CHANCE News 1.02
              (8 Sept to 15 Sept 1992)



>>>>>==========>> "Quote of the week in a note from Chris Thron" I've been congratulated repeatedly for my participation in CHANCE. People here are impressed by the other institutions that are involved. The college president, quoting one of our more elderly (and colorful) faculty members, said "That's slopping at a pretty high trough". I was embarrassed to discover that I had not looked at the last two issues of CHANCE magazine. My subscription stopped for some reason. The Fall issue Vol. 4, No. 4 has a farewell from the editors William Eddy and Stephen Fienberg who say among other things " Perhaps the best compliment we have received has been the use of a series of articles drawn from Chance as one of the primary materials for an introductory course in statistics, one that has been taught with considerable success at a number of major American universities during the past year." <<<========<<

>>>>>==========>> This issue of Chance has a number of interesting articles: An article that asks the question "Which record is more likely to be broken DiMaggio's streak or Williams 406 batting average? An article titled "Flawed Reasoning in Court" The authors argue that the fact that in a Vermont case the prosecution could not produce a DNA database for people of the same ethnic background as the subset was not a valid reason to rule that the DNA evidence put forward by the FBI was inadmissible. (So ruled in a pretrial hearing) Stephen Fienberg's last word on Adjusting the Census. "Resampling with More Care" an article that attempts to tone down some of Simon and Bruce's claims for the magic of sampling. Simon and Bruce get to reply. The Vol 5 no 1-2 issue seems less interesting and is obviously just a filler until the next issue that will be the first joint issue with the new editor and as a joint publication with ASA. <<<========<<

>>>>>==========>> Browsing through old issues of the Economist I was reminded also of two interesting articles: The Economist July 4, 1992 A critique of pure reason This article discusses theories of Cosmides and Tooby related to the why people give different answers to the same problem phrased in two different ways. An example is the difference in the number of people who correctly answer the question "If a test to detect a disease, whose prevalence is 1/1000, has a false-positive rate of 5%, what is the chance that a person found to have a positive result actually has the disease, assuming you know nothing more about the person's symptoms?" In a class of Stanford students three times as many people gave the correct answer to this question when it was phrased in terms of "How many of the 1,000 people who tested positive actually had the disease?" The authors claim that this suggests that people think better as frequentists than as Bayesians. (Its not clear what correct answer means in the absence of information about false negatives unless I am missing something.) The Economist June 13, 1992 Heads I win, tails you lose. An account of Michael Rossides's proposal to do away with small change by the following gamble. Suppose the purchase price is $1.52. The customer and the machine pick a number between 1 and 100. The two numbers are added. If the result mod 100 is less that 52 the customer pays $2 and if it is more the customer pays $1. <<<========<<

>>>>>==========>> The Cost of Quality Some examples where Total Quality Management seem not to have worked. Newsweek 7 September 1992 Cancer prevention trials open new era in medicine. Los Angeles times, 7 September 1992 Shari Roan An account of the Breast Cancer Prevential Trial. This is one ofthe first big studies to see if a drug given to people who do not have a disease can help prevent the disease. The risks of the drug (Tomoxifen) are discussed. Peter Duesberg makes the news again. <<<========<<

>>>>>==========>> Op-page editorial by Jerome Groopman, a professor at Harvard Medical School, who views Duesberg's claim that HIV is not the cause of AIDS as a dangerous delusion. A Dangerous Delusion About AIDS. The New York Times, 10 Sept 1992 Mystery of AIDS still deepening; Unproven theories about HIV abound. The Gazette (Montreal) 11 September 1992 by Nicholas Regush An account of the various skeptics to the idea that we know that HIV causes AIDS. <<<========<<

>>>>>==========>> Polls Don't Lie. People Do. The New York Times 10 Sept 1992 Seymore M. Lipset, co-editor of Public Opinion Research A suggestion that the 14 point difference in polls taken at the same time could be caused by the fact that the samples are of different populations: those who say they are registered voters and those who are "likely voters." More people say they are voters than actually are. The author argues that using the "likely voters" gives a bias in favor of Clinton that will be corrected as the pollsters screen better in later polls. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! CHANCE News 1.02 (8 Sept to 15 Sept 1992) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Please send suggestions to: jlsnell@dartmouth.edu >>>==========>>|<<==========<<<