A recent report by Meyer and Finney concluded that automobile airbags cause more deaths than they prevent. Using a sample of more than 20,000 people involved in crashes, the authors showed that fatality rates were higher for occupants in seats with airbags than for occupants in seats without. This contradicts much of the airbag research published to date. The present study examines the methods used by Meyer and Finney and identifies a number of problems. Most important is the lack of complete information for more than half of the eligible crash records. It is shown that the direction of the results can be changed by taking advantage of additional data. However, these new results also are suspect. The conclusion is that the missing data problem is so severe as to prevent a reliable analysis. (Author/publisher)
Abstract