|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 PEER INFLUENCE ON ATTITUDES TOWARD CABLE NEWS CHANNEL CREDIBILITY By MATTHEW BEATON A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MAST ER OF ARTS IN MASS COMMUNICATION UNIVERSITY OF FLORIDA 2010
2 2010 Matthew Beaton
3 To Mom, Daddy, Chip, and Jonathan
4 ACKNOWLEDGMENTS I would like to thank Ronald Rodgers, Ph.D., David Ostroff, Ph.D., Cynthia Morton, Ph.D., and Kathryn Gerlach for their enormous help and advice throughout this process. Your efforts brought this project to fruition.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................7 LIST OF FIGURES .........................................................................................................................8 ABSTRACT .....................................................................................................................................9 CHAPTER 1 INTRODUCTION ..................................................................................................................10 2 LITERATURE REVIEW .......................................................................................................13 Credibility ...............................................................................................................................13 Medium Credibility ................................................................................................................14 Source Credibility ...................................................................................................................19 Cable News and Credibility ....................................................................................................24 Cable News Viewership .........................................................................................................26 Credibility Polling Data ..........................................................................................................29 Party Affiliation/Political Ideology and Attitudes Tow ard Cable News Credibility ..............35 Peer Influence Studies ............................................................................................................40 Peer Influence on Individuals Attitudes ................................................................................47 Peer Influence on Attitudes Toward Media Credibility .........................................................50 Resistance to Peer Influence with Age ...................................................................................50 Theories Used in Peer Influence Studies ................................................................................51 Social Cognitive Theory .........................................................................................................53 Social Cognitive Theory in Peer Influence Studies ................................................................58 Justification for Using SCT ....................................................................................................60 Research Questions .................................................................................................................61 3 RESEARCH DESIGN AND METHODS ..............................................................................63 Data Collection .......................................................................................................................63 Instrument and Measures ........................................................................................................65 Independent V ariable .......................................................................................................65 Outcome V ariable ............................................................................................................67 Data Analysis ..........................................................................................................................69 4 RESULTS ...............................................................................................................................70 Descriptive Statistics ..............................................................................................................70 Research Question 1 ...............................................................................................................71 R esearch Question 2 ...............................................................................................................72
6 Research Question 3 and Hypothesis 1 ..................................................................................73 Linear Regression for CNN ....................................................................................................74 Linear Regression for MSNBC ..............................................................................................76 Linear Regression for Fox News ............................................................................................78 5 DISCUSSION AND CONCLUSION ....................................................................................88 Summary .................................................................................................................................88 Explanation .............................................................................................................................91 Limitations ..............................................................................................................................93 Suggestions for Future Research ............................................................................................94 Closing ....................................................................................................................................95 APPENDIX A SURVEY QUESTIONNAIRE A ...........................................................................................97 B IRB DOCUMENTATION ....................................................................................................102 LIST OF REFERENCES .............................................................................................................104 BIOGRAPHICAL SKETCH .......................................................................................................114
7 LIST OF TABLES Table page 51 Results of Linear Regressions Predicting Effect of Independent Variables on Respondents Credibility Ratings for CNN .......................................................................85 52 Results of Linear Regressions Predicting Effect of Independent Variables on Respondents Credibility Ratings for MSNBC ..................................................................86 53 Results of Linear Regressions Predicting Effect of Independent Variables on Respondents Credibility Ratings for Fox News ...............................................................87
8 LIST OF FIGURES Figure page 51 Distribution of Respondents Credibility Ratings for CNN ..............................................82 52 Distribution of Respondents Credibility Ratings for MSNBC .........................................82 53 Distribution of Respondents Credibility Ratings for Fox News .......................................83 54 Distribution of Respondents Perceptions of Peers Credibility Ratings for CNN ...........83 55 Distribution of Respondents Perceptions of Peers Credibility Ratings for MSNBC ......84 56 Distribution of Respondents Perceptions of Peers Credibility Ratings for Fox News ....84
9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts in Mass Communication PEER INFLUENCE ON ATTITUDES TOWARD CABLE NEWS CHANNEL CREDIBILITY By Matthew Beaton December 2010 Chair: Ronald Rodgers Major: Mass Communication Though much scholarly research has been devoted to media credibility, no studies have explored peers potential influence on individuals assessments of cable news channels objectivity and trustworthiness. The current study examines respondents relative credibility ratings for the three most popular cable news channels (CNN, MSNBC, and Fox News) and explores the relationship between respondents perceptions of peers credibility ratings and respondents own credibility ratings. A total of 562 questionnaires were obtained from undergraduate students attending the University of Florida (UF) Results indicate that respondents find CNN and Fox News to be most and least credible, respectively. Their percep tions of peers credibility rat ings mirror their own credibility assessments. A series of linear regressions for each of the three channels show respondents pe rceptions of peers credibility ratings to be the single best predictor of respondents own credibility ratings.
10 CHAPTER 1 INTRODUCTION Hundreds of journal articles and numerous scholarly books have explored media credibility. Additionally, the topic has received considerable attention outside academia. For example, with few exceptions, the Pew Research Center releases annual media credibility polling data. These polls cover a broad range of issues, running the gamut from comparing broadcast networks nightly news programs credibility ratings to investigating highly specific inquiries such as how critical Fox News viewers are of the news media at large (Pew Research Center, 2005; Pew Research Center, 2006; Pew Research Center, 2008; Pew Research Center, 2009). Results of these media credibility polls are far from encouraging. A recent Pew report revealed that Americans assessments of media credibility have reached an all time low (Pew Research Center, 2009). Among a nationally representative sample of American adults, only 29% agreed that the media get the facts straight, a 26% drop from when the poll was first conducted in 1985 (Pew Research Center, 2009). Across the board, Pew data have consistently shown waning credibility ratings for the U. S. news media (Pew Research Center, 2006; Pew Research Center, 2008; Pew Research Center, 2009). Evidence demonstrates that credibility ratings often suffer when the medias constructions of reality contradict the existential and social realities ( Tehranian, 2002, p. 76). These false constructions of reality directly oppose journalisms primary objective: seeking truth and providing a fair and comprehensive account of events and issues ( SPJ Code of Ethics, 1996). This disconnect between espoused tenets and actual practices should alarm those in the communications field since w ithout credibility, media lose their le gitimacy, audiences, power, and ultimately money (Tehranian, 2002, p. 72). Furthermore audience shrinkage subsequent to growing skepticism will invariably lead to a less informed public. An
11 understanding of current events is held to be a forerunner of justice and the foundation of democracy (SPJ Code of Ethics, 1996). Thus, the publics loss of faith in news media may represen t a crack in the foundation of our democratic society. In light of such concerns, the topic of media credibility is both consequential and timely. The current study explore s the role of peer influence in assessments of three cable news channels (CNN, Fox News, and MSNBC). Specifically, UF unde rgraduate students were surveyed to determine if a statistically significant relationship exists between respondents perceptions of peers credibility ratings and respondents own ratings. By querying students regarding their attitudes toward the news media, this study offered greater insight into an important segment of the population: highly educated young adults. Some research suggests that college students are particularly disenchanted with the mainstream media an d are less likely than other groups to acquire their knowledge about current events from newspapers or cable news programs (Diddi & LaRose, 2006). Furthermore, Mitchelstein and Boczkowski (2010) predict that Internet based news sources may complement trad itional media use for older, less educated, and lower socioeconomic status groups, but it may displace offline news consumption for younger and more educated users (p. 4). College students attitudes towards cable news networks are influenced by many vari ables some of which have received inadequate research attention. Despite an abundance of media credibility studies and a plethora of peer influence research, intersections between these two fields of inquiry have not been investigated. Peer influence l iterature as a whole has tended to focus on how and to what extent peoples attitudes towards drug use, body image, and risktaking behaviors as well as their decisions as consumers are affected by others (Bauman & Ennett, 1994, p. 820; Dohnt & Tiggemann, 2006, p. 929; Childers & Rao, 1992, p. 198; Gardner & Steinberg, 2005, p. 625). No studies could be located
12 concerning the role of peer influence in the formation of individuals attitudes toward news media. Not only does a gap exist in the academic litera ture, but, surprisingly, the topic has not been addressed in media credibility polls conducted by nonprofit organizations, for profit organizations, and news outlets (Pew Research Center, 2006; Pew Research Center, 2008; Pew Research Center, 2009; Public Policy Polling, 2009; Public Policy Polling, 2010). One of this studys most si gnificant contributions is the fact that it connect s two disparate (yet related) bodies of literature. Th is research project procure d data regarding respondents perceptions o f their peers credibility ratings for CNN, Fox News, and MSNBC as well as respondents personal ratings of these major cable news channels. When studying the role of peer influence, knowledge of respondents perceptions of their peers attitudes is perhap s of greater value than knowledge of their peers actual attitudes. This studys survey was designed specifically to elicit information from the respondents perspective which may or may not accurately reflect reality. Although causality cannot be establ ished, a statistically significant relationship between respondents credibility assessments and respondents perceptions of peers credibility assessments suggest s that peer influence may contribute to flagging media credibility ratings. A deeper understa nding of the process by which undergraduate students form attitudes toward cable news channels is achieved through examining the extent to which these attitudes correlate with perceptions of their peers credibility assessments.
