TEDTalks: "Uncertainty" Peter Donnelly (2005)

Peter Donnelly: How juries are fooled by statistics

Discusses probability theory in relation to genetics (ie. the patterning of G/A/T/C).
Human Genome Project - attempting to understand how differences make people susceptible to diseases.

Currently studying thousands of individuals with different diseases to look for genetic correlations.

Discusses the idea that if a person is tested as 'positive' with a disease then not only the accuracy of the test is in question, but also the frequency/proportion of the population with the disease. As a result, false positives can be frequent.

"We are not good at reasoning with uncertainty."
Errors are very frequently made in terms of grappling with statistics and arguments of (false) logic. In the early days of DNA profiling evidence was misrepresented, (ie. "the chance that this guy is innocent is one in 73,000,000" - was frequently incorrect).

http://www.ted.com/index.php/talks/view/id/67