Thursday, October 19, 2006

T-19 Days and Counting: Lies, Damn Lies, and Statistics

A number (pardon the pun) of recent stories would seem to be bad news for the Republicans:
Study: War blamed for 655,000 Iraqi deaths (Cnn.com)

War has wiped out about 655,000 Iraqis or more than 500 people a day since the U.S.-led invasion, a new study reports.

Violence including gunfire and bombs caused the majority of deaths but thousands of people died from worsening health and environmental conditions directly related to the conflict that began in 2003, U.S. and Iraqi public health researchers said.

"Since March 2003, an additional 2.5 percent of Iraq's population have died above what would have occurred without conflict," according to the survey of Iraqi households, titled "The Human Cost of the War in Iraq."
Poll Signals More Republican Woes (WSJ.com) (H/T Real Clear Politics)

A new Wall Street Journal/NBC News poll illustrates the political toll Republicans are paying for rising discontent over the Iraq war, as well as a spate of scandals including the disclosure that Republican House leaders knew of inappropriate emails to House pages from Florida Rep. Mark Foley, who resigned late last month. Voters' approval of Congress has fallen to 16% from 20% since early September, while their disapproval has risen to 75% from 65%.
Warm winds of change hit the Antarctic (news@nature.com)

[...] The stronger winds cause warming mostly in the summer. Warming on the Antarctic Peninsula has actually been most intense in winter, but in summer a large part of the extra heat goes into melting ice. This has dramatic consequences.
Percolating meltwater enlarges crevasses and leads ultimately to the disintegration of floating ice shelves.


The largest and most prominent such event happened in March 2002 when the Larsen ice shelf, with an area of 3,250 square kilometres, collapsed. Overall, more than 13,500 square kilometres, an area larger than Jamaica, of floating ice shelves have broken up in the past 30 years. This is expected to speed the flow of inland ice to the coast, accelerating global sea level rise.

The 2002 event can now be pinned down to a specific change in climate, which is in turn linked to human-induced global warming, the authors say. Some argue that this is the first single event proved to have been caused by manmade climate change. "It's close to being evidence," says Ted Scambos, lead scientist of the National Snow and Ice Data Center at the University of Colorado in Boulder.
What's wrong with these stories? The authors make inferences about causality that may or may not be valid. My mother, who studied statistics in the 30's at Columbia University, liked to relate a tale told by one of her professors to illustrate the difference between correlation and causality.

It seems that a certain researcher had found that older women were more likely to walk with their toes pointed outwards than younger women. He concluded that women's feet turned more outward as they aged. A later researcher took another look at that study, did some anthropological digging, and concluded that age had nothing to do with it: older women were more likely to have been taught as young girls to walk with their toes pointing out. The correlation was solid, but what caused the condition was open to debate.

Some counter-arguments for the stories above:
655,000 War Dead? A bogus study on Iraq casualties. (OpinionJournal.com)

After doing survey research in Iraq for nearly two years, I was surprised to read that a study by a group from Johns Hopkins University claims that 655,000 Iraqis have died as a result of the war. Don't get me wrong, there have been far too many deaths in Iraq by anyone's measure; some of them have been friends of mine. But the Johns Hopkins tally is wildly at odds with any numbers I have seen in that country. Survey results frequently have a margin of error of plus or minus 3% or 5%--not 1200%.
An email to Hugh Hewitt (a followup to Hugh's earlier post):

Thank you so much for bringing attention on the all the faulty polling being done. You are not crazy, you are absolutely, 100% spot-on on this. I have worked my entire adult life, 25 years, processing market research and public opinion surveys. I know enough about surveys to be able to construct one that shows people prefer Pepsi over Coke, 60%-40%, or vice versa, and you would have no idea how I got either result even if I gave you the internals and methodology. You don’t have to take it on faith that there is a media conspiracy to misrepresent polling, they admit it in their own poll results. I have yet to see a general population poll that did not show adults 2%-4% more liberal than registered voters and registered voters 2%-4% more liberal than likely voters, and yet the media has no qualms about citing adults or registered voters when they want to give the Dems an added boost. Likewise, most 7 day tracking studies show a liberal 2%-4% bias when collecting over the weekend, yet the weekend seems to be the favorite time to poll. But the most important flaw is the one you have been talking about, the fact that every poll seems to over sample Democrats by 5%-10% consistently.
Climate of Fear (OpinionJournal.com)

To understand the misconceptions perpetuated about climate science and the climate of intimidation, one needs to grasp some of the complex underlying scientific issues. First, let's start where there is agreement. The public, press and policy makers have been repeatedly told that three claims have widespread scientific support: Global temperature has risen about a degree since the late 19th century; levels of CO2 in the atmosphere have increased by about 30% over the same period; and CO2 should contribute to future warming. These claims are true. However, what the public fails to grasp is that the claims neither constitute support for alarm nor establish man's responsibility for the small amount of warming that has occurred. In fact, those who make the most outlandish claims of alarm are actually demonstrating skepticism of the very science they say supports them. It isn't just that the alarmists are trumpeting model results that we know must be wrong. It is that they are trumpeting catastrophes that couldn't happen even if the models were right as justifying costly policies to try to prevent global warming.
Advertising copywriters know that having "factual" numbers in their stories lend credibility; most people lack the technical or scientific knowledge to know how to refute them, or recognize when they've been manipulated. Furthermore, people remember the bold headlines on page one, but not the corrections buried inside in tiny print. So it's an uphill battle to counter the erroneous perceptions people get from the mainstream media.