13 CHAPTER 2 LITERATURE REVI EW Credibility Although media credibility has been separated into a few different subsets by the communication literature, it is typically parsed into two distinct categories: source credibility and medium credibility (Stroud & Lee, 2008). In general term s, credibility references the judgmental space wherein listeners, readers, viewers, etc. judge the communication and the communicator on the basis of several internally held criteria, e.g., honesty, trustworthiness, ability, and so forth (Markham, 1968, p. 58). One of the most salient and recurring factors used to define media credibility is trust. Indeed, the concepts credibility and trust overlap to such an extent that, at times, they have been used interchangeably (Kiousis, 2001; Johnson & Ka ye, 2004; Kohring & Matthes, 2007; Public Policy Polling, 2010). Kohring and Matthes (2007) assert that trust is a hierarchal factor on the second order, comprised of four first order factors which media consumers cognitively perceive and group under the umbrella of trust in news media (p. 247). These first order factors include: trust in the selectivity of topics, trust in the selectivity of facts, trust in the accuracy of depictions, and trust in journalistic assessment (p. 247). Source credibi lity is determined by the sum of all communicator characteristics that can influence the processing of a message (Kiousis, 2001, p. 382). Alternatively, medium credibility refers to the channel through which the content is delivered rather than the sender (or senders) of that content (Kiousis, 2001, p. 382). Some researchers have taken a macrolevel approach to source credibility by examining the media industry as whole, analyzing credibility across all channels (Watts, Domke, Shah, & Fan, 1999, p. 150, 151). Research has established that both individuals and groups can function as sources/communicators. It is, perhaps, more
14 intuitive to identify an individual speaker who communicates directly to the audience and gives his own views on an issue as a sp okesperson deserving of credibility assessment (Hovland, Janis, & Kelley, 1953, p. 19). Nevertheless, when organizations are segmented into their various components, two primary entities relevant to source credibility emerge: channel/institutional credibil ity and individual credibility (Watts et al., 1999, p. 150, 151; Oyedeji, 2007, p. 118). While the primary goal of source credibility research is to discover the impact that factors like presentation style, quality/sound of voice, and word choice have on the judgment of the communicators credibility (Addington, 1971; Markham, 1968), the study of medium credibility concerns determining the role publications or channels play in how messages are received and assessed. Lucas and Britt (1950) were among the fi rst scholars to note the importance of medium credibility. They maintained that advertisements credibility is tied to the media channels, magazines, and televisions networks that carry them (Lucas & Britt, 1950, p. 660). Similarly, Hovland et al. (1953) c laimed that the impact of a message probably depends upon the particular publication or channel through which it is transmitted (p. 19). Medium Credibility In 1964, Westley and Severin noted that very little attention had been given to the r elative credibility of the media which carry the message (p. 325). They attempted to address this gap in the literature by conducting a credibility experiment across multiple media channels. At the time of publication, the researchers wrote: We find no previous studies of the correlates of media credibility. We believe such knowledge will shed some light on the question of why the media are trusted or distrusted (Westley & Severin, 1964, p. 327). Their seminal study examined respondents beliefs concerni ng the credibility of television, radio, and newspapers, while accounting for key demographic variables (e.g. age, gender, income, education, urban/rural) (Westley & Severin, 1964, p. 326330). These scholars used the following question
15 to assess the concept of credibility: As between television, radio, and the newspapers, which one do you feel gives the most accurate and truthful news? (Westley & Severin, 1964, p. 326). Thus, Westley and Severin (1964) used accuracy and truthfulness as the two primary di mensions for defining media credibility and sought to discover which medium was the most credible (p. 326). Their study revealed that respondents considered television to be the most credible, while radio was rated least credible (Westley & Severin, 1964, p. 326). Findings indicated that respondents with higher levels of education considered newspapers to be more credible than television and radio (Westley & Severin, 1964). Additionally, greater education was associated with lower television credibility ratings compared to the study average (Westley & Severin, 1964). Interestingly, the best predictor of credibility was place of residence; urbanites found newspapers to be more credible while those who lived in rural areas ascribed greater credibility to t elevision (Westley & Severin, 1964, p. 330). Farmers, in particular, gave low credibility ratings to newspapers, but assigned high credibility ratings to television (Westley & Severin, 1964, p. 334). Additionally, the study revealed that the amount of time spent with a given media channel was positively related to its subsequent credibility ratings; in other words, those who use a medium more and say they prefer it as their chief news source also tend to assign it greater credibility (Westley & Severin, 1 964, p. 333). Similarly, a study conducted by Abel and Wirth (1977) comparing the credibility of local television news and local newspapers found that respondents perceived television news to be significantly more credible than newspapers. This survey of Detroit residents analyzed credibility according to three dimensions: believability, truthfulness, and importance (Abel & Wirth, 1977, p. 374). The authors concluded that television is perceived to be a more credible, truthful, and important source of lo cal news than the newspaper ( p. 375).
16 Lee (1978) provided further evidence of the higher perceived credibility of television news. In this study, college students completed a survey to ascertain what news medium they trusted most (p. 283, 284). These resp ondents in a three to one ratio r anked television news to be more credible than print news (p. 287). The researchers found that college students appeared to perceive TV news as more trustworthy, authentic, dynamic, expert, objective, intima te, convenient, easy, etc. than newspaper news (Lee, 1978, p. 287). Similarly, Ibelema and Powells (2001) findings ranked cable news as the most credible source of news (measured as trustworthiness). In fact, among the various news outlets br oadcast networks (ABC, CBS, NBC, and FOX), local television news, national newspapers (like USA Today and The New York Times), cable networks (CNN, MSNBC, Fox Ne ws), and local daily newspapers all components of television news outranked the remaining media. Cable news was rated highest in credibility, followed by local news, and then network news (p. 46). National newspapers came next in terms of credibility, then radio, and, finally, local newspapers were considered least credible (Ibelema & Powell, 2001 p. 46). The authors conjectured that cable news stands out as particularly credible because of its ubiquitous ability to provide information cons tantly ( p. 49). Greenberg (1966) defined media credibility according to the sole dimension of believabi lity; respondents were asked which medium they would trust if television, radio, and newspapers gave conflicting reports (p. 667). In Greenbergs (1966) study, television news was judged to be the most credible medium when compared to television news and newspapers (p. 667). Data revealed that gender, age, and education were significantly associated with the outcome variable. While 78% of women trusted television news more than newspapers, only 61% of men concurred (Greenberg, 1966, p. 667). With regards t o age, the researchers found that
17 younger respondents were more trusting of television news. Specifically, 72% of participants under 50 reported that television is the most credible medium; 63% of those aged 50 and older agreed (Greenberg, 1966, p. 668). E ducational attainment also predicted respondents credibility ratings; 80% of the respondents who had not finished high school trusted television news more than newspapers, but only 63% of college educated respondents trusted television news more than news papers (Greenberg, 1966 p. 668). In this study believability is synonymous with credibility. As mentioned previously, ideas such as believability and trust are frequently used to represent the concept of credibility and such terminology is considered by some to be interchangeable (Stroud & Lee, 2008, p. 1). An Edelstein and Tefft (1974) study revealed that more Americans trusted television news over newspaper and radio news during the Watergate scandal (p. 434). Overall, the literature suggests telev ision was considered the most credible medium until about the 1980s, when perceptions began to shift. By the late 1980s, television may have lost its credibility edge. Gaziano and McGrath (1986) performed a media credibility study in which respondents were called via telephone and sent self administered questionnaires. While respondents reported that they found television news to be more factual than newspapers, overall the respondents had similar attitudes toward the two media (p. 460, 461). Myriad studies have been conducted comparing television news, print news and radio news; however, very few studies have compared those traditional media outlets with the Internet. Johnson and Kaye (1998) surveyed politicallyminded Internet users, asking them to rate the credibility of online newspapers alongside traditional hard copy newspapers and online magazines with traditional hard copy magazines (p. 328, 329). Respondents rate online newspapers as more credible than their traditionally delivered coun terparts while no
