Sunday, May 11, 2008

Analysis of competing hypotheses

The US Central Intelligence Agency's reputation for analytical accuracy may have been tarnished lately, but nevertheless it can provide useful generic tools for analysis under uncertainty. The chapter on Analysis of Competing Hypotheses from the publication Psychology of Intelligence Analysis is such a case in point: it describes in some detail the procedure how to arrive at the most likely hypothesis by way of analysing competing hypothesis, knowing that - against intuition - the hypothesis with the least evidence against it is often the best guess. The rigour of this approach may be excruciating, but it is also the most promising in my view.

No comments: