This experiment explores the relationship between television violence and viewer enjoyment. Over 400 participants were randomly assigned to one of 15 conditions that were created by editing five TV programs into three versions each: A graphically violent version, a sanitized violent version, and a nonviolent version. After viewing, participants reported their enjoyment of the content and emotional reactions to the content. Once we controlled for viewer perceptions of action in the program, we found that the nonviolent version was significantly more enjoyable than the two violent versions were. This finding held regardless of participants' sex, level of trait aggression, and sensation-seeking tendencies. Thus, the widely held belief that violence increases enjoyment was not supported.