Two Ideas We Need to Teach the Media (and Everyone Else)

• Activity Number: 102
• Type: Invited
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Committee on Excellence in Statistical Reporting
• Author(s): Howard Wainer*+
• Companies: National Board of Medical Examiners
• Keywords: Bayes Rule ; Lack of Independence ; Finding Causes ; Convenient estimators ; Common flaws

Abstract

The science of uncertainty suffers from the unfortunate combination of its criticality in the understanding of the modern world and the subtlety of some of its ideas. In this talk I will discuss four statistical errors that have yielded serious practical consequences over the past century and will argue that we must work harder to clarify the public's understanding of them. These are: 1. The choice of event probabilities based on their convenience but not their accuracy; 2. The assumption of independence when evidence screams that this is untrue; 3.The calculation of P(data|theory) rather than the more relevant P(theory|data); and 4. The insistence on a cause for randomly occurring clusters.

I will illustrate their occurrence in two trials, one popular film and in the mistreatment of two Massachusetts graduate students.