In today’s installment of TED’s wrongful conviction series, Peter Donnelly discusses how misunderstanding statistics and probability can lead to wrongful convictions. Even the “experts” can really mess this up. The problem is often experts in other disciplines attempting to apply statistics and probability, and failing miserably. However, often no one seems to notice. We would never trust a doctor to build a bridge, but many will trust the same doctor if she spouts statistics, even though she probably knows as much about that as she does engineering. The same holds true for psychologists and others. As a statistician, I find so much truth resonating in this talk. My experience is that no one realizes the danger inherent in doing statistics poorly. Here is a look at it. (This video is from 2005, so the video quality isn’t quite as good as the others in this series.)