Background

In 1998, 49 States, the District of Columbia, and Puerto Rico each reported a statewide estimate of passenger vehicle shoulder belt use, based on surveys conducted following guidelines issued by the the National Highway Traffic Safety Administration (NHTSA) in September 1998. The guidelines were issued to establish a new occupant protection incentive grant program that was authorized under Section 1403 of the Transportation Efficiency Act for the 21^st Century (TEA-21). This grant program, Section 157 of Title 23 of the United States Code, required states to provide estimates of their safety belt use rates that were "accurate and representative."

The guidelines for conducting State Safety Belt Use Surveys, Uniform Criteria for State Observational Surveys of Seat Belt Use (23 CFR Part 1340), are summarized below:

• Estimates must be obtained through a survey of actual observation of shoulder belt use. This means that use rate estimates must be based on physical observation of occupants in vehicles on roadways. Use rates determined from secondary sources, e.g., police crash reports or use reported through telephone surveys, are not permitted.
• The survey must be probability based. Statistical procedures must be employed to select sites at which observation of shoulder belt use are made. Following probability based sampling procedures permits estimates that are "representative" of the use rate in the desired population and makes it possible to calculate their sampling errors.
• The survey must be designed and conducted to permit estimating shoulder belt use for the following population of interest:
• Front seat, outboard passengers, i.e., the driver and right front seat passenger.
• All passenger motor vehicles, i.e., automobiles, pickup trucks, vans, minivans, and sport-utility vehicles must be observed, regardless of the state (or country) of registration.
• Observational sites in at least the largest (in terms of population) geographic areas (usually counties) in the state containing at least 85 percent of the state's population must be included in the sampling frame and have positive probabilities of selection. This criterion permits states with large, sparsely populated geographic areas to exclude the smallest (in terms of population) where few observations were expected.
• Observations must be conducted during all daylight hours and all days of the week and must be scheduled without regard to day-of-week and time-of-day (for daylight hours).
• The survey must be designed to produce an overall estimate of shoulder belt use with a relative precision (the estimated sampling error of the use divided by the estimated use rate) of not more than +/- 5 percent. This ensures that there are a sufficient number of observational sites and observed vehicles to produce a statistically reliable estimate.
• The survey design and results must be properly documented for evaluation of survey results by NHTSA and others and to determine compliance with Criteria 1 - 4 listed above.

Results

Table 1 shows the shoulder belt use rates reported by the states in 1998. Use rates for front seat outboard occupants ranged from a high of 88.6 percent in California to a low of 40 percent in North Dakota. The second highest reported use rate, 82.6 percent, was reported by three states: Maryland, New Mexico, and Oregon.

Each state estimate of shoulder belt use rate was weighted by that state's number of vehicle miles traveled for 1997, the latest year available, resulting in a computed National average use rate of 68.7 percent. This computed National average use rate compares closely to the overall shoulder belt use rate estimated in the 1998 National Occupant Protection Use Survey, 68.9 percent, which was conducted in October 1998.