Subjects were 262 introductory psychology students at the University of Wisconsin - Madison. A video tape of the proceedings of the actual disorderly conduct trial of State of Wisconsin v. M.B. was used. "Newspaper articles" included a description of an incident in which police responded to a report that a man was arguing with his fiancee outside an apartment building. When police arrived, the defendant, who had been drinking, was noticeably upset and was heard swearing at police and telling them to leave. These articles also contained one of the pretrial publicity manipulations. Five types of pretrial publicity were used, including statements about the defendant's character, weak inadmissible statements by a neighbor of the defendant, a prior police record for the defendant, mention of the defendant's low-status job, and strong inadmissible statements by a neighbor of the defendant. In addition to these pretrial publicity conditions, some subjects saw the newspaper stories with no negative pretrial publicity; some subjects saw nothing but the trial. Findings show that pretrial publicity, particularly negative information about the defendant's character, can influence subjects' initial judgments about a defendant's guilt. This bias is weakened, but not eliminated, by the presentation of trial evidence. Character pretrial publicity and both weak and strong inadmissible statements apparently operate by changing subjects' initial judgments of the defendant's guilt. This initial judgment then affects the way subjects assess the evidence presented in the trial and the attributions they make about the defendant. Prior record pretrial publicity apparently has its effects by influencing subjects' inferences about the criminality of the defendant, and this is related to posttrial judgments. 3 tables and 30 references