This paper describes two related studies. The general hypothesis was that the rhythmic patterning of dialogue is related to how onlookers perceive certain psychological characteristics of the speakers. The experimental procedures in both studies involved evaluation by independent observers of the “credibility” of participants. In the first study 30 female college students were assigned to two conditions; each subject was either to read or to listen to monologues and dialogues (between senators and witnesses) from the “Watergate” hearings. There were significant differences in student ratings between listening and reading conditions for witnesses engaged in dialogues. In the second study 53 male and female subjects listened to tapes or read transcripts of dialogue from four senator-witness pairs. As in Study 1 subjects rated witnesses on a number of dimensions including credibility. Non-lexical qualities of speech in the taped dialogues were analyzed using a PDP-12 computer and these data were correlated with subjects' credibility ratings. The multiple regression analyses indicated that mean turn time, variability of pauses, and similarity in duration of pauses and vocalizations were significantly related to ratings of witnesses.
