Tuesday, June 16, 2009

Eerie Rise in College-age Drinking Deaths

This just out:

"Using figures from government databases and national surveys on alcohol use, researchers at the National Institute on Alcohol Abuse and Alcoholism (NIAAA) found that drinking-related accidental deaths among 18- to 24-year-old students have been creeping upward -- from 1,440 in 1998 to 1,825 in 2005."

Before you shriek in horror, consider this truly frightening fact: 1,825 deaths is exactly five a day. What are the odds? Probably exactly the same that the number given for 1998, 1,440 deaths, is exactly one hundred and twenty a month.

Someone needs to say it, and I will: The NIAAA's Ralph Hingson* is pulling these numbers right out of his ass. College drinking deaths are an excruciatingly painful loss -- as the parent of two teenagers, it's starting to prey on my mind -- but this kind of ridiculous exaggeration is an offensive travesty. Take a look at this fact-based perspective on the issue.




* Heard of this guy? Here's some info from Behind the Neo-Prohibition Campaign: The Robert Wood Johnson Foundation, by Dan Mindus, an analyst at the Center for Consumer Freedom**

Boston University sociologist Ralph Hingson runs the RWJF-funded “Join Together Online” program, which serves as a clearinghouse of information for the anti-alcohol movement. He is also one of the most careless--and frequently cited— researchers around. Hingson produced a study declaring that a national .08 BAC law would save “500 to 600 lives a year.” But to come up with that number, Hingson went cherry-picking, carefully selecting only certain .08 states so that the results would align with his prejudices. Perhaps the most highly respected auditor in the world, the GAO, calls Hingson’s study “unfounded.” Nevertheless, it is still cited by MADD and other neo-prohibitionist groups.
• In 2002, Hingson authored a headline-grabbing study that blamed alcohol for 1,400 fatalities among college students each year. Aside from stretching the definition of “alcohol-related fatality” to ridiculous lengths, Hingson did no research whatsoever with college students. Instead, he simply took statistics about alcohol-related fatalities among 18 to 24 year-olds and multiplied by their percentage of the general population. Unfortunately, this blunder didn’t stop
The New York Times from beginning its coverage by reporting: “On an average day, according to a new study, four college students die in accidents involving alcohol.”


**Who are hardly lily-white themselves...though please note that this source, "Sourcewatch," practically gives the ridiculously anti-alcohol-biased Robert Wood Johnson Foundation a free pass...so I question their impartiality, as should you. Far as that goes, you should question mine. Hell, I do.

4 comments:

Steve said...

I'm guessing they never mentioned the number of 18-24 year olds in collegs has also risen? Just pulled this from (http://nces.ed.gov/fastfacts/display.asp?id=98):

"Enrollment in degree-granting institutions increased by 16 percent between 1985 and 1995. Between 1995 and 2005, enrollment increased at a faster rate (23 percent), from 14.3 million to 17.5 million. Much of the growth between 1995 and 2005 was in female enrollment; the number of females enrolled rose 27 percent, while the number of males rose 18 percent. During the same time period, part-time enrollment rose by 9 percent, compared to an increase of 33 percent in full-time enrollment. Enrollment increases can be affected both by population growth and by rising rates of enrollment. Between 1995 and 2005 the number of 18- to 24-year-olds increased from 25.5 million to 29.3 million, and the percentage of 18- to 24-year-olds enrolled in college rose from 34 percent to 39 percent. In addition to the enrollment in accredited 2-year colleges, 4-year colleges, and universities, about 434,000 students attended non-degree-granting, Title IV eligible,1 postsecondary institutions in fall 2005."

The increase is so minimal compared to the increase in enrollment, that it's relatively moot. It increased approximately 0.7%.

Like Homer Simpson once said, “Oh, people can come up with statistics to prove anything. 14% of people know that.”

Russ said...

Good point, Steve. There are all sorts of questions one can raise about the veracity of the data (one issue I have is that, according to the abstract, the study used NHTSA data, and the NHTSA categorizes a drunk 60-year-old hitting and killing four sober teenagers as four alochol-related teenage deaths). Beyond this, though, is the issue of causation. If the numbers really have gone up, what do we blame them on?

What really infuriates me is they no matter what happens the neo-prohibitionists will claim victory. Deaths go up? We've got a problem; time to enact more restrictive policies. Deaths go down? Restrictive policies are working, so let's enact more of them! Deaths stay the same? "The persistence of college drinking problems underscores an urgent need to implement [restrictive policies.]" That's taken right from the press release. Yet in all this, nobody questions our overall strategy of zero-policy until you're 21. We don't tell hunters to never touch a gun until they turn 21, at which point they should run out and bag as many deer as possible. Why aren't we questioning this approach when it comes to alcohol?

Harry Spade said...

This study has convinced me to not send my kids to college.

Carey said...

Lew and Steve makes good points. As someone with quite a bit of mathematics as part of my life I see the gross misuse of statistics constantly, even by people who should know better, in the media, to further research, etc. There is the sample size as Steve mentions, and then the matter of the two numbers being statistically significant as Lew alluded to (e.g., 2.3% = 2.45% when the means and variances are considered).

Then there is probably a good yard of slop in the way the local officials employ their taxonomy from case to case and year to year (does it go to alcohol, car accident, other drugs, a fall..?)

What if the following statistic came out? Would this ancillary information change the perception? Of the population of alcohol related campus deaths in 2008 versus 1998, 65% of the former versus 11% of the latter were taking prescription anti-depressants or anti-anxiety and imbibing alcohol.

Humans aren't good unless it is either a linear storyline or a dimensional reduction. We are limited by not being able to compute on too many degrees of freedom. Any time you reduce the dimensionality of the problem it is easier for humans, computers, or algorithms, but you loose some information in the precess.

I hope the panjandrumancy at the NIAAA employ some good researchers to get good data out in general.

Anyway, very sad regardless.

<\ end rant>