Do Safety Seats Protect All Kids? New Research, New Debate
It's something that's ingrained in us early on in our roles as parents and caregivers: Young children should be buckled up in a safety seat whenever they are in a car. Recently, however, that commonly held belief has been challenged by Steven D. Levitt, an economics professor at the University of Chicago and co-author of the best-selling book, Freakonomics. His research indicates that today's child safety seats are no more effective at protecting children above age two than lap and shoulder belts. Levitt and co-author Stephen J. Dubner published his research in a July article in the New York Times Magazine.
Levitt's statements fly in the face of years of research by the child safety community, automakers and medical professionals that conclude child safety seats are far more effective than seatbelts alone at preventing death and injury to children in car accidents. The National Highway Traffic Safety Administration (NHTSA) concluded that car seats have saved an estimated 6,000 lives between 1975 and 2003. As a result of these studies, every state in the country has enacted child occupant protection laws.
Even so, many parents and caregivers may be wondering if they can — or should — abandon their car or booster seats. However, child safety experts within the automotive community feel that Levitt's findings need further scrutiny before anyone takes a child out of a safety seat.
Don't Throw It Away Yet
No one is suggesting parents or caregivers stop using these devices altogether. All researchers, including Levitt, agree that child safety seats are the best solution for kids age two and under. It's with children over age two that the opinions diverge.
"My research shows that car seats are not any more effective than regular lap and shoulder belts [at protecting children over age two]," Levitt says. "It doesn't suggest they are any worse."
To arrive at this conclusion, Levitt analyzed the Fatality Analysis Reporting System (FARS), a national database of accidents from 1975 to 2003 compiled from police reports. A full explanation of his analysis is in an academic paper he published on the Freakonomics Web site, but at the core of his research are two statistics: Based on Levitt's analysis of the FARS data, 18.3 percent of children ages two to six involved in auto accidents who were restrained in child safety seats died, versus 19.2 percent of children ages two to six restrained by a lap and shoulder belt only.
The child safety community contends that car seats may be more critical in preventing injuries, which are far more common than deaths in auto accidents. In 2003, about 241,000 children under age 16 were injured in the U.S. in auto accidents, while about 1,800 children under age 16 were killed.
"While preventing the fatality of a child is very important, car seats and boosters are really very effective at preventing serious internal and spinal injuries that often affect children in a crash," says Rae Tyson, NHTSA spokesperson.
Levitt also analyzed children who were injured in crashes and concluded that there was no substantial reduction in injuries to children in car seats versus in seatbelts alone. However, he describes this conclusion as "uncertain," because the FARS data includes only crashes in which at least one fatality occurs. Levitt is currently studying other data sets to determine if this conclusion holds true for less severe crashes.
In the meantime, other studies indicate that car seats are far more effective at preventing injury than lap and shoulder belts alone. Data from the Partners for Child Passenger Safety (PCPS), a research effort between the Children's Hospital of Philadelphia (CHOP) and State Farm Insurance, shows that child safety seats reduce injury risk by 71 percent (as compared to seatbelts alone) in children ages one to four. For kids ages four to seven, the data shows a 59 percent reduction in injury risk by using a booster seat versus a seatbelt alone. Those 2003 statistics come from the PCPS database, the largest database on injuries from car crashes in the U.S.
Different Data, Different Conclusions
Child safety researchers often use different data, so it's standard practice for a researcher to present his data analysis to an objective peer group within the medical and scientific community before it's published. The review process can last six months or more.
At the time of our interview, Levitt had not participated in a peer-review process, nor had he discussed his findings with anyone from the automotive or child safety community.
"[Levitt] is clearly a brilliant economist who has done a thorough analysis of the data and developed a method for compensating for the data's limitations," said Dr. Dennis Durbin, an emergency room doctor at CHOP and a leading PCPS researcher. "But because he has not held his research up for peer-review, we at CHOP can't be confident that his method really compensates for those limitations. We are not so sure that his methods make the data representative of all crashes. That may be why there is such a difference between his research and what has already been published by CHOP and others."
