Wednesday, February 08, 2006

Iron Monkey's Rule of the Broken Bell Curve

Normally you aren't dealing with "the average person" or a "cross-section of people," you are dealing with the type of person who would get into the type of situation you're both in.

The percentage of soccer fans at a soccer game is much higher than the percentage of soccer fans in the general population. This example is easy to understand, but what is harder to remember is that almost every real-life situation works the same way, and involves a nonrandom and to some degree self-selected group. We do not live inside carefully controlled studies.

7 comments:

Anonymous said...

Speak for yourself... some of us work in cubicles.

Anonymous said...

Your point illustrates why people (in general) don't understand statistics: they're thinking too small. Because we're used to dealing with the type of person who would be in that kind of situation, we are blinded to the enormous variety of everyone else.

Conversely, I think there's a "squeaky wheel gets the oil" phenomenon, too: sometimes, because a small group makes noises about something, they stand out, even though they are a minority, thus distorting the overall picture.

So--when are you going to publish the Mahomet cartoons? Buy Danish!

P.S.: Go Metric!

Caryn said...

Yeah! I was never able to put it into words like that, but yes.

P.S. The last 3 bananas I ate, I opened backwards. And I thought of you. And I don't even know who you are!

LeesMyth said...

Interesting point. But it seems less true on NYC subways, where the commonality of the group is so broad (a desire to get from Point A to Point B in a cost-effective way, without traffic snarls) that you really get quite a cross-section of the population.

If you travel on the subway very late at night, however, the crowd becomes much less diverse, as it will (to some extent) be people who value money over safety. In other words, late at night the subway-riding population tends to shift to poor people, students, and drunk or crazy people. That somehow reminds me of the apocryphal Texan who said: "Honey, it don't take all kinds, we just got all kinds."

Unknown said...

Caryn - I'm glad to hear that you tried it! What did you think? Was it better?

Squidley - I agree. And I'm planning an original cartoon of sorts for a future post.

Leesmyth - You're right, there are some situations like public transportation that have a wider variety of people than many other situations. However, in your example it still is made up of the subset of people who choose to live in New York City.

Random Michelle K said...

However, in your example it still is made up of the subset of people who choose to live in New York City.

And that's what people miss. Unless it's the census, where they actually go out and hunt down individuals to attempt to get everyone (and even there they fail) You are *always* looking at a subset of the population.

I feel like my Survey Methods class is haunting me. No matter where I turn, I'm reminded that I should be studying. :)

LeesMyth said...

I do agree with your Rule of the Broken Bell Curve - it is a good insight.

And I was also going to agree with you that NYC subway riders are a subset of the people who choose to live in NYC, but now that I put finger to keyboard, it occurs to me that tourists, commuters who live in the suburbs, and out-of-town guests frequently travel by subway as well.

As for those of us who do live in NYC, there are all sorts of reasons why we do it - reasons that divide us more than they unite us.

In fact, I'd almost think that there's more commonality about people who choose to live in Portland than those who choose to live in Manhattan, the Bronx, Queens, Staten Island or Brooklyn (five boroughs with multiple neighborhoods and cultural identities, all served and connected by one giant subway system).

Sorry to ramble on like this. It's clearly past my bed time.