A DINOSAUR THAT IS 65,000,038 YEARS OLD.

A DINOSAUR THAT IS 65,000,038 YEARS OLD. John Allen Paulos had a review in the Washington Post in 2010 of Charles Seife’s PROOFINESS (compare Colbert’s “truthiness”). He recommends it as a good book on the problems of interpreting and explaining statistical results. Paulos retells a useful story about the failure to recognize that statistical estimates involve uncertainty. When asked how old a dinosaur was, a museum guard said that it was 65,000,038 years old. He had been told it was 65,000,000 years old when he was hired, and he had been working there 38 years. Paulos says: “Seife begins by pointing out that numbers in the news…result from very fallible measurements that are often based on vague definitions or faulty assumptions.” A number is only as good as the estimations and assumptions that produced it. We need to bear in mind the uncertainties associated with the numbers, and it is easier to do that if journalists report those uncertainties when they report the numbers (perhaps even a standard disclaimer that only a certain percentage of such studies can be replicated).