18 differences exist for news magazines and issueoriented sources (p. 334). The study also reaffirmed that a mediums credibility is strongly linked to the degree to which people rely on it, and the researchers stated emphatically that indeed, in this study, reliance is linked to credibility (p. 334). Also, this study revealed a marked increase in the amount of credibility given to the Internet, showing that credibility can rapidly fluctuate (p. 335) However, Johnson and Kaye (1998) spe culated that this increase in credibility may be due to the fact that Internet users were surveyed exclusively, for they likely rely on the Internet a great deal (p. 335). In a comparison study of the 21st centurys three primary medium channels (i.e. new spapers, television, and the Internet), Kiousis (2001) found that respondents viewed newspapers as being significantly more credible than television news programs, the reverse of findings from research conducted in previous decades (p. 393). In this study, respondents credibility ratings did not vary dramatically by medium, as all three channels fell within the questionnaires moderately credible response category (p. 394). Another study, Abdulla, Garrison, Salwen, Driscoll, and Casey (2002), compared the credibility of television news, newspapers, and online news (p. 3). The researchers followed the same model which Gaz iano and McGrath (1986) created a 12 item Likert news credibility scale, with factors on the scale including trustworthiness, currenc y, bias, fairness, completeness, objectivity, honesty, up to date, believability, balance, accuracy, and timeliness (p.14). The study found that of the three mediums, users rated online news as the most credible (p. 16). This article, coupled with Johnson and Kayes (2004) Wag the Blog study, suggest that media consumers credibility evaluations for online news and journalism are becoming increasingly more positive over time. Johnson and Kaye (2004) created a credibility index based on four components believable, fair, accurate, [and] depth. Not surprisingly, this research revealed a
19 positive relationship between media usage and credibility ratings: those who used Weblogs judged them as highly credible and even sig nificantly more credible than other media (p. 634). Furthermore, almost three quarters of respondents view Weblogs as moderately to very credible, while only 3.5% of respondents rate Weblogs as not very or not at all credible ( Johnson & Kaye, 2004, p. 633). Of all news mediums incl uding televis ion, internet, radio, and print Weblogs rate as the most credible news source (p. 630). Credibi lity ratings for cable news are intermediate (lower than print news and online versions of print news, but higher than radio news and broadcast television news sources ) ( Johnson & Kaye, 2004, p. 631). Again, reliance i s a determining factor in Weblog credibility, explaining between 12.7% and 14.6% of variation in respondents ratings of Weblog credibility ( Johnson & Kaye, 2004, p. 632). Conversely, a surve y of print journalists reveals that Weblogs and other forms of online news are only thought to be moderately credible (Cassidy, 2007, p. 154). In recent years, newspapers outrank television in terms of credibility. A study by Flanagin and Metzge r (2000) reveal s that newspapers were rated significantly higher in credibility than the other media (p. 524). Simil arly, a Kiousis (2001) study shows that when respondents compare newspapers and television news sid e by side, they perceive newspapers to be significantly more credible (p. 393). Source Credibility The basi c relationship of communication between source and receiver define s a system wherein a source influences the states or actions of another system the destination or receiver by s electing among the alternative signals that can be carried in the channel connecting them (Osgood, Suci, & Tannenbaum, 1957, p. 272). One of the earliest source credibility studies, conducted by Hovland and Weiss (1951), demonstrates that trustworthine ss i s an essential dimension of source credibility; so much so that
20 they discovered changes in (viewers/readers) opinion[s] are significantly related to the trustworthiness of the source used in the communication (p. 647). A follow up study revealed a second prominent dimension of source credibility, coined expertness (Hovland et al., 1953, p. 20). Hovland et al. (1953) concluded that when acceptance is sought by using arguments in support of the advocate view, the perceived expertness and trustworthin ess of the communicator may determine the credence gi ven them ( p. 20). Hovland et al. (1953) discovered that trustworthiness was an essential ingredient to source credibility because recipients may still be inclined to reject the communication if he susp ects the communicator is motivated to make nonvalid ass ertions ( p. 21). Trustworthiness was affected by beliefs about (the sources) knowledge, intelligence, and si ncerity ( p. 20). Furthermore, perceptions of the communicators intentions to persuade h is audience may affect judgments of his cred ibility (p. 25). Factors affecting perceptions of expertness include the communicators age, whether he/she is in a position of leadership within a group, and similarities between the communicator and recipient in terms of status, values, and interests (p. 22). Additionally, when test subjects receive d information from a communicator with a high level of credibility, they generally experience d some level of opinion change based on the information communicat ed. Specifically, the greater the communicators perceived credibil ity the more opinion change was expected to occur ( p. 40). Public relations professionals are acutely aware of these findings and seek to maximize the perceived credibility of those who r elay their organizations messages. For instance, they actively seek out highly credible journalists to report on their companies news stories and recognize that advertisements are not as effective as positive press from third parties since the latter are thought to be more impartial and hence more credible ( p. 23).
21 Schweitzer and Ginsburg (1966) maintain that while Hovland et al.s (1953) dimensions of credibility (i.e. trustworthiness and expertness) are valid, such an operationalization is overl y simplistic. Schweitzer and Ginsburg (1966) wrote that it does seem clear that the Hovland, Janis, and Kelley model is incomplete (p. 99). They went on to speculate that perhaps the aforementioned factors might serve as a basic foundation on which more precise studies of source credibility could be constructed. Schweitzer and Ginsburg (1966) maintained that source credibility is extremely complex more complex than the Hovland et al. (1953) recognized and that the particular cues, or perceived charac teristics, which influence the recipients judgment of credibility will vary across communication contexts and across populations of recipients (p. 99). Building on the Hovland et al. (1953) and Schweitzer and Ginsburg (1966) studies, Whitehead (1968) pe rformed a study which used sixty five bi polar semantic differential scales to perform a factor analysis which yielded results showing competence (or professionalism) and objectivity as part of the underlying factor structure of source credibility (p. 5961). Competence was also described as having experience and a professional manner, while objectivity was described as being open minded (Whitehead, 1968, p. 63). The study also reinforced the presence of the trustworthiness factor found in previou s source credib ility studies ( p. 60, 62). Kelman and Hovland (1953) used the polar descriptions positive and negative to operationalize source credibility. Positive sources were described as trustworthy, prestigeful, or wellliked and well inform ed; conversely, negative sources were deemed untrustworthy, less prestigious, or not well liked and poorly informed; additionally, sources judgment, competence, and fairness were assessed (p. 327, 328, 330). The study, which surveyed
22 high school st udents, showed that positive sources can persuade respondents to agree with their messages while negative sources messages are often rejected (Kelman & Hovland, 1953, p. 335). The study also built on the Hovland, Lumsdaine, and Sheffield (1949) concep t of the sleeper effect in which the impact of the communication wears off with time or has a delayed effect in changing the subjects attitude (p. 188). Kelman and Hovland (1953) found statistically significant evidence which showed that the impact of sources, both negative and positive, fades, but a reintroduction of the source after a three week time lapse will reestablish the effects which the source originally made on the test subjects (p. 335). Berlo, Lemert, and Mertz (1970) built on Hovlands research and isolated three dimensions that receivers use when evaluating message sources; these dimensions of source credibility are safety, qualification, and dynamism ( p. 570). Comparing their dimensions of source credibility with Hovlands dimensions, the researchers stated that the three factor definition is not incompatible with Hovland, Janis, and Kelleys earlier conceptualization of credibility as expertise and trustworthiness (Berlo, Lemert, & Mertz, 1970, p. 574). Berlo, Lemert, and Mert z (1970) noted, though, that their analytic results provide a clarification of what is meant by those (Hovlands) terms (p. 574). They asserted that safety is a more appropriate, all inclusive dimension than trustworthiness and that qualification is e asier to interpret and more accurate than expertise (Berlo, Lemert, & Mertz, 1970, p. 575). Their third dimension dynamism can be conceived of as an intensifier, heightening the polarity of past judgments regarding safety and qualification if the source is viewed as having high dynamism (Berlo, Lemert, & Mertz, 1970, p. 575, 576). Thus, dynamism, as a dimension, reemphasizes the two previous dimensions, clearly showing whether or not the source is credible. Additionally, the authors asserted tha t the sources image in a television newscast or the sources voice on the
