I just read SuperFreakonomics by Steven Levitt and Stephen Dubner, and I enjoyed the book as much as the best-selling original Freakonomics book. While much has been made of the chapter on global warming, I actually found the material on children's car seats to be more fascinating, particularly as a father of three young children. Levitt's research suggests that car seats may not prevent fatalities more effectively than traditional lap and shoulder seat belts for children older than the age of 2. To examine the empirical work behind the book, take a look at this scholarly paper. Here's an excerpt from the conclusions to that research paper:
"The empirical evidence presented in this paper, however, suggests that for children aged two and up, child safety seats provide no discernible advantage over traditional lap and shoulder belts, and only a marginal improvement relative to lap-only seat belts in preventing fatalities. These conclusions are robust to the inclusion of a wide array of covariates, analyzing a variety of sub-samples of the data, including vehicle fixed-effects, and correcting for sample selection in the way the FARS data set is constructed. An obvious question to ask, although one which is beyond the scope of the FARS data, is the extent to which the failure of child safety seats to outperform seat belts is a consequence of child safety seats frequently being improperly installed. Indeed, NHTSA (1996)estimates that more than 80 percent of all child safety seats are incorrectly installed. Based on crash tests, Kahane (1986) argues that properly installed car seats reduce fatailities by 71 percent, compared to 44 percent for improperly installed safety seats. Thus, there may be potential gains to achieving better installation. On the other hand, it is worth noting that when I conducted my own crash tests at an independent lab using lap and shoulder belts on dummies corresponding to children aged 3 and 6, the seat belts performed well within the guidelines the federal government has established for child safety seats, and just about as well as the (properly installed) child safety seats that I tested. While far from definitive, the crash tests I conducted suggest that even with proper installation, there may not be clear advantages of car seats over seat belts."
I must say that I admire Levitt for taking on such controversial topics. I also liked his response to Transportation Secretary Roy LaHood's criticism of his research.