A common belief about teenagers is that they implicitly assume that they are invincible or immortal and think little about their own deaths. A new study just published in the Journal of Adolescent Health shows this to be a myth, however, as they vastly over-estimate their chances of dying within the next year.
By the mid-teens, our ability to judge the likelihood of uncertain events is usually equal to that of adults, so we might expect that adolescents can judge the chance of death as accurately as grown-ups.
This study, led by psychologist Baruch Fischhoff, surveyed 3,436 14-to-18 year-old adolescents and a local group of 124 seventh graders and 132 ninth graders asking them to estimate their chance of dying in the next year, enquiring about what sort of neighbourhood they lived in, whether they’d experienced or witnessed any violent events and whether they’d had any serious health problems.
Although the statistical death rate is 0.08%, the most common estimates where that they had a 5% of 10% chance of dying within the next year. Interesting, there was a larger than expected number of teens who judge their chance of dying within the next year as 50%, although this likely suggests that they were indicating a sort of 50/50 answer as a way of expressing “I don’t know”.
Adolescents assumptions about how likely they were to die were strongly related to their reports of how much crime they expected to experience and not or only very weakly related to if they’d experienced violent events or had health problems.
In other words, teenagers seem to be personally pessimistic and live in a world where they perceive themselves to have a high chance of dying despite the relatively small actual risk.
Link to PubMed entry for study.
Mind Hacks 
Explains Goths.
What would be more interesting is to show that teenagers overestimate their chances of death *relative to other age groups*. As Baruch knows, people in general have a hard time making these kinds of numerical estimates.
The first thing that sprung to mind was how we tend to “medium-size” figures. If the question was “what is the percentage of likelihood that you will die in the next year?” you will probably get different answers than something like:
“In what order of magnitude is the likelihood you will die in the next year? a) between 1:100 and 1:1,000 b) between 1:1000 and 1:100,000 c) 1:100,000 and 1:1,000,000 and d) smaller than 1:1,000,000”.
Then there is the issue of whether or not a 14-year-old even understands a question like that, so crafting its representation would be important. Of course there are also potential issues of framing/anchoring and other similar fun.