23 radio should only be judged subjectively according to the receivers perceptions instead of using researcher created operationalized measurements (Berlo, Lemert & Mertz, 1970, p. 573, 576). Other studies built upon the initial Hovland studies and added other dimensions to source credibility, including: competence (or reliability), trustworthiness, and dynamism (or showmanship) (e.g. Addington, 1971, p. 242; Markham, 1968, p. 6162) For a message to be effective and believable, its source must have a high level of credibility, which depends largely upon the willingness of the audience to accept its source as trustworthy and competent (Addington, 1971, p. 242). Hovland, Janis, and Kelley (1953) also made this clear by stating the effectiveness of a communication is commonly assumed to depend to a considerable extent upon who delivers it (p. 19). In a study during which subjects listened to audiorecordings of newscasts, Addington (1971) found no significant difference between the credibility ratings of male and female sources, indicating that if there is a credibility difference between male and female speakers it is unlikely that the difference is due to the differences in their voices (p. 245). Similarly, a study by Andsager (1990) revealed that subjects ascribed no more credibility to a male author of a political column than to a female author of a political column (p. 488). In the study, Andsager used two political columns, both written by males, and attached a fictitious female name to one and a fictitious male name to the other. The study allowed Andsager (1990) to study the level of gender bias held by his sample population of college students at Kansas State University (p 487). According to Andsager (1990), male and female bylines received the same rating overall (p. 488). Source credibility can also be viewed in terms of a media outlet or organization; again, Hovland, Janis, and Kelley (1953) explained that the source can be an individual, a group, or an
24 organization (p. 19). When investigating grouplevel source credibility, researchers began breaking television news into strata: local news, network/broadcast news, and cable news ( Ibelema & Powell, 2001, p. 45) Although it excluded network news, Howard, Blick, and Edwards 1987 study found that local television news was preferred to cable news as a source for specialized news news tailored to specific topics (politics, finance, entertainment, etc.), specific audien ces, or specific areas (p. 622). Decades later, respondents trust in and preference for cable news seemed to have increased, as Ibelema and Powells 2001study demonstrated that cable news earned the highest trust rating from participants in their phone survey (p. 46). The perceived trustworthiness of local news ranked second, while network news was thought least trustworthy (p.46). These data suggest that cable news networks have progressed significantly as legitimate and credible sources of information since their inception. Henke (1985) queried college students about the frequency with which they viewed CNN (the only cable news network in existence during the mid1980s). Results showed that the students used CNN as a supplement to their regular media consumption, rather than a replacement to broadcast news and newspapers (p. 434, 435). Later, in 1996, MSNBC and Fox News were launched, heightening competition in the cable news market (Mifflin, 1996, para. 1; Huff, 1996, para. 4). Cable News and Credibi lity Television news, in particular cable news, has increasingly become Americans primary news source. In 2009, a Pew Research Center poll found that when Americans were asked how they learned about national and international issues, 71% of respondents l isted television news as their main news source. And, of those who listed television as their main source, 40% tuned in to one of the three major cable news networks as their primary news outlet (Pew Research Center,
25 2009). CNN and Fox News split the lion s share of viewership 22% and 19% respe ctively while 6% of respondents indicated that they watched MSNBC (Pew Research Center, 2009). An earlier Pew Researcher Center (2008) survey found that cable news had become a widely popular news medium, second only to local news in frequency of exposure; its consumption outranked newspaper, radio, network morning news, network evening news, and online news (Pew Research Center, 2008). Of the respondents surveyed, 39% indicated that they regularly watched cabl e news, while 52% reported that they regularly watched local news (Pew Research Center, 2008). Furthermore, the poll showed that between 2002 and 2008, those who regularly watched cable news had increased from 33% to 39% while those who regularly watc hed network evening news declined from 32% to 29% (Pew Research Center, 2008). According to Pew, cable news viewership has steadily increased and surpassed network newss regular audience, making it the primary source for information on national and inter national issues (Pew Research Center, 2008). Pew research has also revealed a significant relationship between age and television news viewership, as older adults are more likely than younger adults to watch both cable and network news programs. A mere 21% of respondents aged 1829 regularly watched network news; a larger proportion (36%) indicated watching cable news with regularity (Pew Research Center, 2008). Of those in the 30 49 age group, just 22% regularly watched network news, while 38% reported watching cable news on a consistent basis. In the 50 64 age bracket, 34% and 42% regularly watched network and cable news, respectively. Among respondents 65 and older, 46% watched network news and 44% reported viewing cable news routinely (Pew Research Center, 2008). In this study, 56% of the overall cable news audience was male and 44% was female (Pew Research Center, 2008). Gender distribution was similar across the three cable news
26 channels. Specifically, 51% of CNNs viewers were male and 49% were female (Pew Research Center, 2008). At the other channels, female audience members slightly outnumbered male audience members. Fox News viewers were 48% male and 52% female, while 47% of MSNBCs viewers were male and 53% were female (Pew Research Center, 2008 ). The greater popularity of cable news relative to network news, particularly among younger adults, may be a consequence of cables twentyfour hours a day format. The Pew Research Center concluded that rather than setting aside time to watch the networ k news at a specific hour what's known as appointment television younger adults are more likely to go to cable, which is available any time they choose to tune in; this is very important because the sporadic nature of television viewing means that t his age group will be more likely to assign credibility to a channel rather than an individual news anchor/reporter or source (Pew Research Center, 2004). Another report demonstrated that less than half of respondents (48%) spent 30 minutes or more watching TV news per day, averaging 54 minutes per viewer (Pew Research Center, 2008). Cable news has allowed all viewers the opportunity to consume news at their leisure, but young people, in particular, take advantage of the flexibility that a 24hour news cycle offers (Pew Research Center, 2008). Dinner hour is no longer considered the time to get informed about current events. In 1998, 59% of 1824 year olds consumed their news during their so called dinner hour, but 10 years later, in 2008, only 39% of 18 24 year olds used the dinner hour for ne ws consumption (Pew Research Center, 2008). Cable News Viewership The three most viewed cable news channels in the 21st century have been CNN, Fox News, and MSNBC (Stroud, 2007, p. 16). Using the 2004 National Annenberg Election Survey results, Stroud (2007) found that 92% of those who claimed to watch cable news watched Fox
27 News, CNN, and MSNBC most frequently (p. 16, 17). According to the 2004 data, 34% [of respondents] reported viewing Fox News, 45% CNN, and 12% MSNBC (p. 17). Pew Research Center findings also demonstrate that Fox News is substantially more popular than competing cable news networks (Pew Research Center, 2005). The survey found that in 2004 Fox remained cable news's undisputed le ader in ratings, or the number of people watching at any given time (Pew Research Center, 2005). Fox News held about 55% of the cable news audience during the total day the aggregate ratings for all 24 hours of programming in 2004; CNN finished second with 30% of viewers for the total day, and MSNBC came in third with 15% (Pew Research Center, 2005). In primetime, Fox News was also the leader growing its median viewership by 10% 1.34 million in 2003 to 1.47 million in 2004 (Pew Research Center, 2005). In primetime, CNN held second place in median viewers, t hough it had a dropoff of 2% 815,000 in 2004 from 832,000 in 2003 (Pew Research Center, 2005). MSNBCs median primetime viewership is a distant third (but growing) with 341,000 viewers in 2004, a strong 19% increase over its 287,000 viewers in 2003 (Pew Research Center, 2005). Looking deeper into the demographics of these three cable news channels, revealed that, as of 2009, 51% of CNNs viewers are male (49% are female) (TV Dimensions, 20 09, p. 89). Twenty two percent of its adult audience falls into the 1834 age range, while 35 49 year olds make up 30% of the audience (p. 89). Fortyeight percent of CNN viewers are 50 and older (p. 89). The channels median viewer age is 49.2, while medi an household income is $66,763 (p. 89). The majority (61%) of CNNs audience has completed one or more years of college (p. 89). Fox News viewer demographics are similar. Specifically, half of Fox News viewers are male and half are female (p. 89). Persons aged 1834 constitute 21% of the Fox News audience, while
