CHANCE News 1.08
              (23 Oct to 29 Oct 1992)


Please send suggestions for  articles to include to

All of the newspaper articles and some of the journal articles are 
available in full text on NEXIS/LEXIS which your library probably 



>>>>>==========>> The sure thing that got away. The New York Times Oct 25 1992 David W. Moore Moore is director of the Survey Center at the University of New Hampshire and author of the recent book "The Superpollsters: How they Measure and Manipulate Public Opinion." Here and in his book he remarks that the way the questions are asked leads to a significant underestimate for the number of undecided votes and can lead to significant changes in the polls as people do make up their minds. He points out that, in particular, this can lead to significant fluctuations in the polls towards the end as people who have not made up their mind can be influenced by unpredictable events in the final campaign such as a particularly effective negative ad, etc. <<<========<<

>>>>>==========>> Polls apart: Why the surveys differ. Boston Globe 30 Oct 1992 Adam Pertman A critique of the Gallup CNN poll that estimated a close race by throwing out voters least likely to vote (actually most polls do this but apparently Gallup did more than most on this poll). A number of explanations are given for the variation in polls including the wise statement "We have to accept some randomness in this business" EDSTAT-L this week Marshal Mangan asked how margin of error is computed especially when you have three candidates. This led to an interested discussion of the possible ways that it might be done and the observation that probably not many people really know what margin of error means. I was told that Harris once asked the pollsters at a professional meeting what they thought it meant and got an interesting variety of answers. <<<========<<

>>>>>==========>> Earthquake prediction now shaky. Washington Post 23 Oct 1992 John E. Yang More about the attempts to predict a level 6 earthquake in Parkfield California by a rather formidable amount of equipment set up there for this purpose. Regarding their failure last week they point out that "Nobody said there was 100 percent certainly". Apparently, they can still win their 95 percent prediction that there would be such a quake sometime between 1984 and January 1993. <<<========<<

>>>>>==========>> Two studies conflict on when to test for breast cancer. Toronto Star, 24 Oct 1992 Marilyn Dunlop A recent Boston study concluded that women should be screened each year starting at age 40. "Dr. Anthony Miller, who led a major Canadian breast screening study scheduled to be published next month, says that the Boston team's study does not show the benefits claimed by the researchers. The Canadian study is expected to show that mammograms do not cut the death toll among women younger than 50." <<<========<<

>>>>>==========>> Federal Health Official Propose an Expanded Definition of Aids. The New York Times, 28 Oct 1992 Lawrence K. Altmon "The Center for Disease Control is proposing a more comprehensive definition of AIDS that include a test of immune function and add three illnesses to 23 other complicating ailments listed in the current definition, which was devised in l987." "The three new conditions are invasive cancer of the cervis, pulmonary tuberculosis and two or more episodes of bacterial pneumonia. The new definition will include any adult infected with H.I.V. who has 200 or fewer CD-4 cells per microliter of blood, or about one-fifth the normal level. No such measurement is part of the current definition" It is expected that the number of AIDS cases in 1993 would go from the previously projected 50,000 to about 100,000, 30 to 40 thousand coming from including individuals with CD-4 counts less than 200 and the rest from those with the new complicating illnesses. <<<========<<

>>>>>==========>> There's no 'Whites only' sign, but. Business Week, 26 Oct 1992 Geoffrey Smith; Mike McNamee "The Federal Reserve Bank of Boston has come up with a definitive answer to a question that has dogged the banking industry for decades --- whether banks discriminate against minorities. The answer: Yes. A study released Oct. 8 found that mortgage applications from blacks and Hispanics are 60 percent more likely to be turned down than those from whites from similar financial backgrounds. Says Boston Fed President Richard F. Syron" <<<========<<

>>>>>==========>> Courtroom conundrum. U.S. New and World Report, 26 Oct. 1992 Joannie M. Schrof An account of some of the difficulties that judges and juries have in dealing with cases whose resolution depends upon deciding the present state of scientific knowledge such as DNA fingerprinting and medical trials. Solutions are discussed such as providing judges and jurors with DNA primers and crash courses on statistical methods. <<<========<<

>>>>>==========>> Follow the money. Forbes 26 Oct 1992 Charts and other statisical devises to show that tax increases raising any large quantities of revenue will have to come mostly from the middle class despite the wishes and claims of politicians. <<<========<<

>>>>>==========>> Searching for Markers on the AIDS Trail. Science 16 Oct 1992 Jon Cohen The difficulty in doing clinical trials for drugs to cure a disease like AIDS that takes many years to develop has led to trying to find surrogate markers to allow clinical trials over shorter time periods. For AIDS the best-known surrogate marker has been CD4 counts. This article discusses the difficulties of getting reliable predictions using a particular marker like the CD4 counts and the search for other markers. <<<========<<

>>>>>==========>> Mortality: Overturning Received Wisdom. Science 16 Oct. 1992 Marcia Barinaga An account of two papers in this issue of Science that challenge the theory that the death rate in monotone increasing. Gompertz law claims that in humans the death rate increases exponentially. It has been roughly verified for humans but scientists have pointed out that it is hard to check for very old people because of limited data. These studies carried out with millions of fruit flies under quite controlled conditions suggest that Gompertz law is not appropriate for fruit flies and may well not be correct for humans. <<<========<<

>>>>>==========>> Environmental Exposure to Lead and Children's Intelligence at the Age of Seven Years -- The Port Pirie Cohort Study. New England Journal of Medicine 29 Oct 1992 See also editorial on this subject by Katheryn Mahafey in same issue This study shows that exposure to lead in early childhood can cause identified intellectual defects in children even after a seven year period. This assertion has been previously challenged in the press and scientific literature on the grounds that if there are such defects they are probably due to factors common to to children with a greater exposure to lead -- poverty, urban environment etc. This longitudinal study together with two forcomming longitudinal studies, one on an affluent population in Boston and another on disadvantaged children in Cincinnati, are said by Mahafey to "demonstrate that the lead-associated decrements in intelligence are persistent across cultures, racial and ethnic groups, and social and economic classes. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! CHANCE News 1.08 (23 Oct to 29 Oct 1992) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Please send suggestions to: jlsnell@dartmouth.edu