WeightOfEvidence

Human beings are by nature quite incompetent when it comes to statistics. Until recent millennia we haven't really needed to be good at more than crude estimation of probabilities and correlations --- hence the growth of the gambling (aka "gaming") industry. Gut feel is particularly unreliable when a situation touches upon one of our cherished beliefs. Beware, for instance, attempts to explore the connections among guns and violence, or the relative risks of tobacco when smoked versus chewed, or the chances of getting into an accident in a monster SUV versus a smaller but more agile vehicle, or the impact of human activities on climate, or the hazards of various forms of electricity generation ... or any of the infinitely delicate issues surrounding group averages of almost all parameters when viewed along dimensions of race, sex, religion, nationality, age, political alignment, etc. There's a reason that civilization needs actuaries and statisticians.

Evidence is a particularly slippery but important concept. What it really means is conditional probability. How much does knowing one fact change the chances of another fact being true? That, in a nutshell, is evidence. There are solid mathematical methods to analyze and add up evidence. The application of those methods is straightforward but nonobvious. The inputs to those methods are subject to debate in any particular case. But like gravitation, the math works whether we want it to or not.

This all came to mind again recently when an otherwise-highly-intelligent person was ranting against a particular anti-terrorist proposal. (The details aren't important; it was some complicated procedural-technical notion that's unlikely to ever see the daylight of actual implementation.) The argument was doubtless well-intentioned, but it flouted the simple truth that evidence --- even when noisy, inconclusive, and fragmentary --- still provides clues that can improve judgments in crisis situations.

It's ok to refuse to consider a suggestion because it's disgusting, or inhumane, or unæsthetic, or too expensive. But it's stupid (or disingenuous) to close one's eyes to information that can help increase the odds of making good decisions, even when that information is incomplete and unreliable.

(see also WebsOfEvidence (15 Feb 2000), PickyAboutFacts (11 Mar 2003), AbsenceOfEvidence (17 Mar 2003), ...)


TopicScience - TopicThinking - Datetag20040321



(correlates: AbsenceOfEvidence, SherlockHolmes, FourTypes, ...)