28 an additional 30% of viewers are 35 49 years old; median viewer age is 49.7 (p. 89). The Fox News audiences median household income is $62,964 (p. 89). More than half (54%) of Fox News viewers hav e completed a year or more of college (p. 89). MSNBCs audience consists of 51% males and 49% females; 23% of its viewers are between the ages of 18 34 (p. 90). Viewers aged 3549 account for 32% of its total audience, and median viewer age is 47.7 years old ( p. 90). The average household income of an MSNBC viewer is $68,752, and 64% of its audience has completed a year or more of college (p. 90). The most recent data show that cable news viewership is steadily increasing (State of the News, 2010). In 2009, cable news median viewership grew by 7%, resulting in an average of 3.88 million primetime viewers compared to 3.64 million primetime viewers in 2008 However, primetime mean viewership has recently suffered a 9% decline, falling from 4.17 million viewe rs in 2008 to 3.81 million viewers in 2009. But daytime median viewership grew by 16% in 2009, putting it at 2.16 million viewers, up from 2008s 1.86 million viewers (State of the News, 2010). Additionally, daytime mean viewership grew by 7% to 2.19 milli on, up from 2.05 million i n 2008. Between 1999 and 2009, the mean primetime viewership for the three cable networks grew from 1.26 million to 3.81 million viewers, an increase of over 300% Daytime mean viewership grew at a similar rate from .75 million to 2.19 million viewers, an increase of nearl y 300% In terms of viewership by channel, Fox News dominated CNN and MSNBC, as it averaged 2.14 million viewers in primetime during 2009, an increase of 6% over 2008s 2.02 million viewers (State of the News, 20 10). Fox News mean daytime viewership grew as well, by 17%, up to 1.2 million viewers from 2008s 1.03 viewers In 2009, CNN followed Fox News with a mean primetime viewership of 891,000, a surprising 30% percent drop from their 1.27
29 million viewers in 2008. But CNNs daytime mean viewership rose 3% in 2009, up to 645,000 over 2008s 627,000. MSNBC suffered, too, with a primetime mean viewership decline of 12% 889,000 viewers in 2008 down to 785,000 in 2009. Daytime viewership was no better; MSNBC lost 1 3% of their daytime viewers, registering 344,000 viewers in 2009 down from 398,000 viewers i n 2008. Websites corresponding to the three major cable news channels all ranked among the top ten news websites in terms of unique visitors in 2009 (State of the News, 2010) MSNBCs website ranked highest of the three cable networks, finishing second on the list of most popular news sites, behind Yahoo News, with 35.57 million unique visitors in 2009; CNN ranked fourth with 20.73 million unique visitors; Fox News ranked seventh with 12.65 million unique visitors These numbers accounted for more than 68 million total unique visitors over the course of 2009. Credibility Polling Data According to the literature, believability represents a primary component of media credibility (Greenberg, 1966; Markham, 1968; Addington, 1971; Abel & Wirth, 1977; Gaziano & McGrath, 1986). For 10 years, the Pew Research Center tracked believability in cable news between 1998 and 2008. In 2008, CNN attained higher credibility scores than all other cable or other media news channels and also outranked print, local, and network news in terms of credibility ( Pew Research Center, 2008). In 2008, 30% of those surveyed (N=928) believe(d) all or most of what CNN reported. In contrast, 24% and 23% of respondents believe(d) all or most of information relayed by MSNBC and Fox News, respecti vely Although CNN did lead all media channels in credibility in 2008, viewers faith in the network had declined substantially from 1998 when 42% of respondents believe(d) all or most of what C NN reported CNNs main competitors were not included in the 1998 poll, but in the 2000 poll, Fox News was
30 believed by 26% and MSNBC was believed by 28% of respondents, while CNN was believed by 39% of those surveyed that year (Pew Research Center, 2008). Believability ratings across different news channels vary widely according to political party affiliation. Overall, Democrats find the media more believable than do Republicans (Pew Research Center, 2008). Fox New s, is the exception, as only 19% of Democrats believed all or most of the networks reports, but 34% of Republicans were confident that all or most of the stories repor ted by Fox News were believable. This 15% gap in Republican/Democrat believability ratings for Fox News is the largest difference by political affiliation for any of the three main cable news net works CNN showed a 13% gap (Republicans at 22%; Democrats at 35%); MSNBC showed an 11% gap (Republicans at 18%; Democr ats at 29%) It should be noted that Fox News Republican/Democrat believability gap has grown during the eight year time period for which data are available (Pew Research Center, 2008). When Fox News was first rated in 2000, more Democrats believed all or most of what was repor ted than Republicans (27% compared with 26%, respecti vely) In subsequent years, the responses changed 16 percentage points, with Republicans and Democrats moving in opposite directions Meanwhile, both CNN and MSNBC showed convergence between Republican a nd Democrat responses. Between 2004 and 2008, the believability gap of Republicans and Democrats for CNN dropped from 19% down to 13% while MSNBCs gap dropped from 15% down t o 11% The Pew study also showed that the cable news networks viewerships have comparable gender make ups: Fox News (48% male, 52% female), CNN (51% male, 49% female), MSNBC (47% male, 53% female) (Pew Research Center, 2008). Cable news viewers as a group tend to be intelligent, mature, and informed. In 2008, 31% were college graduat es, 44% were 50or -
31 older, and 25% were categorized as possessing a high knowledge based on their awareness of current e vents MSNBC viewers had the highest numbers in two of three of these categories: 37% were college graduates, 41% were 50 or older, and 25% possessed high knowledge of current events. Fox News and CNN finished with very similar numbers: 25% and 32%, respectively, were college graduates, 47% of both channels viewers were 50or older, and 19% of both channels viewers possessed high knowledge of current e vents One of the most recent studies of cable news credibility was conducted by Pew in 2009. The Pew study compared party affiliation with favorable/unfavorable ratings of the three primary cable news networks CNN, Fox News, and MSN BC (Pew Research Center, 2009). Of the 1,506 adults surveyed on the phone, the gap between what registered Republicans found to be favorable and what registered Democrats found to be favorable was quite signif icant Most polarizing was the number of Democr ats who found CNN to be favorable (75%) compared to the number of Republicans who found it favorable (44%). The substantial gap of 31% suggest s that media consumers perceive CNN to be more sympathetic to the Democ ratic party Fox News, widely thought to be a divisive force in cable news, yielded a similar gap between Republicans and Democrats, but in the opposite direction. Of Republicans queried, 72% gave the cable news network a favorable rating, while only 43% of Democrats labeled it as favorable, a 29 point difference (Pew Research Center, 2009). MSNBCs numbers nearly mirrored CNNs statistics, as 34% of Republicans labeled it favorable compared to 60% of Democrats who labeled it favorable, a gap of 26 points (Pew Research Center, 2009). Meanwhile, 35% of Republicans labeled it unfavorable compared to just 7% of Democrats, a 28poin t gap. In terms of all respondents both Democrat and Republican
32 CNN was the cable news channel with the strongest polarity, with a 41% gap in fa vorable and unfavorable ratings 60% and 19% respectively. The same Pew Center survey also included two items designed to assess consumers perceptions of journalists credibility (Pew Research Center, 2009). The first question asked respondents in general, do you think news organizations get the facts straight, or do you think that their stories and reports are often inaccurate? Only 29% of respondents indicated that reporters generally get the facts straight; an astounding 63% of those sampled maintained that news stories w ere often inaccurate, while the remaining 8% were u nsure Longitudinal data reveal that consumers have become less trusting of the media over time. In 1985, 55% of respondents thought news organizations generally got the facts straight and only 34% suspect ed that stories and reports were often inaccurate Plus, 11% of those surveyed in 1985, reported being unsure Additionally, the results of a 2004 study revealed that about half (54%) of survey respondents trust some (news outlets) more than others, and 45% of the sample consider all news outlets as the same in terms of credibility (Pew Research Center, 2004). Whitehead (1968) concluded that professionalism is a vital component of credibility. Current data suggest that consumers find the media to be les s professional than in previous decades. In 2009, when the Pew Center asked respondents if news organizations were highly professional or not professional, only 59% of respondents chose highly professional, while 27% thought not professional was a more accurate descriptor. Twenty four years earlier (in 1985), 72% of respondents indicated that news organizat ions were highly professional 13% higher than the most recent statistic (Pew Research Center, 2009). The number of those answering not professional increased by 16 points during this same period, as only 11% of the sample reported viewing news organization this way in 1985 (Pew Research Center, 2009). The
33 Republican/Democrat gap for those answering not professional grew between 1985 and 2009: 11% of Republicans in 1985 and 39% in 2009 indicated a lack of professionalism in the media; the Democratic numbers for these year s are 18% and 21%, respectively. The Pew study (2009) also revealed that the public no longer trusts the media to avoid bias or to practice fairness, and independence. In 1985, 36% of respondents believed the press was careful to avoid bias, but in 2009 just 26% thought this was the case. Furthermore, 60% of those surveyed in 2009 observed bias in the press, up 15% si nce 1985 when only 45% saw bias (Pew Research Center, 2009). Political affiliation has an influence on consumers perceptions on this issue. In 1985, 49% of Republicans and 43% of Democrats believed that the media was b iased In 2009, 78% of Republicans sa w the media as biased, while 50% of Democrats indicated that this was the case. This same year, only 18% of all Pew survey respondents reported feeling that the media deals fairly with all sides (Pew Research Center, 2009). Twenty four years earlier, 34% agreed with this statement Alternatively, the number of respondents who believe the media tends to favor one side jumped from 53% in 1985 to 74% in 2009, a 21 point incre ase. Those who saw the media as independent constituted a 37% minority in 1985; the number dropped to an abysmal 20% in 2009. Furthermore, those who thought that the media was influenced by powerful people/organizations accounted for 53% of the 1985 sample, but a massive 74% of the 2009 sample Of the respondents in the 2009 survey, 40% cited cable news as their primary source for national and international news (Pew Research Center, 2009). Of those obtaining news from cable channels, the Pew 2009 study revealed correlations between their party affiliation and their regular news sources. Fox News loyal Republican viewership increased from 31% in 2003 to
