This study had limitations though, such as including children up to six years of age even though the appropriate restraint at that age would be lap-shoulder seat belt, fatal crash data that lacked a good control variable for crash severity, and fatality data with nonfatal injury information collected from police reports.