The Search for the Elusive Asshole...

So, my school always likes to stress the practical, real-world applications of what we learn in class. After all, what's public policy good for if it only exists on paper?

One of the things you hear time and time again from the female population is that all men are assholes. Their ex-boyfriends are all jerks (but note that they seldom think that when they first start dating them...perhaps there is a reverse causality?) and clearly there is a lack of justice in this world and male assholes (who supposedly abound in the wild) should be shot, locked up, castrated, or something of the sort. Now, I can't address the justice or corrective measures, but I can certainly take a look at the normative statement that "all men are assholes."

Now, one way of looking at this is to test the null hypothesis that all men are assholes:

Ho: P(Asshole|Male) = 1

That is to say that across the entire human population, if all men are assholes, then the probabilty of being an asshole, given that you are male, is 100%, or 1.





Here's a nice venn diagram to illustrate:





The only problem with this is that it's just too easy to invalidate: just find one good male soul and the statement falls apart. Surely one would not classify the Ghandis, Mandelas, and Dalai Lamas of the world as assholes, but perhaps their numbers are quite few. Given the 3 billion or more males on the planet, a 1 in 10 chance of being a non-asshole is still good enough for there to be a decent non-asshole population for women (and gay men) to fight over. So lets be slightly conservative with our estimate and say something like 90% of men are assholes, which leaves us with 300 million good men in this world. This would look somewhat like this:



P(Asshole|Male) = .90




So, one of the ways we can test this is to take a random sample of men, see if they are assholes, and test if the rate of asshole-ishness among men in the sample is significantly different from the population mean of 90% that we have asserted. To do this, we can use a simple statistical t-test (I'm going to use non-traditional symbols simply because I don't know and am too lazy to figure out how to do the fancy signs that statistics likes to use):

(X-Y)/SE
Where:
X = population mean, which is .90;
Y = sample mean (the observed percentage of male assholes);
SE is the standard error, which is Z/SQRT(N), where Z is the standard deviation of the sample mean and N is the size of the sample.

The diagram above shows a normal distribution curve for the percentage of assholes among males. It is centered on the population mean, 0.90. The blue shaded region is the 95% confidence interval, which means in laymen's terms that if the sample mean that we obtain falls outside the shaded region, we are 95% certain that the sample mean is different from the population mean. In otherwords, if we get a percentage of 80% or something outside that blue area, there is a 1 in 20 chance that our figure was simply a fluke, but a 19 out of 20 chance that it really means something significant: that the percentage of assholes among males is not 90%.

But here we have a dilemma: how do we know what makes a person an asshole?

Before we can even attempt to test what the true percentage of assholes is among men, we must first be able to identify them accurately. Of course, many of you will say that assholes are pretty easy to distinguish, but then again, if they were, you wouldn't have dated them in the first place! So we need to first determine what factors in a person give rise to asshole-ishness. This requires regression analysis (and a future post!)

To investigate, a friend of mine and I are conducting a survey of women that asks them to describe the characteristics of their boyfriends, past and present. Through this, perhaps we can obtain the causal relationships that relate to asshole-ishness among males. And the world will be better off.

If any altruistic woman out there would like to add their data (err... history) to this study, please email me, and I will furnish you with a copy of the survey for you to fill out.

For the good of science!