34 34% in 2009. Conversely, 17% of Democrats in 2003, but only 10% of Democrats in 2009, indicated that Fox News was their main source for national and international news (Pew Res earch Center, 2009). CNN saw the reverse of this trend. The number of loyal Republicans CNN viewers dropped from 26% in 2003 down to 13% in 2009 (Pew Research Center, 2009). Their Democratic viewership decreased slightly from 32% in 2003 down to 29% in 2009 (Pew Research Center, 2009). MSNBC data on this issue are unavailable. A Public Policy Polling (2009) study, which sampled registered voters in North Carolina, found that Fox News was the most trusted cable news channel for accurate and fair reporting o f political news. The Fox network was chosen by 42% of respondents; 39% trusted CNN, and 16% trusted MSNBC (Public Policy Polling, 2009). Of the surveys self identified liberals, 85% indicated that they trusted either CNN or MSNBC the most out of the t hree cable news channels (59% and 26%, respectively) (Public Policy Polling, 2009). Of the surveys self identified conservatives, 62% chose Fox News as the most trustworthy cable news channel (Public Policy Polling, 2009). An even more recent poll, conduc ted by Public Policy Polling (2010) and released in January, reveals that Fox News was the most trusted of all television news networks, including broadcast networks. Results of the national telephone survey, which included 1,151 registered voters, show that 49% of respondents trust Fox News, while 37% do not trust the network and 15% are unsure (Public Policy Polling, 2010). Also, of the respondents queried, 39% reported trusting CNN, 41% did not trust CNN and 20% were not sure (Public Policy Polling, 2010). No questions about MSNBC were included. The poll showed a representative distribution by political party Democrats comprised 36% of the sample; 35% were Republicans; and 29% were independents (Public Policy Polling, 2010). Age had a significant influen ce on perceptions of
35 trustworthiness. The majority of adults in the 1829 age bracket trusted these cable news networks (61% trust Fox and 50% trust CNN) (Public Policy Polling, 2010). A 60 Minutes/Vanity Fair poll released May 2, 2010, revealed similar r esults regarding viewers trust in CNN and Fox News. MSNBC was not mentioned in the survey. When asked which was the most trustworthy of news sources in a list that included all broadcast news networks, the New York Times, and The Wall Street Journal, Fox News and CNN rated highest (60 Minutes/Vanity Fair Polling, 2010). Of those queried, 32% responded that CNN was the most trustworthy, while 29% rated Fox News as the most trustworthy, and the networks, as a whole, were chosen as the most trustworthy by jus t 13% of respondents (60 Minutes/Vanity Fair Polling, 2010). As expected, respondents answers were split along party lines; 46% of those who chose CNN were Democrats, 23% were Republican, and 31% were Independent, while 50% of those who chose Fox News wer e Republican, 12% were Democrat, and 38% were Independent (60 Minutes/Vanity Fair Polling, 2010). Party Affiliation/Political Ideology and Attitudes Toward Cable News Credibility Stroud and Lee (2008) analyzed the media use and credibility portion of 38,443 surveys from the Cooperative Congressional Election Study taken before the 2006 midterm election (p. 11). Their analysis unequivocally revealed that people discriminate between outlets [CNN and Fox News], and just 6% of respondents rated CNN and Fox News identically in terms of credibility (p. 15). Additionally, the study showed that those who identified themselves as conservatives or Republicans found FOX to be more credible while liberals and Democrats found CNN to be more credible (p.16). In a pr ior study, Stroud (2007) found that 64% of conservative Republicans consume[d] at least one conservative media outlet; this was contrasted by 28% of liberal Democrats who consumed at least one conservative media outlet (p. 23). Conversely, 42 percent of conservative
36 Republicans consume[d] at least one liberal outlet while 75 percent of liberal Democrats consume[d] at least one liberal outlet (p. 23). An even earlier study conducted by Lee (2005) found that those who identify themselves as conservatives and support the Republican Party trust the media less and tend to rate its credibility lower (p. 50, 51). Conversely, liberals and strong Democrats tend to trust the news media to report news fairly (Lee, 2005, p. 52). Another study conducted by Jones (2004) had similar findings: while 18.5% of strong Republicans indicate that they almost never trust the media to report news fairly, only 5.1% of staunch Democrats feel the same (Jones, 2004, p. 66). Of all Republicans strong, weak and Republicanlean ing independents 40% did not trust the media, compared to just 16.5% of Democrats (i.e. strong, weak, and Democrat leaning independents) (Jones, 2004, p. 66). A later study, which analyzed Pew Research Center data from 1998 through 2004 lends credence to these results (Morris, 2005). The Pew data revealed that between 2002 and 2004 a dramatic change and stark polarization in viewership occurred along party lines (Morris, 2005, p. 65). Democrats were more likely than Republicans to watch CNN, and R epublicans were much more likely to watch Fox News (Morris, 2005, p. 65). In the 2008 survey, 65% of conservative Republicans and 71% of liberal Democrats indicated that there were definitely some news sources which they trusted more than others (Pew Res earch Center, 2008). Nevertheless, 32% of conservative Republicans and 27% of liberal Democrats reported that they thought most news sources were similar (Pew Research Center, 2008). Pew survey results from 2008 showed that of those who viewed Fox News 39% were Republicans and 33% were Democrats (Pew Research Center, 2008). MSNBC and CNN were much more skewed in terms of political party affiliation. Only 18% of MSNBCs viewers were
37 Republican (45% Democrat) (Pew Research Center, 2008). CNN had comparable numbers: 18% Republican and 51% Democrat (Pew Research Center, 2008). Thus, overall, the data point to an obvious difference in cable news channel preferences by Republicans and Democrats. Studies conducted over the past few decades demonstrate increasing polarization in this regard. In 2004, 44% of Pew research participants who identified themselves as Democrats claimed that they were regular viewers of CNN, while only 25% of those who claimed to be Republicans watched CNN regularly (MSNBC data were not collected) (Pew Research Center, 2004). Furthermore, 41% of Republicans in this study stated that they were regular viewers of Fox News, while just 29% of Democrats tuned into Fox News consistently (Pew Research Center, 2004). A recent content analysis conducted on the three primary cable news channels showed that they focus heavily on political coverage, which maximizes the potential for perceived or actual bias: On average, 9% of cable news stories (5% of network news stories) are devoted to political topics (State of the News, 2010). In election years, as much as 56% of cable news coverage focuses on politics (State of the News, 2010). Issues affecting the United States received greater coverage on the three cable news channels compared to network news stations (State of the News, 2010). Eightytwo percent of cable news focused on domestic issues, while network news devoted 74% of their airtime to domestic issues (State of the News, 2010). Furthermore, the content analysis showed little variance in the types of stories on which the three cable news networks reported. According to study authors, the list of the top five subjects covered by each of the channels was the same (State of the News, 2010). Although the importance placed on each topic varie d, CNN, Fox
38 News and MSNBC all covered the economy, health care, the Obama administration, Afghanistan, and terrorism (State of the News, 2010). Just as striking as these findings were the Public Policy Polling (2010) data on trust in the media along pa rty and political ideology lines. Seventyfour percent of Republicans indicated that they trusted Fox News while just 30% of Democrats and 41% of Independents reported trusting Fox News (State of the News, 2010). Conversely, 15% of Republicans, 52% of Demo crats and 44% of Independents distrusted Fox News (Public Policy Polling, 2010). Not surprisingly, CNNs numbers were nearly the reverse with regard to party affiliation: while 59% of Democrats, 23% of Republicans, and 33% of Independents trusted the network, 18% of Democrats, 62% of Republicans, and 45% of Independents felt they could not trust the network (Public Policy Polling, 2010). The statistics were even more disparate when the labels conservative and liberal were used: 75% of conservatives trusted Fox News while 66% of liberals did not; 63% of liberals trusted CNN while 60% of conservatives did not (Public Policy Polling, 2010). Also, when collecting data on undergraduate student opinions throughout the country, the geographical opinion variance among the nations universities must be considered. Significant variance may occur according to which university is surveyed and where that university is located. Researchers who design studies that query young adults about their political affiliation and attitudes toward media credibility must be cognizant of these factors. Results of past presidential elections indicate that Northeastern states tend to favor Democratic candidates; in turn, Democratic voters have more positive attitudes toward CNN and MSN BC (President Map, 2008; Pew Research Center, 2008; State of the News, 2010; Public Policy Polling, 2010). Conversely, Southeastern states appear more favorable to Republicans candidates and
