Did you want t comment on my assertion that you could choose 100 different attributes and reference that with who is "co-operative" and who isn't, and ONE of the two groups you choose is going to be more co-operative or they will be EXACTLY the same (which is the most unlikely outcome of all).
Well, that's right.
But don't you think that's something that applies to basically all research? For instance you probably take it for granted, as the medical establishment does, that things like smoking dramatically increase your risk of cancer. But your criticism applies just as easily to that research, as well - you can take your sample size, re-index them into groups based on nose size or hair color or something stupid like that, and you're certainly going to have one hair color turn out to get cancer at a greater rate than another. In fact
every time you take a group of people and organize them by one trait, they're going to be in groups that necessarily differ along some other quantifiable trait. Two groups of people organized by zip code will differ in weight. Two groups of people organized by eye color will differ in how much they like country music.
You're by no means the first person to figure this out. (Frankly you should never suspect you're the first person to figure
anything out. That goes for the rest of us, too.) As a result, statisticians have a series of tests and tools to determine when two groups really differ in a
statistically significant way. It's based largely on the notion of the standard distribution - a pattern that naturally varying traits fall into - and the expected variation within a completely random group compared to one selected based on a certain trait. If you find the idea of such tools and tests interesting, then you could do worse than audit an Intro to Statistics course at your local college, or barring that, read Larry Gonick's
The Cartoon Guide to Statistics. I own it, it's great.
I largely agree with you that symmetry is not the be-all and end-all of attractiveness. But it's widely-understood to be
the most quantifiable association with attractiveness. Short of parading study participants in some kind of "Hot Or Not" cattle call, I can't think of a better quality on which to objectively measure who is more attractive than who. Can you? The best part about statistical analysis is that it is error-counteracting; some people in the "not attractive" group will probably be more attractive than some people in the "attractive" group since symmetry or lack thereof is the sole characteristic for grouping in these studies. But, the "attractive" group will be on average
more attractive than the "not attractive" group, significantly so, and that's sufficient to draw conclusions about the behavior and characteristics of attractive vs non-attractive individuals in the aggregate.
It's not hand-waving. It's a known, mathematically-verified application of statistical tests. Much more interesting would be for you to talk about the actual paper, it's actual flaws and actual conclusions, rather than hand-wave over the fact that you haven't read the research you're talking about. You know, since we're talking about hand-waving.
™ Version 4.2
