So, here are the gory details of the t-test. No, not that kind of t-test, silly. This t-test is a statistical test. Very mathy and geeky and such. Two characteristics, by the way, that cause me to be hopelessly categorized as an "Androgyne" by that foolish t-test known as the COGIATI. Well, that and the fact that I can parallel park, am good with directions, and don't get migraines. But I digress.
A t-test compares an observed set of data with a set of data that fits your hypothesis. Based on how statistically different the hypothetical is with the observed, it gives you the probability that your hypothesis is valid. The output of the t-test is called a "p-value." The p-value tells you how likely it is that taking a random sample of the population and evaluating your variable for that sample would produce a result which obeys your hypothesis.
For my test, I assigned values of 0 for straight, 1 for bisexual, and 2 for lesbian. For the observed set of data, I used percentages taken from a study which found 5% identified as lesbian, 2% bi, 93% straight. For the hypothetical set I used 34% straight, 33% bisexual, 33% lesbian. To perform a t-test, you need the mean of the two data sets, the sample size, and the standard deviation of the observed set. I chose a sample size of 200, since bigger numbers would have taken too long to figure. The data looks like this:
Observed: {2,2,2,2,2,1,1,0,0,0,0,0,0,0,0,0,................,0}
Hypothetical: {0,1,2,0,1,2,0,1,2,......................,0,1,2}
Group | Str | Bi | Les | Mean | Standard Dev. |
Observed | 186 | 4 | 10 | 0.12 | 0.4545 |
Hypothetical | 67 | 66 | 66 | 1.01 | don't need |
When you plug these numbers into a T-Test calculator, the p-value comes out to be 1.023 * 10-69. That means there is a one in 102,300,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000 chance that the sexual preference of cisgender women is random, or unrelated.
The probability of winning the Powerball jackpot n consecutive times is 195,249,054n. 195,249,0548 is about 2.11*1066. There is a better chance of winning Powerball eight consecutive times than this of hypothesis being valid. "But wait, Teagan," you say. "This test only compares this single specific set of data. What if the percentage of straight women is really 60%? That screws up your stupid test. This test doesn't prove anything. What then?"
I considered that. So I ran and re-ran the test until I got a probability of 10% that the hypothesis was valid. I had to reduce the numbers to 4 bisexual, 14 lesbian, 182 straight before this happened. That's the biggest variance in sexual orientation which might match the observed results. And don't forget, this was only a tiny sample of 200 people. The larger the sample gets, the less room for difference between the sets is allowed.
Did you get to the bottom of this? I hate to be the bearer of bad news, but if you have, the COGIATI says that you're no more than an Androgyne. :)
Anyway, this was kind of fun to do. I'm a nerd.
