Thursday, August 21, 2008

90% of Shark Attacks Happen in Less than 6 Feet of Water

CNN has the story of a website that allows comparisons by death rate in the cases of pneumonia, heart attack, and heart failure. Currently the CNN online poll shows 60% saying they would not be interested in this information, which is consistent with other research that says people are unwilling to compare hospitals.

However, I think that people would be quite right to heavily discount information about hospital death rates because there is a clear difficulty in separating causality and correlation. What does a high death rate mean? For a similar reason, Economists studying housing markets have long noticed that crime rates are not always big negatives for housing prices. Well, does the crime rate reflect a very good police force that people trust to report crimes to, or does it actually reflect higher crime? Does a high death rate mean that people with difficult cases disproportionately seek out that hospital for their treatment, or is it that they are not good at handling basic cases?

No comments: