Date of Award

Spring 1987

Degree Type

Thesis - Restricted

Degree Name

Master of Arts (MA)




Using a 2 x 2 factorial design, this study examined whether a source using a dialect similar to the receiver would be perceived to be more credible than a source using a dissimilar dialect. It also tested the effect of source - receiver dialect similarity on stereotyping and social attributions. Two-way analysis of variance and correlational analysis were used to test five hypotheses regarding source credibility, attitude change, recall, stereotyping and social attribution derived partly from a review of past studies in credibility, attitude change and information gain, stereotyping, speech styles and social evaluations, and social attribution. It was hypothesized that the source perceived to be similar by the. receivers would be rated as more credible than the dissimilar source, and that, attitude change in the desired direction would be greater in the group who heard a similar speaker than in the group who heard a dissimilar speaker. It was hypothesized that a greater amount of the information presented in the message would be recalled by subjects who had heard a similar speaker than by subjects who had heard a dissimilar speaker. It was also hypothesized that a favourable stereotype would be held of the similar speaker, based on the ethnic group he was perceived to belong, and that attributions for the positive behavior of the similar speaker would be internal and dispositional, whereas those for the dissimilar speaker would be external and situational. Black and White subjects heard a speaker with a Black dialect and a speaker with a White dialect. The experiment thus had four cells, with a total number of 134 subjects. An initial analysis of the results showed that none of the hypotheses proposed for the study were supported. However, when only the data of the subjects who had correctly perceived ethnic similarity with the speaker were used for analysis, it was found that the similar source was rated as being more credible than a dissimilar speaker, and that subjects held a more positive stereotype of the similar speaker (presumably due to the positive stereotype of the group he was perceived to belong) than the dissimilar speaker. There were still no significant differences in the attribution of behavior, the amount of information recalled and the amount of attitude change across the four cells. The study questions the function of source credibility. It recommends a methodological refinement in the scale measuring stereotypes. The results also suggest that the manipulation of speaker dialect was not very successful.