In contrast, CHOP has published over 40 peer-reviewed papers since 1999. Safety researchers from NHTSA, the Insurance Institute for Highway Safety (IIHS), and automakers such as General Motors, Ford and Volvo also regularly participate in conferences and meetings to share research findings.
Can Crash Tests Lie?
As part of Levitt's analysis, he and co-author Dubner hired a firm to conduct independent crash tests. Using the same three-year-old and six-year-old dummies as in federal crash tests, they tested the dummies in seatbelts alone, as well as in a car seat (for the three-year-old) and in a booster seat (for the six-year-old). Their results showed "nominally higher" head and chest acceleration for the three-year-old dummy in the seatbelts alone than in the car seat, and nearly identical results for the six-year-old dummy in both restraints.
But these types of tests don't simulate real-world scenarios as well as adult crash tests do. While child dummies' hard plastic pelvises actually help hold the lap portion of the belt in place, real children's pelvic bones, which aren't fully developed, cannot. Further, children are more comfortable if their knees bend at the edge of the seat, and as a result, they often slump, causing the lap belt to ride too high and rest on their abdomens. In a crash, these kids often slide down and out of the belt, an effect called "submarining." None of the child crash test dummies currently in use are sophisticated enough to allow experts to collect numbers on submarining, yet it is one of the most common injuries to children in car crashes.
The Right Fit
The submarining issue is pivotal among children in this age group. "Because the geometry is different in a child than an adult, you need to have the lap belt over their thighs. By elevating the child in the seat [via a booster], it helps position the belt properly," says Ingrid Skogsmo, director of Volvo Cars Safety Center.
"For older kids, using a restraint so they do not scoot their bottom forward will improve the fit of the seatbelt and reduce their potential for injury," agrees Dr. Steve Rouhana, group lead for occupant protection at Ford's Research Science Lab. (For more on older kids and booster seats, see "Booster Seats: Fight the Good Fight."
Using a booster seat isn't a "one size fits all" proposition and doesn't guarantee improved seatbelt fit for every child. First, not every rear vehicle seat will accommodate every booster seat. Beyond that, not every booster seat will properly fit every child. "Your best bet is to take the child, the booster and the vehicle and test out all three [at once]," says Susan Ferguson, senior vice president of research at IIHS.
Plus, children's bodies change shape rapidly. The same belt-positioning booster seat may change the seatbelt fit for your child as he or she grows. Lorrie Walker, chief training manager for National Safe Kids, said, "Parents really need to assess the child every few months in the safety seat and in the different vehicles he or she travels in."
Two Sides, Come Together?
Levitt is not advocating getting rid of booster seats. Instead, his main concern is that car seats and booster seats are not built into vehicles at the factory. Chrysler, Dodge and Volvo offer built-in booster seats as an extra-cost option in some of their vehicles. If you want a built-in seat with a five-point harness, though, you're out of luck.
"I'm not advocating using adult seatbelts for kids," he says. "What I'd like to see is some of the $300 million that's invested in the car seat industry allocated to further study and improve the technology already available in cars — seatbelts — to better suit children."
But designing a belt that fits both kids and adults is easier said than done. According to Artie Martin, technical integration engineer for child safety at General Motors, "It's not just adjusting the shoulder. It's also the location of the lap belt anchorages. By designing to one end of the spectrum [like kids or 95th-percentile adults], you shortchange a whole range of people."
CHOP's Durbin sees both sides. "Customizable restraints in rear seats are a better long-term solution than today's child restraints. However, even if we had fully customizable restraints in every new vehicle coming on the market right now, we would still need a solution for all the vehicles that are on the road today. Child safety seats are the best solution to ensure the safety of our kids when they are in a crash."
Of course, to do their job, car seats have to be installed correctly. Since only about 20 percent of them are, according to the AAA, we suggest that parents and caregivers double-check their car seat installations (see "How to Install a Car Seat"). Also consider letting an expert inspect your work: Free car seat checks are available by calling 1-866-SEAT-CHECK or visiting Seat Check's Web site.