There are two primary goals with this research. First, we examine whether news media were biased in coverage of the candidates or issues during the 1996 U.S. presidential campaign, as Republican Party candidate Bob Dole and others claimed. Second, we use an ideodynamic model of media effects to examine whether the quantity of positive and negative news coverage of the candidates was related to the public's preference of either Bill Clinton or Dole. The model posits that a candidate's level of support at any time is a function of the level of previous support (as measured in recent polls) plus changes in voters' preferences due to media coverage in the interim. This model allows exploration of whether news media coverage, alone, could predict the public's presidential preference in 1996. Using a computer content analysis program, 12,215 randomly sampled newspaper stories and television transcripts were examined from forty-three major media outlets from 10 March to 6 November 1996. Findings reveal both remarkably balanced media coverage of the two principal candidates, Clinton and Dole, and a powerful relationship between media coverage and public opinion.