by a falling bucket is that by chance I walked under it when it fell." Of course, the actual "things-happen-for-a-reason" crowd can be divided between those who are successful, and those who are so oppressed by bad luck that they need to adopt a philosophy that is intended to give them the courage to go on (usuually by promising them later rewards) so that they don't just commit suicide right now.

4) In any group, a person who goes first or who is more insistent or more persuasive can have more influence in the group dynamics than he deserves. Often this is for ill, but occasionally it is for good. I observe that any group Bobby* is a part of ends up treating all its members more decently and humanely than most groups of teenagers customarily do - because Bobby is a living embodiment of a fierce commitment to listen carefully to others and treat their ideas and preferences with respect, no matter how much he disagrees.

*Richard's son, my nephew. Tom Ewing is the Bobby Kogan of our corner of the Web.

Poptimists discussion

[koganbot]

poptimists discussion on related issues here.

Re: Poptimists discussion

[blue-russian.livejournal.com]

Good column, Frank.

Re: Poptimists discussion

[koganbot]

Thank you!