39 Republican voters generally regard Fox News more positively (Pres ident map, 2008; Pew Research Center, 2008; State of the News, 2010; Public Policy Polling, 2010). For example, Florida voters were split 50.9% Democratic to 48.4% Republican in the 2008 presidential election compared to voters in New York where the Democr atic presidential candidate received 62.2% of the vote and Republican candidate 36.7% (President map, 2008). Also, this Northeastern partiality toward Democratic candidates is evidenced through New Yorks representation in Congress; the state sends two Dem ocrats and no Republicans to the Senate for representation, and it sends twentysix Democrats and two Republicans to the House of Representatives (U.S. Senate, 2010). Meanwhile, Florida has one Republican senator and one Democratic senator, plus fourteen Republican representatives and ten Democratic representatives (U.S. House of Representatives, 2010). Furthermore, the Southeastern voters consistently vote Republican, particularly in presidential elections, where many Southeastern Democratic voters will cross party lines to vote for the GOP candidate (The Republican South, 2004, p. 7) Plus, Democratic majorities in the Northeast have continued to grow in recent years evidence by the 2008 election while the southern Republican majority has swelled, unabated, in recent years (Hopkins, 2009). Also, 2008 presidential election exit polling exemplifies the disparity between northern and southern states. Of those 1829, who voted in New York, 76% voted for the Democratic candidate, Barack Obama, and 21% voted for the Republican candidate John McCain (ElectionCenter 2008 Exit Polls, 2008). However, in Florida, 61% of those 1829 voted Democratic in the presidential election, and 37% voted Republican (ElectionCenter 2008 Exit Polls, 2008). Data suggests that stud ents in the Northeast are more likely to vote Democratic. Thus, they are also more likely to watch and trust cable news channels that are perceived to have
40 liberal leanings (i.e. CNN and MSNBC). Conversely, in the South, one would expect a greater number of students to vote Republican, as well as watch and trust Fox News. Peer Influence Studies Though great disparities exist in the media credibility realm, peer influence, an often studied topic, has not been analyzed juxtaposed with media credibility. However, the literature on the subject of peer influence in other areas is quite diverse. In a seminal work, Parsons (1963) defined influence as a way of having an effect on the attitudes and opinions of others through intentional (though not necessarily ra tional) action, (p. 38). He maintained that influence could be viewed as a generalized medium of persuasion or a means of persuasion (p. 44, 48). Parsons (1963) conceptualized influence as being driven by intrinsic persuaders which have a foundati onal basis in facts (p. 48). The intrinsic persuaders are attached to individuals whose level of influence is governed by their reputation (p. 50). In essence, the same statement will carry more weight if made by someone with a high reputation for c ompetence, for reliability, for good judgment, etc., (p. 50). According to Parsons, the critical common factor [for influence] is a mechanism of persuasion (p. 58). Another early study on influence looked at adolescents pre marital sexual behaviors. Re sults showed that adolescents pre marital sexual behaviors generally reflected the attitudes and opinions of those in their reference group (Mirande, 1968, p. 573). The author reported that respondents who had not engaged in coitus are far more likely to have reference groups which disapprove of intercourse with anyone and those who have experienced coitus, on the other hand, generally provide limited or unlimited approval of premarital intercourse (Mirande, 1968, p. 573). Peer influence can have a profound effect on adherence to social norms, especially as individuals reach late adolescence. Biddle, Bank, and Marlin (1980) conducted a study in which
41 149 adolescents at a public high school were interviewed about parental and peer pressure and how it affected their alcohol consumption/academic achievement. The study found that peer norms usually have more impact on older adolescents, when peers have had more time to influence their friends and when friendship ties are stronger (Biddle, Bank, & Marl in, 1980, p. 1072). Peer influence can also be seen in the consumer decision making process. Bearden and Etzel (1982) performed an initial study that examined how peers influenced an individuals purchase of public and private goods as well as luxury and necessity goods. They found that subjects relied heavily on peers for information regarding these purchases, particularly when choosing among several brands (p. 192). Childers and Rao (1992) replicated the Bearden and Etzel study. They conducted a similar survey in which respondents were queried about the level of influence peers had on their decisions to purchase specific consumer items (Childers & Rao, 1992, p. 202, 203). A total of 196 MBA graduates were questioned (p. 203). The researchers learned that peers had the largest influence on the purchase of conspicuous products items seen by the general public such as cars, boats, houses, clothing, etc., but their influence on the purchase of private products was significantly less pronounced (p. 208). Chil ders and Rao (1992) concluded that these findings demonstrate the relatively large influence of peers for public products and luxuries (p. 208). Mangleburg, Doney, and Bristol (2004) conducted a nother study on consumer purchases and peer influence which demonstrated that the effects of peer influence are substantial when shopping is a group activity. In surveying high school students, they found that peers exert social pressure to conform and even create less pleasant shopping experiences when buyers pur chase products the group deems psychologically or socially risky (p. 112). In fact, when making such
42 purchases, buyers consciously choose not to include their peers in the shopping trip (p. 112). Also, the study found peers influenced teens decision makin g when shopping through information dissemination (p. 112). In fact, information dissemination was an even more powerful influence factor than the typical pressure from peers to conform to social norms (p. 112). These findings echoed previous research in t hat they showed people are influenced more by the information that groups provide rather than by group pressures to conform (p. 112). With regard to peers influence on political opinions, evidence shows that when an issue has great importance to both the individual and the peer group, he or she is apt to be affected. Tedin (1980) found in a study of 183 recent high school graduates that as issues grew in perceived importance, peers opinions became more valuable (p. 153). After reviewing data from a surv ey of undergraduate students political attitudes, Dey (1996) asserted that p eer and faculty normative contexts appear to be strong significant influences on the development of student political orientations (p. 548). Furthermore, the author found that students with a high degree of interaction with their peers were especially likely to adopt the views of their peers (p. 549). Interestingly, interpersonal interaction was only a minor contributor to the socialization process (p. 550). Dey (1996) suggested that other subtler peer influences like general social trends and the campus media might be more important factors in the peer influence process (p. 550). In another study, Dey (1997) tracked University of California Los Angeles undergraduate st udents political attitudes (p. 401). The study revealed that students peers have a strong and consistent influence upon the development of student political orientations (p. 408). Dey added that these peer normative contexts consistently influence the political orientation of students, regardless of the social era in which students attend college (p. 409). Perhaps more noteworthy
43 is the tendency for students to change in the direction of institutional peer norms, a finding that is consistent with the socialization process; in essence, regardless of original political orientation, individuals tend to adjust their views to be more in line with those of fellow students and the institution at large after four years of school attendance (p. 409). Hallin an and Williams (1990) analyzed 20,000 friendship dyads from the High School and Beyond survey; their sample consisted of high school sophomores and seniors (p. 125). After the initial surveys, follow up surveys were conducted every two years (p.125). The researchers learned that the closer the peers friendships were and the more similarities between th e respondents and their friends in areas such as schooling and background the more likely it was that peers were able to influence one another; in essen ce, solidarity tends to lead to greater peer influence (p. 130). This was especially the case when it came to planning for the future, as closer peer relationships especially peers who shared the same school track, gender, and reciprocated desires fo r friendship were highly influential in peers college aspirations and college attendance (p. 130). Gunther and Storey (2003) found a pronounced thirdperson effect in which individuals perceive that others are more influenced by media messages and peer pressure than they are (p. 212). Ba sed on this third person effect in which one believes that their peers have been influenced by a media message Gunther and Storey (2003) found that an individual will adjust his or her attitudes and opinions to closer reflect the attitudes and opinions held by his or her peers (p. 212). Similarly, Chia (2006) proposed that adolescents develop presumptions about the medias influence on their peers, and, because adolescents are susceptible to peer norms, they align their attitudes or behaviors with these presumptions (p. 586).
