Monday, February 23, 2015

Convincing correlations

Given the complexity of human subjects in economics, sociology, and medicine it is hard to find data that clearly indicates a correlation between two variables. I read The Economist each week and many of the graphs that they present just look like random noise to me. However, a year ago they showed the curves below, that I found rather convincing.


The graphs are in an article Tobacco and health: Where there's smoke that marked the fiftieth anniversary of the USA Surgeon Generals report which contained the graph on the left showing a correlation between life expectancy and smoking. The graph on the right shows how cigarette consumption in the USA has changed over time, reaching a maximum around the time of the report and decreasing after the banning of broadcast advertising.

Unfortunately, the response of tobacco companies to their reduces markets in affluent countries has been to shamelessly promote smoking in the Majority World and to intimidate those who oppose them. This humorous/disturbing video from John Oliver describes their antics. I was very proud to see that the previous Australian government led the way in standing up to the companies.

No comments:

Post a Comment

Emergence and protein folding

Proteins are a distinct state of matter. Globular proteins are tightly packed with a density comparable to a crystal but without the spatia...