Wednesday, February 9, 2011

Causation or Correlation

In the past thirty years or so, there have been more than two dozen research papers published in American medical journals "linking" guns with a) homicide, b) suicide, or c) injuries. But as even the New York Times recently admitted (amazingly enough!) there has never been any finding of a direct cause-and-effect relationship between gun ownership and elevated risk of homicide, suicide or injury. These medical studies led one critic (was it Dave Kopel?) to ask, "On what basis do the researchers recommend not keeping a gun in the house when they have not documented any increase in mortality or morbidity as a direct result of having done so?"

One of my favorite comparisons is to a (hypothetical) study to determine the causes of obesity. The researchers recruited a thousand voluteers who were overweight to participate in the study. One of the things that the researchers did was to look into the eating habits of the volunteers, and when they did, they found that many of them consumed large amounts of diet soda. Their conclusion? "Consumption of diet soda is a risk factor for obesity."

It's a "reductio ad absurdum" joke, of course. No one, especially a team of Ph.D.'s, would be stupid enough to fall for that.

Oh, really?

New research that links diet soda consumption with an increased risk of heart attack and stroke has doctors urging caution about the controversial and preliminary results.

According to a study of more than 2,500 people presented today as a poster at the American Stroke Association International Stroke Conference in Los Angeles, people who drank diet soda daily had a 61 percent increased risk of cardiovascular events compared to those who drank no soda, even when accounting for smoking, physical activity, alcohol consumption and calories consumed per day.

"This study suggests that diet soda is not an optimal substitute for sugar-sweetened beverages, and may be associated with a greater risk of stroke," Hannah Gardener of the University of Miami and her colleagues reported at the conference.

ABC News


Josh Kruschke said...

There are no standards of proof in non-engineering fields. This really sad. Question, why did a peer not ask, "how so; and if so, how do we prove it?" This is why peer review doesn't work. The stander should be like in engineering. You test to destruction or failure. Papers should be put up for ridicule.
In the peer review process everyone knows there going through it so they're nice so others will be nice to them or if doesn't match what the majority thinks it's not published out of hand.

Just my thoughts,

Josh Kruschke said...

Sorry, my standard of writing was poor. I see I left some words out of that last paragraph.

My apologizes,