Objective To assess how local television news programs' reporting of injuries and deaths from traumatic causes compares with coroners' records of deaths and the estimated incidence of injuries in the same geographic area during the same time. Methods Using epidemiologic methods, we identified the underlying cause of death or injury in each of 828 local television news stories broadcast in Los Angeles during late 1996 or early 1997 that concerned recent (<3 days) traumatic injuries or deaths in Los Angeles County. Odds ratios were computed using deaths by homicide or injuries sustained in assaults as the referent group. Results The number of persons depicted as dead amounted to 47.8% of the actual total number of traumatic deaths occurring in Los Angeles County during the study period. In contrast, the number depicted as injured represented only 3.4% of injuries due to traumatic causes. Both injuries and deaths due to fires, homicides, and legal interventions were proportionally well represented. However, injuries and deaths from accidental poisoning, falls, and suicide were significantly underrepresented. Conclusions Some types of events receive disproportionately more news coverage than others. Local television news tends strongly to present only those events concerned with death or injury that are visually compelling. We discuss reasons for concern about the effect that this form of information bias has on public understanding of health issues and possible counteractions that physicians can take.