Car Crashes: Toll on Children

Car Crashes: Toll on Children

In the United States, motor vehicle crashes are the leading cause of death for children over the age of one. Those facts are especially chilling to Michael Elliott, an assistant professor of biostatistics and a father of two, who studies the incidence of child injuries in automobile accidents.

“If we’re going to reduce the number of children who are injured,” Elliott says, “we need to know more about how and why they get hurt.”

For the past six years, Elliott has been working with Partners for Child Passenger Safety, a study aimed at reducing the childhood injuries and deaths that result from motor vehicle crashes. Launched by the Children’s Hospital of Philadelphia and the auto insurer State Farm, the study combines in-depth telephone interviews, on-site crash investigations, and computer crash simulations with interdisciplinary analysis and interpretation.

Elliott helps design and analyze the study so that its statistical findings are reliable. For example, to understand the nature of injuries and their risk factors, researchers sample a disproportionately large number of children in more severe crashes. So Elliott and his colleagues “weight down,”or count less heavily, children in overrepresented crashes.

And because standard methods to account for over- and underrepresentation make for less-stable estimates, Elliott is developing statistical “weight-trimming” methods to make more efficient use of the information from the study.

He is also using and extending causal model methodology to untangle the effects of confounding factors. For instance, if a particular restraint type is effective, it will reduce the risk of death. But studies that focus on fatal crashes will underestimate its effectiveness since survivors won’t be in the data. On the other hand, studies that include all kinds of crashes may overestimate the restraint’s effectiveness because inherently safer drivers may be more likely to use the device to restrain children.

Elliott and his colleagues on the study hope to quantify the effectiveness of child restraints, identify potential problems with certain types of vehicle construction, track changes in restraint usage and injury rates, and look at driver statistics. Elliott notes, for example, that a “non-trivial fraction” of passengers under 16 are in cars driven by teenaged drivers, and that children in crashes with teen drivers are at substantially elevated risks of injury—a finding that suggests some sort of policy action may be warranted.

Since its inception, the study has provided empirical evidence to support new seatbelt and child-restraint laws. But as Elliott points out, with a million and a half U.S. children still involved annually in motor vehicle crashes, more work is clearly needed.

--Photo by Peter Smith

Send correspondence about this or any Findings article to the editor at You will be contacted if your letter is considered for publication.

Michael Elliott was lead biostatistician on the January 2007 report "Driving: Through the Eyes of Teens." In this national survey, 5,665 high school students describe their experiences and perceptions as drivers and passengers.