Richard Baker ([info]sharp_blue) wrote,
@ 2006-12-15 11:44:00
Previous Entry  Add to memories!  Tell a Friend!  Next Entry
Population filters; or The Wine Bar Drake Equation

A post by Razib on the weblog Gene Expression about meeting a pretty sf fan[1] who was working in a wine bar seems to have created quite a storm in the science blogosphere in the last day or so. Shelley Batts, Suzanne Franks, Jennifer Ouellette, Tara Smith and PZ Myers all have interesting comments threads that are worth reading. However, I think that almost everyone is missing a key point.

We can consider the predicates is intelligent, is physically attractive and likes sf as filters on populations. Given a group of people we can pass them through the is intelligent filter and select out the intelligent subset, and similarly for the is attractive and likes sf filters. Each of the filters is a little bit subjective, because there are no absolutely agreed upon standards for how intelligent or attractive someone is or what counts as sf. For the sake of argument, let's say someone passes the is intelligent filter if they're in the top 10% of the general population ranked in intelligence as discerned by Razib[2], that they pass the is physically attractive filter if they're in the top 10% of the general population ranked in hotness by him, and that 10% of the general population like what he or I would consider science fiction. Furthermore, let's suppose that the three filters act independently[3]. (If you have more stringent filters then any degree of surprise will, of course, be greater.)

Razib's surprise at meeting an attractive, intelligent[4] sf fan working in a wine is the surprise of a person who's met someone who passes through three quite stringent filters. If only one in ten people get through each then a mere one in a thousand will get through all three. If we assume that people working in wine bars are an accurate reflection of the general population with respect to these three filters then you'd have to visit an awful lot of bars to find an attractive, intelligent sf fan working in one. Hence his post.

Most of the commenters seem to be implying that it shouldn't be a surprise. Many of the examples given are along the lines of "I work in a science department and there are plenty of hot, smart people here!" But, of course, that's working from a biased population. People found in a science department - excluding many support staff - have already passed the is intelligent filter. (If they work in your science department they've also most likely passed the is interesting filter that I haven't mentioned until now.) This means that around one in ten people in the department will pass the joint is intelligent and is attractive filters, rather than one in a hundred in the general population. Hanging around science departments is a much better idea than hanging around bars if you want to meet smart, attractive people!

Furthermore, I think that many people are confusing making observations with perpetuating stereotypes. This is clearly very easy to do, and it's also very easy for people to read comments like Razib's as attempts to push stereotypes even if the comments are not intended to do so. I think I'm willing to give Razib the benefit of the doubt on this front, even though his choice of words would make it very easy to take what he's said the other way. But I'm not going to wade into that debate as last time I made some reasonable, rational comments on the subject I was accused of being a brain-damaged purveyor of "pseudo-intellectual bullshit"[5].

The argument I've been making above can be seen in less charged ways too. For example, suppose I posted that I was quite surprised to meet someone on the bus-stop who was well-read in both history and science. I think that would be quite surprising as perhaps 1% of the general population is what I'd consider well-read in history and similarly for science. I'd have to talk to ten thousand people at the bus-stop - assuming people riding the bus from my suburb are representative of the general population - before meeting such a person. (Actually, probably quite a lot fewer than that as in this case I don't think the two filters are anywhere near independent.) Given that I've only spoken to perhaps a dozen people on the bus-stop, it would be a surprise indeed! I don't think anyone would find me making such a post objectionable[6] because it stays far from stereotypes of groups of people.

[1] Not quite his choice of words...

[2] Of course, I'm really the ultimate arbiter of such things but I'm feeling magnanimous today.

[3] Which they probably don't, but I think any correlations are likely to be fairly small.

[4] Okay, so he did specify that she was intelligent, but she clearly likes not just sf but good sf and I can read between the lines as well as any other blogger.

[5] I think my critic might have a point with that last part ;)

[6] Dull, perhaps, but not objectionable.


(Post a new comment)


[info]i_ate_my_crusts
2006-12-16 08:02 am UTC (link)
You know, I read and liked this, as an analysis, but I think it overlooks the fact that many people seem to have the belief that intelligence and good looks (in women) are negatively correlated.

Example 1: http://www.myfoxboston.com/myfox/pages/Home/Detail?contentId=1758678&version=1&locale=EN-US&layoutCode=VSTY&pageId=1.1.1 (video) wherein the commentator says, about a fictional co-ed at Harvard "she’s smart...so how good can she look?"

This is said, out loud, on a major TV station and isn't censored. I think this demonstrates this attitude isn't a isolated example. And it's not just women -- "nerdy" and "geeky" guys are traditionally considered unattractive.

Similarly, I think there's a tendency to think that liking science fiction is also negatively correlated with attractiveness.

(On the other hand, psychological studies of sterotypes suggest that shown pictures of attractive people and unattractive people will rate the attractive people as more intelligent.)

I think it's these sterotypes -- of the "nerd" that factor into the reaction generated by the original post.

