Statistical hypothesis testing

I got 14/16. Not bad right?

Sometimes there are concepts in school that stay just that, abstract ideas lacking intuitive real-life examples. Sure, there are illustrations of Type I and Type II errors in the detection of disease in biological samples, but the concept never meant as much as when I got 14/16.

Can you tell?

Sure, 14/16 positive positives are good. I've never complained about getting an 88% on an exam. The test itself is none too brilliant nor scientific, but in this example, it *really* does matter whether the 2/16 was Type I or Type II. Would I rather mistake a womanly looking man for a woman, or a manly looking woman for a man? Stats can be so complex and fraught with... icky-ness. Unfortunately, my 2/16 were Type I.

I'm never going to a bar without X-Ray glasses or a SRY test kit again.

An example worthy of Stats-101 textbooks.

Thanks to Screenhead.