All of the rear-end impact accidents occurring in the city of Rochester, New York, in a three-month period were surveyed by tabulation of the police accident reports. Special police information forms, telephone interviews, and mail questionnaires were used for further data acquisition. Vehicle photographs and medical examinations were conducted for approximately every 20th vehicle.During the data collection period, 691 rear-end impacts occurred. Although a computer program revealed 1371 accidents, defects in the program accounted for the large difference. Whiplash injury frequency based on telephone interview and mail questionnaire data obtained one to seven days after the accident revealed a whiplash injury frequency of 38%, which was approximately twice that determined by on-scene police investigators. Head restraints reduced whiplash frequency by 14% and fixed head restraints appeared to be more effective. Seventy percent of adjustable head restraints were in the downmost position. Women sustained whiplash injury more frequently (51%) but benefited from head restraints more (whiplash injury frequency 38%).An extensive review of the literature related to whiplash injury and head restraint design and effectiveness is presented.