<- file stat smoke.html -> Blailar, on Smoking Here is something from an epidemiologist on some smoking issues. I have previously posted this in some note, to some .stat. NetGroup, when the issue once arose. -- RFU. _Chance_, Vol 8 #1, page 25, 28,20, John C. Blailar III, "How the News Media can help Statisticians do a better job". 'Second, neither science, nor news reports, nor the public can benefit from an unfortunate tendency of the news media to give dead controversies an appearance of still being active.... ..'The Surgeon General's report on smoking and lung cancer, published in 1964, presented an early version of criteria that were (apparently independently) developed by Bradford Hill in his famous paper of 1965, which laid out procedures and criteria for the rational interpretation of observational data in epidemiology. The evidence about the destructive health effects of tobacco was sufficiently clear in 1964, now 30 years ago, to support strong public action, but action was blocked then and continues to be substantially hampered by the tactics of the tobacco industry. Those tactics simply could not succeed if there were not a substantial part of the population that continues to think that the matter is really in some doubt.' <deleted, third point, about relative strengths of arguments.> <deleted, fourth point, about keeping a skeptical attitude toward pronouncements with some ulterior motive.> ... 'Another example of a grand pronouncement that needs to be treated with greater skepticism is the recent conclusion by the Environmental Protection Agency that secondhand smoke (environmental tobacco smoke) is a substantial cause of lung cancer in nonsmokers. That conclusion may well be true, but the meta-analysis used to suport it applies flawed methods to flawed data in such a way as to supply little support to the conclusion. (There is quite solid evidence that secondhand smoke has other kinds of harmful effects, such as chronic lung disease in both adults and children; it is the specific linkage to lung cancer that does not appear to be adequately established.)' * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Document by Rich Ulrich. E-mail to wpilib+@pitt.edu

FAQ top.

Ulrich home page.

Ulrich FAQ. http://www.pitt.edu/~wpilib/stats99.html