Research findings are based on information obtained through the taping of more than 150 hours of criminal trials held in a North Carolina court, interviews with legal professionals, and a review of trial practice manuals. Analysis revealed that witnesses who used hedging remarks such as 'I guess' and 'uh' were seen as less competent than witnesses who used a more powerful language style. The narrative style of testimony was seen as more convincing than the fragmented style; legal tacticians appear to have been correct in recommending the use of a narrative style to enhance witness credibility. Witnesses who were hypercorrect in that they misapplied certain rules of formal grammar and used a more formal and technical vocabulary than seemed necessary or appropriate were viewed less favorably by jurors. Concerning simultaneous speech indicating a struggle for testimony control, results indicate that the lawyer is seen as losing some of his control over the testimony whenever he is drawn into verbal conflict with a witness. From these studies it was concluded that the style in which testimony is delivered can have substantial effects on the social perceptions and beliefs of those hearing the testimony. A final study revealed that jury instructions were not effective in controlling such style effects. Two tables, 11 footnotes, and 38 references are provided.