Schrage writes:
...his columns have quickly become "must reads" for entrepreneurs and executives who grasp that mathematics is now the "lingua franca" of serious business analysis. There is no better English-language explicator of complex quantitative concepts than Steven Strogatz. His work is a model for how mathematics needs to be popularized
Here's Strogatz on doctors, the OJ Simpson case and probability. After reading this, you will realize that statistics can not only lie, but really confuse.
That was an eye-opening article. Reminded me of a presentation I saw on TED by Hans Rosling. Although he draws some absurd conclusions (Mao brought health into China!), his presentation is still pretty exciting, especially his information on income distribution:
ReplyDeletehttp://www.youtube.com/watch?v=RUwS1uAdUcI
I must disagree and urge your readers to review Strogatz's article with an open mind. His suggestion about using frequencies rather than percentages is an important way to clarify the statistics. Just as important is to be careful about focus. Regarding cancer screenings its very helpful to focus on those who have a positive reading and than, within that group how many actually have cancer.
ReplyDeleteThis is completely different from the presentation by Rosling who uses sophisticated computer graphics to present population and income data in a time series. This was clearly used to sell while Strogatz was more about understanding statistical significance. Aside from the questions about realiability of data Rosling used which was compiled by the UN (think climate change or any of the failed development programs in Africa), Rosling's data series suggested that as countries moved toward market economies their income grew and their health improved. His comment about the health improvement in China during Mao's rule was out of context and he didn't dwell on it. The fact was that Mao arrived when China was coming out of a very primative time and almost any change could lead to improved health. More important was the fact that Mao's policies resulted in millions of deaths due to starvation which was probably not part of the data set. If millions die the remaining population improves its access to resources.
This is what Peter Orszag wants to do in health care with his call for econometrics, instead of double blind clinical studies.
ReplyDeleteHe made the comments last year at the Brookings Institution.
Econometrics and quants helped drive Wall Street into the abysmal quality range. Expect the same for health care.