Emily Oster, you idiot.

So I’m taking a break, eating my Waitrose pesto pasta salad, feet up on the table, and decide to watch something on TED.com.  Emily Oster’s talk, about “flipping” our thinking on AIDS in Africa catches my eye and, I have to say, it was hard to keep myself from gnawing at my keyboard.

She’s wrong.  Oster sounds like someone who has never been to Africa, yet conducts epic analyses with data presumed to be accurate, then claims to know much more about the continent than Africans themselves.  Don’t believe me?  See for yourself.

A huge problem here is that she’s making lots of assumptions about the data.  AND SHE’S TAKING CORRELATIONS TO MEAN CAUSATION.  Just because one data set looks like another when you plot them on a graph DOES NOT MEAN THEY ARE RELATED.  It only means they’re correlated.  Correlations might suggest predictive ability, but they DO NOT PROVE predictive ability.

This is science at its worst.

This is economics at its worst.

Emily Oster, if you want insight into African crises, move to Africa.  Live there.  Not forever, if you’re not inclined, but at least a couple of years.  Long enough to learn things, like how people don’t die when they get malaria (there’s an immunity developed amongst most locals, which is actually the reason for its prevalence), and how disparate beliefs and contexts play a massive role in HIV infection rates — a role that easily confounds your assumption that Egyptians and other Africans differ only in that Egyptians don’t die early from AIDS.