Tracing a Statistic: Teen Suicide

By Kevin

Two-thirds the way down on page 15 of Alexandra Robbins' The Overachievers, I come across this statistic:

[Overachiever culture] has diminished leisure time for all ages. It is believed to be a major factor in the 114 percent spike in suicide rates among fifteen-to-nineteen-year-olds between 1980 and 2002.
I'm not an expert on these types of important social issues, but that number didn't look quite right to me, so I flipped to the endnotes, and found:
15.114 percent spike: The important article by Stepp, Laura Sessions. "Perfect Problems/" Washington Post, May 5, 2002.
Unfortunately, that article is not online (for free), but using my local library's electronic database, I found it:
"I've seen a dramatic change in the stress level of these kids," says Carolyn Callahan, who has worked with high achievers for 30 years, currently at the National Research Center on the Gifted and Talented at the University of Virginia. "They're going through the motions and not enjoying what they're doing." The perfection machine, what she calls a treadmill, "has created a situation where they don't feel they have a choice to get off."

One report in a newsletter by Callahan's center worries that the characteristics of these students, including "perfectionism" and "supersensitivity," put them at risk for suicide, and notes that the proportion of young people ages 15 to 19 who have taken their own lives has jumped 114 percent since 1980.

We're getting warmer. A simple Google search revealed the report in the newsletter. It starts off like this:
The rate of suicide among children 10 to 14 years of age increased 100% between 1980-1996. Among youngsters 15-19 years of age, the rate of increase was 114%, making suicide the fourth leading cause of death for this age group (U.S. Department of Health & Human Services, 1999).
So it is not true that the suicide rate for 15-19 year olds increased by 114% from 1980 to 2002 -- 2002 being when the WaPo article was written --, but from 1980 to 1996.

But I had to keep going. Is that just as depressing statistic even true? We're still not sure how this data was created and validated. According to the footnote, the HHS report is actually the Surgeon General's Call to Action to Prevent Suicide 1999. Another Google search reveals the SG report, in which we find.:

From 1980 to 1996, the rate of suicide among persons aged 15-19 years increased by 14% and among persons aged 10-14 years by 100%
Umm.... It's 14%, not 114% percent!?! Let's go to the source of those statistics! I would, but those aren't sourced at all in the SG document!

Fortunately, hard data do exist, but I found it difficult to source the original CDC table ("Death rates for 72 selected
causes by 5-year age groups, race, and sex: United States, 1979–1997. Worktable GMWK 291 Trend B"). These folks do cite the data, and do me one better by graphing it:


We're interested in the white squares. So the real growth in the 15 to 19 year old suicide rate is 14%, not 114%. The red line shows where a constant suicide rate would have ended up in 1996. The blue line shows where the alleged 114% increase would have left us in 1996.

What an utter refusal to check sources and validate simple statistics! THIS IS NOT MY JOB, nor the job of any of Ms. Robbins' readers. It's the job of the author and editors. I don't know if I should even bother continuing to read the book at all, as I've spent 1/2 hour tracking down just one horrendously wrong data point. How many more will be this wrong???


I'm not saying teen suicide isn't a problem, or that we should pretend that overachievers don't have problems. I am saying that understanding the actual scope of the problem is vital in arranging our political priorities. And for that we need solid data, not this crappy series of citations of tertiary sources.

Moral: Next time you hear a politician say we need a "national discussion" on an issue, realize that the type of "facts" they want to talk about are frequently no better sourced than the example cited here.


paul wrote:

Does the statistics include attempted suicides?

-- November 18, 2006 4:59 AM

Tim Worstall wrote:

Far worse is this:

'[Overachiever culture] has diminished leisure time for all ages.'

Leisure time has been increasing for all age groups for the past centruy: ever since we started measuring it actually.

-- November 18, 2006 5:39 AM

Kevin Brancato wrote:

The data represent deaths due to suicide.

-- November 18, 2006 10:05 AM

William Hart wrote:

Nice job on the fact-checking. Part of the hurrah over the book "The Skeptical Environmentalist" was over this very issue. What are the correct numbers and who generated them? Once, as an exercise in a grad class I taught in Community Nutrition, I asked the students to look up how many homeless there were in St. Louis. They had to have a citation from somewhere to back-up their number. Official answers ranged from 800 (US Census Bureau) to 12,000+ (Local advocacy group for the homeless). Only the Census Bureau told the students how they generated their estimate. Everyone else sort of danced around that. Led to a great in-class discussion about how estimates are made and how the number chosen as real can affect dollar allocation.

-- November 20, 2006 3:17 PM

pacific_waters wrote:

My grandfather had so much more leisure time on his hands than I do after he finished his 60 hour work week.

-- November 23, 2006 5:51 PM

thats what she said wrote:

same as paul.
do these statistics include attempted suicide?

-- November 15, 2007 9:49 PM

MV wrote:

To: that's what she said

See the third comment. No, this is CDC's data collection on types of death. Attempted suicides are much harder to gather good data on.

-- December 18, 2007 7:56 PM

henry rollins wrote:

all this makes no sense at all because im looking for a number of deaths not a whoever made this little article about nothing really why did you make it???

-- December 1, 2008 1:58 PM

lala wrote:

This graph is not as great as i expected it to be but i love the other comments.

-- April 24, 2009 1:13 PM

Post a comment