I think your analysis is on the mark, bu if you were to analyse people's *perceptions* of those ratios, as opposed to analysing the actual ratios, I suspect you'd get a different picture.

ie, what people think isn't necessarily true, but if

1. people *think* that intelligence is positively correlated/not correlated with science fiction fans
2. people *think* that intelligence is negatively correlated with science fiction fans
3. people *think* that intelligence is negatively correlated with attractiveness

Then they will consider that an intelligent science fiction fan is pretty standard, that an attractive intelligent person is rare, and that an attractive science fiction fan is rare, and that the combination very rare.

Add in an additional assumption -- that science fiction fans tend to be male. And a stereotype -- that only smart chicks are ugly. Voila!

That the ratios are *actually* as stated in your post then means that the person holding those incorrect assumptions will be surprised more often (and perhaps qualitatively more) than someone who doesn't hold those assumptions.

(Reply to this)(Thread)

[info]sharp_blue
2006-12-16 12:32 pm UTC (link)
Those are good points? How do you think those stereotypes are generated and perpetuated?

I think that people in general are very bad at intuitively reasoning about probability. If I remember correctly, it was Carl Sagan in The Demon-Haunted World who argued that our innate, unconscious apparatus for reasoning about probability evolved in a world of small populations with fairly balanced probability distributions and uses some kludges that don't apply to a world of large populations and wildly skewed distributions. His point was about people assigning supernatural causes to things that seem freakily unlikely but are actually fairly likely given large enough sample sizes, but I think something similar is happening here. People fallaciously think that something like P(hot, sf fan) being small when considering the general population implies that P(hot|sf-fan) or P(sf-fan|hot) are similarly small when in fact they're substantially larger. From there it's only a small step to thinking that no sf fans are hot or that no hot people are sf fans.

Mixed in with this, I think there's an intuitive notion that some people have that life is in some sense fair - I see this contributing to religiosity too, but that's another story - and that people with some set of positive qualities must have a compensating set of negative qualities. Someone who is smart can't then be hot. Someone who is smart and hot must be desperately unhappy. And so on. But the world just isn't fair in that way.

(Reply to this)(Parent)

[info]truecatachresis
2006-12-17 10:27 pm UTC (link)
The initial post clarifies the filtering as:

the key issue for me was the intersection of science fiction && female physical hotitude

so, frankly, we can drop it down to two filters. Which makes it, according to your numbers, a 1% probability. But, more importantly, the issue was really the *surprise* that Razib attributed to this To quote:

Is this the Twlight Zone??? Am I a freak to think this is freaky?

What this really boils down to, then, is that Razib seems to believe that the intersection between the sets of "science fiction" and "female physical hotitude" is far smaller than the 10% your estimates imply, i.e. that he believes the filters do not act independently. The reactions to this point are that, in fact, the reality does not support his reaction, i.e. that he is, indeed, a freak to think this is freaky.

However, there is another statistical point here; even if the chance is 1 in 1000 of meeting such a person working in a wine bar, given the number of customers that go to wine bars and the frequency of turnover of staff in such establishments, it is not in the slightest surprising that a random person would happen upon such a situation.

(Reply to this)(Thread)

[info]sharp_blue
2006-12-17 11:00 pm UTC (link)
My numbers are actually slightly optimistic. Someone elsewhere has quoted data that states that 6% of readers read sf. I don't know how many people read at all, but I can't really imagine that the number of people in the general population who read sf can be much more than about 2%. There's also another factor: what fraction of meetings between sf fans and random people develop in a way that indicates to the random person that the sf fan likes sf? I'd guess that for meetings of an hour or two in a wine bar, this is probably quite a lot less than 10%.

I think I'm going to stop this line of argument now as it's turning in the Wine Bar Drake Equation.

Your last point isn't quite relevant, however. What we're interested in isn't the probability that such a situation will arise for random pairs of people but the probability that the situation will arise for a specific person. After all, we're discussing Razib's surprise and whether he's right to be surprised. Your argument is more akin to trying to calculate my surprise that somebody whose weblog I read (or skim) - of which there are two hundred or so candidates - has such an experience. Clearly I'm much less surprised by Razib's experience than he is for just this reason.

And moving away from the WBDE, I think the real problem is the spin that Razib has put on his experience. If it had been me and I'd been surprised enough to post something - as if I'd post on a topic that wasn't the grand unification of forces or something else equally pretentious! - then my spin would've been that I was pleasantly surprised to overhear someone working in a stylish wine war enthusiatically recommending sf to a fellow employee, and that this was yet more evidence that sf was no longer the ghetto it was when I was growing up, and how this new widespread acceptance of science fiction was surely a good thing for all concerned. (Doubtless by now I'd be fighting off the massed hordes of fandom who all felt slighted...)

(Reply to this)(Parent)

Create an Account
Forgot your login?
Login w/ OpenID
English • Español • Deutsch • Русский…