!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CHANCE News 2.06
(15 March to 29 March 1993)
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Prepared by Laurie Snell
Please send suggestions for articles or format to
dart.chance@dartmouth.edu
Back issues of this chance news and full text of the articles referred
to are provided on the Chance gopher. Just point your gopher to
chance.dartmouth.edu
===================================================
The race may not always go to the swift, nor the battle to the strong
but that's the way to bet it.
Damon Runyon.
===================================================
ARTICLES ABSTRACTED
>>>>>==========>> Bob Norman and Steve Comer noted that our discussion of the article by Wainer and Steinberg on sex differences in performance on the Math SAT was slightly garbled. It should have said that women get on average 35 points less than men on the Math SAT rather than 35 points more then men. I have since obtained their article and, while they say that one solution to the problem would be to add points to the women's scores for admission purposes, I do not see that they advocate that. Indeed they say that a better solution would be to figure out what is going on. Of course, I was pleased that two people had read the news-letter to the end. <<<========<<
>>>>>==========>> Tests for AIDS and drugs: how accurate are they? Parade Magazine, 28 March 1993 Marilyn Vos Savant Marilyn devotes her entire column to the problem of false positives in testing. A quite reasonable discussion. <<<========<<
>>>>>==========>> Scientific evidence: court may decide when jury can hear it. LATimes 30 March 1993, A5 David G. Savage The Supreme Court starts today (March 30) the Daubert suit which involves the kind of scientific evidence that should be allowed in the courts. The article suggests that the court may choose between the Frye rule which suggests that scientific evidence has to have been screened through the traditional peer-review process or the more liberal guidelines provided by Congress when it adopted the Federal Rules of Evidence in 1975 to guide federal judges which seemed to favor letting the juries hear and weigh the conflicting evidence. <<<========<<
>>>>>==========>> Studies find AIDS virus 'hides' in lymph system. LATimes, 25 March, 1993, Part A, page 34 Sheryl Stolberg An account of two articles in the current issue of Nature that seem to solve the mystery of why it has not been possible to detect signs that HIV virus was reproducing and spreading during the prolonged "latent" period despite the fact that the T-cell counts decline during this period. Two groups of researchers have found that HIV virus, instead of going into an inactive period, resides in the lymph system and attacks the immune system from within, like a Trojan horse, remaining undetectable. The discovery suggests earlier treatment with drugs but the toxicity of current drugs like AZT make this not a feasible until newer drugs are tested. It also suggests additional difficulties in developing an effective vaccine. <<<========<<
>>>>>==========>> Facing the cruel truth about HIV Times Newspapers Limited. John Maddox (Editor of Nature) A discussion of recent evidence such as the previous article that Maddox feels should make Duesberg give up his argument that HIV virus is not the cause of AIDS. <<<========<<
>>>>>==========>> Report calls cholesterol tests in young unjustified. The New York Times, 17 March 1993, Sec B, Page 8 Gina Kolata An article published in the Journal of American Medical Association by Dr. Stephen Hulley concludes that there is no justification for routine cholesterol tests before age 35 in men and age 45 in women. This contradicts national guidelines for routine cholesterol screening even in young adults. Hulley argues that this screening will lead to the use of drugs which cause more harm than good for people in this age group whose risk of heart problems during this period is negligible. They give arguments that early lowering of cholesterol is not necessary for possible beneficial lowering of cholesterol in later years when risks are significant. The editorial supporting Hulley's recommendation has a colorful discussion of the way that results of studies are reported to influence the need for screening and treatment. This article has lots of quotes from people who disagree with Hulley. For example, Dr. Claude Lenfant, director of the Heart, Lunge, and Blood Institute was "appalled" by the article. <<<========<<
>>>>>==========>> Conversations James E. Muller; Eat foie gras, live longer? A cardiologist sadly begs to differ. The New York Times, 21 March, 1993, Sec 4 Page 7 Elisabeth Rosenthal A conversation with Dr. James Muller well-known cardiologist who did a classic study in 1985 showing that heart attacks tend to occur in the morning. Dr. Muller suggests that people will just throw up their hands if they continually read articles that tell them baldness is bad, foie gras is good, margarine is bad, walnuts are good etc. for prevention of heart attacks. He advises us not to worry about these things saying that we will be told by the American Heart Association and the National Heart, Lung, and Blood Institutes when the research results reach a level where it should cause a change in our life style. He also suggests that some of the studies have been supported by suspect sources. For example, research showing the value of walnuts was supported by the California Walnut Commission, and the baldness study was supported by Upjohn which makes a cream to promote hair growth. <<<========<<
>>>>>==========>> Study suggests aspirin may inhibit some cancers. The New York Times, 21 March, 1993, Sec 1 Page 32 AP A study published in the journal Cancer Research found that death rates from stomach, esophagus, colon and rectal cancers were about 40% lower among people who took aspirin at least 16 times a month than among non aspirin users. This is in agreement with a previous study that looked only at colon cancers. These were epidemiological studies on 635,000 Americans. An expert on cancer prevention found the new findings puzzling saying that it would be unusual for one thing to be a protectant against all four cancers and suggested the need for clinical trials. Such trials for the case of colon cancer are already underway. <<<========<<
>>>>>==========>> What price risk? The Economist, 6 Feb, 1993 Page 81 The beta of a stock is a standard tool for investment managers. It measures a share's relative volatility. If a share price moves in line with the market is has beta value 1; if it rises by 20% when the market rises by 10% it has beta value 2; but if it rises only 5% when the market rises by 10% it has beta of .5. The theory is that the only way you can do better than the market is to choose risky (volatile) stocks and the beta is helpful in doing this. An article, reported last spring, looked at data between 1963 and 1990 and found that differences in beta did not explain the performance of different shares. This article reports on more recent articles that have come to the defense of beta. To show that beta is a predictor of riskiness one study looked at the ten worst months ever for American shareholders and found that the highest beta (most risk) shares performed far worse that the market as a whole and the least risky did significantly less badly. Similarly, in the top ten months the low risk shares rose significantly less than the average and the high beta shares significantly higher. <<<========<<
>>>>>==========>> Fallacies, flaws, and flimflam. The College Mathematics Journal Ed Barbeau Inspired by all the fuss over Marilyn Vos Savant's discussion of the Monty Hall problem, the author gives an exhaustive bibliography (63 articles) that relate to this problem or one of its more or less equivalent versions such as the shell game, the prisoner's paradox, Bertrand's three box problem, the paradox of the second ace, the second sibling paradox, and the restricted choice bridge paradox. <<<========<<
>>>>>==========>> Medical Uses of Statistics, 2nd Edition. Edited by John C. Bailar III and Frederick Mosteller NEJM books, Boston, Mass 1993 At the suggestion of the New England Journal of Medicine, Bailar and Mosteller organized a study of research papers in the NEJ to see what statistical techniques are used in medical research and how they are used. Reports of this study are woven into a discussion of the basic statistical concepts that the authors found most used in this medical research. Emphasis is on the ideas involved in the various statistical methods used and not on the technical aspects of how to actually carry out specific tests etc. Despite the variety of authors, the exposition is uniformly clear and free of jargon making the book readily available to the non statistician. A remarkably discussion of what it all means. It is interesting also to see how few basic ideas are actually used in the medical literature. <<<========<<
>>>>>==========>> !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! CHANCE News 2.06 (15 March to 29 March 1993) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Please send suggestions to: jlsnell@dartmouth.edu