[escepticos] RV: Dubious stats of 2000

["planetario de pamplona" <planetario@...>]

From: Bill Steele <ws21 en cornell.edu>
Sent: Friday, December 29, 2000 7:05 PM
Subject: Dubious stats of 2000


> Thought this might provide inspiration. A press release from 
> Statistical Assessment Services:
> 
> 
> THE DUBIOUS DATA 2000 AWARDS
> 
> The Top Ten Silliest, Most Misleading Stories of the New Millennium 
> (at least, we think it's ten, and we're pretty sure that the 
> Millennium has already started)
> 
> Yes readers, it's that time of year again when we look back with 
> amusement at the glitches and goofs served up by the media over the 
> last twelve months. It was a banner year for dubiousity, ranging from 
> failed presidential pop quizzes (Vajpayee, Lee, and um, that general 
> from Pakistan -- we knew that) to the amazing electoral counting 
> catastrophe (we thought those were boxes of Florida grapefruit; turns 
> out that's just what happens to a pregnant chad). So, return with us 
> now to those thrilling stories from yesteryear...
> 
> 1). "Well, OK, maybe it's the SECOND time in 50 million years..."
> 
> The August 19 New York Times front page was a real scorcher - 
> complete with color photograph. "The North Pole is melting" read the 
> first sentence. It seems tourists on a Russian ice-breaker saw open 
> water in the middle of the polar ice, clicked the shutter, and rushed 
> right to The Times with "evidence that global warming may be real and 
> already affecting climate." It was a sight "presumably never before 
> seen by humans... the last time the pole was awash in water was more 
> than 50 million years ago." National Public Radio also heralded the 
> news (Aug. 22), but their science reporter, Richard Harris, started 
> to notice the story's own thin ice, and got a skeptical response from 
> other Arctic experts. It turns out during a typical summer, about 90 
> percent of the high Arctic is covered with ice, but about 10 percent 
> of the time there's open water over the pole. The Times started 
> backtracking, and on August 29 revisited the entire matter with a 
> Science Times article altering the claim and
> 
> 2). Perhaps you should stick to the swimsuit competition?
> 
> Could nuclear power plants be a cause of infant mortality? This 
> charge was leveled in Washington, DC (Reuters Apr. 27) by activist 
> groups coupled with the star power of supermodel Christie Brinkley. 
> Though infant mortality rates have been in sharp decline at the same 
> time that nuclear power spread around the country, the activists had 
> an angle -- improvements in infant health were linked to the closing 
> of nuclear plants. But a quick call to the National Cancer Institute 
> (NCI) by Newsday reporter Earl Lane pulled the plug. An NCI study 
> that examined 900,000 cancer deaths in counties near nuclear 
> facilities showed that childhood deaths from diseases such as 
> leukemia were actually higher before, not after, their construction. 
> If anything, the facilities were associated with better infant health.
> 
> 3). Migrating Monarchs Take Wrong Turn?
> 
> What was the notorious Butterfly Ballot doing in wintry Canada, so 
> far from its Palm Beach electoral home? The December 7 issue of 
> Nature presented research from a Canadian psychologist who offered 
> ballot choices to shoppers at the Bonnie Doon Mall in Edmonton, 
> Alberta. It seems that 3 out of 53 Canadian shoppers made the alleged 
> "Buchanan error" on the ballot form; that is, they got confused and 
> voted for the wrong candidate. The San Francisco Chronicle (Dec. 1) 
> and other papers were quick to net the story, reporting evidence of 
> "systematic confusion" and arguing the data "call into question the 
> validity of the presidential election results." A story with, er, 
> wings, or just a lepidopteran let down?
> 
> It turns out the methodology was meandering. The shoppers were only a 
> "convenience sample" not even representative of Canadians, much less 
> Floridians. Second, the study acknowledged that "there was no 
> relation between the amount of confusion and errors" made (so much 
> for "systematic"). Besides, the Associated Press (Nov. 10) had 
> already reported butterfly ballot "research" from Bossier City, 
> Louisiana. Down there two fourth grade teachers had tested 22 kids on 
> the same confusing format, with zero errors.
> 
> 4). "So, have you stopped beating your wife yet?"
> 
> Reuters reported a poll (Feb. 4) which illustrated that "although 
> most smokers in the US know that cigarettes can cause heart and lung 
> disease, few have been able to kick the habit." In fact, out of the 
> 70 percent of the sampled respondents who had ever tried to quit, 
> none had succeeded. Unfortunately, this should have come as little 
> surprise, since the poll specifically sampled "more than 1,000 adult 
