This paper presents an overview of recent estimations of safety belt effectiveness obtained using a new technique—the double-pair comparison method. By effectiveness is meant the reduction, expressed as per cent, in fatalities to a presently unbelted population that would result if all of its members were to use belts, but not otherwise change their driving behavior. The double-pair comparison method is presented more simply, with less mathematical detail and rigor, than in earlier descriptions of it. The method is applied to determine the effectiveness of three-point lap/shoulder belts in preventing fatalities to drivers and right front passengers in passenger cars of model year 1974 or later, averaged over the distribution of crashes which occur in U.S. traffic. The method is then applied to investigate the dependence of effectiveness on a variety of vehicular, accident, roadway, and environmental factors. It is reported that if all presently unbelted drivers and right front passengers were to use the provided lap/shoulder belts, but not otherwise change their behavior, fatalities to this group would decline by (43 ± 3)%. Safety belt effectiveness for drivers is higher in single-car crashes than in multiple-vehicle crashes, but does not depend much on a variety of other vehicle factors (for example, car size), accident factors (travel speed), roadway factors (posted speed limit), and environmental factors (light versus dark).