No wonder people hate statistics

Ok, so you have some theory about something and you have some data and you want to test your theory with your data? Start by assuming the opposite of what you want to prove. Then, test whether the data you saw were unlikely to have occurred if the oposite if the thing you're trying to prove were true. If so, then either you've observed something unlikely or you were wrong when you assumed the opposite of the thing you wanted to prove. Since unlikely things don't happen very often, you then conclude you were wrong back when you assumed that the thing you wanted to prove was wrong, thus proving your theory. Well, not really proving it. Providing evidence for it in the form of a probabilistic statement about how unlikely the data you saw would be if your theory were wrong.

That's the basic logic of a large area of the field to which I've dedicated the last 15ish years of my working life (joder), and that is what I need to start explaining to 200 loud young Catalans in Spanish next week, and it makes me want to cry. Do you know how many different verb tenses there are in that convoluted logic? I should have started explaining it today, in fact, but I didn't have my act together so I punted and dragged out a review of less confusing shit and ended class a little early.

If someone who isn't me could explain p-values in Spanish to a room full of teenagers and then tried to tell me they didn't really speak Spanish, I would think they were full of shit. Still, I think there's a pretty good chance that I actually can't explain p-values in Spanish to teenagers (hence the chickening out and wanting to cry).

I bet the students who cried during exams the last time I taught a statistics class would love to see me now. And I'm really glad they can't.