February 12 2004 Cleavage Among the Voters? USA Today's poll on Jackson's breast baring
We usually think that spotting an error in a professionally administered poll takes some extra degree of training, or some knowledge of higher math. But sometimes spotting a major problem in a poll published by usually reputable news organizations is unbelievably easy. Take a poll published in USA Today on Wednesday, which reported that 55 percent of Americans who watched the Super Bowl half-time show were not "personally" offended by the baring of Janet Jackson's breast, and that 45 percent were. It seems like a distinct split - but the poll also had a margin of error of +/- 5 points. People often don't realize that the margin of error applies to all the percentages given in a poll, and that it can work in either direction. So, really, the USA Today poll shows a statistical dead heat: the percentage not offended could be as low as 50 percent, the percent offended could be as high as 50. The poll's results are still meaningful, but only to show how ambivalent America is about seeing Jackson's breasts on TV - not how divided. http://www.stats.org/logentrybrowse.jsp?type=logentry&date=true&orderby=date +desc&limit=11&start=0 In 1994, an epidemiological study on the relationship between induced abortion and breast cancer risk, published in the Journal of the National Cancer Institute, [FN2] made national headlines. [FN3] Dr. Janet Daling and a team of researchers at Seattle's Fred Hutchinson Cancer Research Center reported that "[a]mong women who had been pregnant at least once, the risk of breast cancer in those who had experienced an induced abortion was 50% higher than among other women." [FN4] When women underwent abortions before the age of eighteen or at age thirty or older, the study found more than a twofold (150%, or 110% higher, respectively) increase in risk. [FN5] Since an average American woman's lifetime risk of developing breast cancer is already about twelve percent, [FN6] a twofold increase would imply an "absolute effect" [FN7] from a single *1597 induced abortion that is comparable to the risk of lung cancer from long-term, heavy smoking. [FN8] The Daling study is just one of many published since 1957 showing a statistical link between induced abortion and the occurrence of breast cancer. http://www.johnkindley.com/wisconsinlawreview.htm
