For successful functioning of society it is important that people become informed about public affairs. The press plays an important role in disseminating factual information related to those issues, although there may be some distortions at times. News consumers, therefore, need to closely analyze news coverage, because there is a chance that news media organizations are presenting incorrect information--either intentional or not. News consumers can make those believability evaluations based on the news content and its news source. Yet, those are not two independent judgments. The messenger and the message, together, create a specific context that people can utilize to make decisions about the believability of news content. Kelley (1972) proposed attribution analysis of persuasion to examine the role of expectancy violations on levels of believability. He contended that the message recipient cannot be sure that a source is being candid when a message confirms expectancies, but the receiver may conclude that a source has overcome his or her self-interest when expectancies are disconfirmed. Such act of perceived honesty would then lead to a higher believability. The scholarly literature that is based on Kelley's theory makes it appear that expectancy violation always leads to higher believability than expectancy confirmation, but this dissertation is challenging that notion. The argument is made that trusted news sources with expected news statements are actually highly believable, because trusted news sources are expected to provide a perspective that those news consumers agree with. Therefore, this dissertation explores the possibility of an interaction effect between news source trust and news statement expectancy, in such that when news statement expectancy increases, the news statement believability increases with an increase in news source trust for trusted sources; and when news statement expectancy increases, the news statement believability decreases with a decrease in news source for distrusted sources. This dissertation examines the believability of a news statement in the form of a headline about an alleged population increase of polar bears at the North Pole. A randomly-selected group of 1329 U.S. adults was asked during an online survey experiment to evaluate the believability of the news statement. It was unknown to the participants that the news source of the statement was manipulated. One-third was told that the headline was from CNN, another third was told it was taken from Fox News, and the remaining respondents did not receive source information. A theoretical model is tested in which variation in news statement believability is predicted by news source trust, news statement expectancy, the interaction between news source trust and news statement expectancy, political ideology, party identification, value-relevant involvement, the interaction between value-relevant involvement and political ideology, and the interaction between value-relevant involvement and party identification. An ordinary least squares (OLS) regression analysis demonstrates that news source trust, news statement expectancy, and the interaction between those variables are important predictors of variation in news statement believability. News source trust and the interaction variable played a more dominant role in explaining variance in the Fox News condition; A Chow test indicates that the coefficients for news source trust are unequal between the two conditions. Moderate effect sizes were obtained in support of the model for both news organizations, but the Fox News condition predicted statistically significant more variance in news statement believability than for CNN.
Read
