Freakonomics in the Times Magazine: The Seat-Belt Solution

The July 10, 2005, Freakonomics column, “The Seat-Belt Solution: How Much Good Do Car Seats Do?” is about the efficacy of child car seats versus plain old seat belts. Click here to read the full article. This blog post supplies additional research material.

There are three main components of this Freakonomics column: 1) a visit to a four-day training seminar at which police officers learn proper car-seat installation; 2) original research that compares auto fatalities for children riding in car seats versus children riding in seat belts; and finally, 3) a Freakonomics-commissioned crash test.

  1. The four-day training seminar described at the beginning of the column used a training manual assembled by the National Highway Traffic Safety Administration (NHTSA) and was organized by Wendy Shindler, a trauma nurse coordinator at Good Samaritan Hospital in Suffern, N.Y. Shindler runs the New York State chapter of a national movement called Safe Kids, which is at the forefront of promoting legislation concerning car seats and booster seats. Click here for a state-by-state rundown of current child car-seat laws.
  2. The research in the column is further detailed in a paper by Steven D. Levitt, a preliminary version of which is available here. The data come from the Fatality Analysis Reporting System (FARS), which compiles police reports on all fatal crashes in the U.S. since 1975, and is housed on the NHTSA website. You don’t need a Ph.D., however, to crunch the numbers that might lead to skepticism about the benefit of car seats. See for yourself: start by clicking here to access the FARS web-based encyclopedia. In the yellow box on left, click “Create a Query.”
    • Under “Step 1,” leave “Select a Year to Query” as is (2003) and click “Submit.”
    • Under “Step 2,” click the following boxes in the “Persons” column: “Age,” “Injury Severity” and “Restraint System Use”; now click “Submit.”
    • Under “Step 3,” go to the “Age” box and click on “2 Years,” “3 Years,” “4 Years,” “5 Years,” and “6 Years” (hold down Control key for to click multiple ages). Leave all other settings as they are. Now click “Cross-Tab.”
    • Under “Column,” select “Restraint System Use.”
    • Under “Row,” select “Injury Severity.” Leave other settings as they are.
    • Now click “Submit.”

    And there’s your report. The bottom row of numbers (under “Total”) represents the total number of children ages 2-6 involved in fatal crashes (that is, crashes in which someone died, not necessarily a child) in 2003, broken down by category of restraint. To calculate rate of death, divide the number of children in each category into the number of fatal injuries (fourth row from the bottom) in each category.

    Of the 941 children in properly-installed child safety seats, 95 died (10.1%); of the 51 children in improperly installed child safety seats, 23 died (45.1%). Combine these two categories (992 children and 118 fatalities) and you arrive at a death rate of 11.9% for children traveling in car seats.

    Now for children wearing seats belts: of the 566 children in lap-and-shoulder belts, 59 died (10.4%); of the 246 children in lap belts only, 27 died (11.0%); of the 2 children wearing only a shoulder belt, 1 died (50%); of the 16 children using a seat belt improperly, 5 died (31.3%). Combine these four categories (830 children and 92 fatalities) and you arrive at a death rate of 11.1% for children wearing seat belts.

    While this raw data is far from definitive, the FARS website allows you to calculate the results while controlling for any number of variables – seating position in cars, types of crashes, old cars versus new cars – and the results remain extremely robust.

    (Also: click here to read the Insurance Institute of America’s 2001 memo to NHTSA concerning booster seats that was quoted in the column.)

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.

 

COMMENTS: 1

View All Comments »
  1. Slavek says:

    I’m sure these “unexpected” conclusions sell books, but unfortunately they are far too biased to be of any significant value.

    The counter point:

    http://intlpress.com/site/pub/files/_fulltext/journals/sii/2009/0002/0004/SII-2009-0002-0004-a005.pdf

    Thumb up 0 Thumb down 0