> smokers." No quitters allowed.
> 
> 5). The View From... Lake Rudolf?
> 
> It's not exactly "Eurocentric," but there's definitely something 
> wrong here. According to the BBC (Oct. 30), research into the last 
> universal common ancestor of human males living today "gives an 
> intriguing insight into the journey of our ancestors across the 
> planet, from eastern Africa into the Middle East, then to southeast 
> and southern Asia, then New Guinea and Australia, and finally to 
> Europe and Central Asia" (emphasis added)
> 
> It seems that modern man hasn't yet reached North or South America. 
> Perhaps the controversy over Kennewick Man goes deeper than we think.
> 
> 6). Kindergarten Cop-out
> 
> An Associated Press story, "Federal study shows kindergarten improves 
> all young minds" (Dec. 4), seemed to suggest that kindergarten was a 
> very good thing. From a sample of 22,000 children who attended 
> kindergarten, the study found that five times as many of them could 
> do simple sums as in the previous year.
> 
> But the fifth paragraph of the story reveals how the study doesn't 
> really prove anything about children's education in general: "The 
> Education Department-funded study offers no comparison with children 
> who do not attend kindergarten." In other words, we don't know 
> whether kindergarten-educated children are better off than other 
> children. All the study showed was that kindergarten helps educate 
> children who are in kindergarten. But "all young minds?" - we just 
> don't know.
> 
> 7). Fuzzy Math
> 
> Nobody at the AP raised an eyebrow when its Oct. 28 story, "Clinton 
> signs law to combat violence against women," repeated the President's 
> claim, "'Every 12 seconds, another woman is beaten,' he said. 'That's 
> nearly 900,000 victims every year.'"
> 
> Errr, no. One incident every twelve seconds translates to over two 
> and one half million incidents a year. Or, looked at the other way, 
> 900,000 incidents a year is one every 35 seconds. Either way, those 
> two figures don't add up.
> 
> 8). Cashew, Cashew, We All Fall Down
> 
> New York Times columnist Paul Krugman found another way to criticize 
> anti-globalization protestors in his Apr. 19 column, "A Real Nut 
> Case." He claimed that the World Bank's intervention in Mozambique's 
> cashew industry benefitted the country's poor farmers, who had 
> suffered compared with the nation's 10,000 nut processing workers.
> 
> Unfortunately for Mr. Krugman, and for Mozambique as well, 
> investigations later in the year by the Washington Post ("A Less Than 
> Helpful Hand; World Bank, IMF Blamed for Fall of Mozambican Cashew 
> Industry," Oct. 18) and Knight Ridder ("World Bank Policies Had Mixed 
> Results in Mozambique," Sep. 17) found that the World Bank's policies 
> had not only put over 7,500 factory workers out of a job in one of 
> the world's poorest countries, but that the farmers who were supposed 
> to have benefitted had lost out to nut speculators, many of them 
> foreign. (Thanks to TomPaine.com for initially drawing our attention 
> to this one).
> 
> 9). Ancestral Vices
> 
> A sense of perspective is important when you deal with statistics. A 
> spokesman for the White House Office of Drug Control Policy clearly 
> lost his when he responded to a study on the number of drug offenders 
> in prison by saying, "Over the same period of time, drug use has gone 
> down and crime is at an all-time low." ("Drug Offenders Jailed at 
> High Rate, AP, Jul. 27)
> 
> While crime has gone down recently, it still has not reached the low 
> levels it began to leave behind in the late Sixties. Murder rates are 
> lower than in the gangster-ridden 1920's and 1930's, but far above 
> the levels of the 1950's and the first two decades of the century.
> 
> Of course, we don't have the data to talk about crime levels before 
> that, but perhaps the spokesman had something else in mind. When Cain 
> murdered Abel, after all, the homicide rate peaked at 25,000 per 
> 100,000 individuals. And the Garden of Eden suffered a 50 percent 
> larceny rate, which was, naturally, motivated by a desire for illegal 
> substances.
> 
> 10). And Finally...
> 
> "Hyper-hyperbole. It's massive!"-article title in the UK newspaper 
> The Observer, Feb. 27, criticizing news media which make exaggerated 
> claims to bolster their arguments. "It's apocalypse now as world 
> boils over."-headline in the same newspaper, same day.
> 
> The Statistical Assessment Service (STATS) is a nonprofit nonpartisan 
> think tank in Washington, D.C. examines the way that scientific, 
> quantitative, and social research are presented by the media.
> 
> Contact:
> Howard Fienberg
> Statistical Assessment Service
> 202-223-3193
> http://www.stats.org
> -- 
> 
> Bill Steele
> ws21 en cornell.edu
>