44 N umerous studies demonstrate peer influence on subjects behavior and show that individuals act differently when observed by peers than when acting without their knowledge. Gardner and Stein berg (2005) found that individuals are more likely to take risks, view risks in a more positive light, and make riskier decisions when in the presence of their peers (p. 632). Dohnt and Tiggemann (2006) found that peers affect young girls body satisfaction and self esteem. The girls, who ranged in age from 58, perceived that their peers valued thinness, causing them to desire a thin appearance (p. 934). Dohnt and Tiggemann (2006) wrote girls perception of their peers desire for thinness was significant ly related to their own desire for thinness (p. 934). Oliver and Thelen (1996) also studied body image, specifically how it relates to food consumption among third and fifth grade elementary students. The researchers collected data about dietary habits, f ood intake, and body image for 142 third and 122 fifth grade students (p. 28). The study found that peer influence affected these young subjects eating and body image concerns (p. 35). As expected, girls were more likely than boys to espouse the idea th at being thin would make them more likable among their peers (p. 35). Ward (2009) examined peer influence on young adults consumption of energy drinks. The survey queried 199 Utah State University students about their energy drink consumption and the factors involved in the formation of these habits (p. 77, 79). The results were not like those one would expect with a younger sample, as the data showed that peers did not seem to influence the behavior of energy drink consumption; in fact, respondents who claimed that they did not consume energy drinks with friends drank more of these beverages, presumably alone (p. 104). Wouters, Larson, Kremers, Dagnelie, and Geenen (2010) found that simple activities like purchasing and consuming snacks and soft drinks at school are influence by ones peers. Based on data procured using surveys from 749 adolescents, ages 1217, the researchers learned that
45 i ndividual snack and soft drink consumption was high when peers proximate to the adolescent had a high consumption (p. 2, 3, 5). Peer influence effects on snack and soft drink consumption were greatest among male students and those with lower levels of education (i.e. younger students) (p. 5). Neuwirth and Frederick (2004) surveyed 397 college students at a Midwest university about peer and social influence on drinking related communication behaviors (p. 689). The study found that peers do influence college students discourse surrounding drinking activities (p. 689). They found peer influence to be a consistent predictor of college students opinions regarding excessive alcohol consumption (p. 685). Their findings indicated that peer influence is more powerful than social norms in motivating and dictating student opinion expression on binge drinking (p. 689). In fact, Neuwirth and Frederick (2004) suggested that if students were to speak out against binge drinking on campus, the campus culture could be changed and would ultimately create an environment that would discourage excessive drinking (p. 692). Another s tudy on alcohol consumption was conducted by Gaughan (2006), who studied how peer influence affects adolescent alcohol consumption by analyzing data from 2,980 best friend dyads collected by the National Longitudinal Study of Adolescent health (p. 55, 56). The study found that in same sex friendships both individuals drinking habits are equally influenced by that friend or peer, but in male female friendships only the female is influence in her drinking habits, while the male is not (p. 57). ). Also, Lars en, Engels, Souren, Granic, and Overbeek (2010) conducted an experiment dealing with beverage consumption which analyzed how individuals were influenced by their peers when consuming both alcoholic and nonalcoholic beverages. The test subjects included 70 women and 65 men ages 1828, with a mean of 21 (p. 50). The study found that when participants consumed beverages both alcoholic and non-
46 alcoholic in a social situation, they imitated confederates which the researchers had surreptitiously placed in th e study (p. 51). The authors also found that participants were more likely to imitate the confederates sip pattern when the participants were drinking alcohol (p. 51). Additionally, this study showed that men were more inclined to imitate the sips of a s ame sex partner than women (p. 50). Another prominently studied peer influenced behavior is smoking. Paek (2009) surveyed 519 student s, who were noncurrent smokers had not sm oked in the last 30 days at a S outheastern United States university, asking them about smoking and the injunctive norms associated with smoking (p. 439, 440). The study found that if peers were considered close friends, they functioned as a significant predictor of respondents own smoking intention, but if the respondent de fined peers merely as members of his or her age group, their impact on smoking intention was not significant (p. 448). Other risky behaviors have also been shown to be heavily affected by peer influence. A longitudinal survey of 714 incoming freshma n college students (279 males and 435 females) examined how previously unknown roommates affected their suitemates sexual behavior and marijuana/alcohol use after one year ( Duncan, Boisjoly, Kremer, Levy, & Eccles, 2005, p. 378). The study found that thes e first year roommates had no perceivable influence on marijuana use or sexual behavior; however, pairing up young men who binge drank in high school appears to promote binge drinking in college (Duncan et al., 2005, p. 384). No such peer influence on bi nge drinking was shown among female freshmen roommates (Duncan et al., 2005, p. 384). The study mirrored Valliants (1995) findings. In this study, first year male college students were more susceptible to peer influence with regards to alcohol consumption than their female counterparts (42.9% versus 23.8%) (Valliant, 1995, p. 402).
47 Peer Influence on Individuals Attitudes In 1967, Reiss published a foundational study regarding peer influence on individuals attitudes. The findings of this research reveale d that the best predictor of adolescents attitudes toward sex were the attitudes which they perceived their friends to hold about the subject; respondents attitudes aligned closer to what they perceived their friends thought about sexuality than their p erceptions of parents attitudes (Reiss, 1967, p. 5). In a study measuring peer influence on attitudes, behavior, and performance, researchers conducted an experiment using 39 male and 11 female subjects (mean age = 27) (Galletta, Ahuja, Hartman, Peace, and Teo, 1994, p. 233). The subjects were asked to work in groups to make calculations using computer software and subsequently complete a survey (p. 234). Though subjects were instructed to maintain silence during the computer tasks, confederates in each group would make outbursts, either in approbation of or in anger at the computer program (p. 236). The study found that subjects who were exposed to unfavorable wordof mouth statements appeared to adopt unfavorable attitudes toward the software (p. 237). On the other hand, subjects who had confederates make positive statements about the software in their presence scored about the same as control group subjects (p. 237). From this, researchers concluded that negative word of mouth comments are more pote nt than positive comments (p. 237). Eveland and Shah (2003) suggested that individuals perceptions of their peer group are often incorrect because they have, over time, engaged in a biased sampling (p. 106). Nevertheless, close knit peer groups have be en shown to influence individuals political views (MacKuen & Brown, 1987). MacKuen and Browns (1987) study on the degree to which others attitudes (particularly neighbors opinions) influence individuals attitudes revealed that social influence has sig nificant effects which begin to stand out in individual attitude changes over
48 time (p. 484). The study showed that friends in neighborhoods rather than amorphous community norms are the proximate cause of attitude changes (MacKuen & Brown, 1987, p. 484, 485). Eveland and Shah (2003) concurred, stating partisans who surround themselves with people sharing their political views may develop a distorted view of news bias (p. 106). They also found that we are more likely to talk with individuals with who m we agree, whether this is intentionalor simply structural (individuals tend to associate with those similar to themselves in age, race, and social status variables that are correlated with opinions) (Eveland & Shah, 2003, p. 106, 107). Study results caused the authors to conclude that individuals' perceptions of media bias are at least partly shaped through their interactions with others (Eveland & Shah, 2003, p. 113). Larson (1972) noted that when adolescents equally value their parents and pe ers opinions, they are distinctly less parent compliant; in other words, they are far more influenced by their peers opinions than their parents opinions (p. 72). Such a finding conforms with what is known about childhood development. In what is refer red to by psychologists as the separation individuation process, when humans move into adolescence, they both interact increasingly with peers and allow peers to have a greater impact on their attitudes and opinions, while allowing their parents less influence ( Meeus & Dekovic, 1995). A study conducted in the United Kingdom analyzed 1114 year olds views on school science and how these were impacted by their peers (Breakwell & Beardsell, 1991). The study found that both attitudes toward and performance in science class were affected by having peers who were interested in, and liked, science (the more ones peers liked science, the better a childs science grades) (Breakwell & Beardsell, 1992, p. 191). Conversely, being surrounded by
49 a peer group who did not like and did not perform well in science caused the respondent to enjoy science class less (Breakwell & Beardsell, 1992, 191). Overall, having scientific peers was associated with both a more positive science attitude and greater science activity ( Breakwell & Beardsell, 1992, 193). In a study of perceived social norms and adolescents attitudes toward sex, Chia (2006) found that perception of peer norms affects, in turn, adolescents own sexual attitudes and consequent sexual behaviors (p. 600). Chia (2006) surveyed 18to 19 year olds and found that they believed based on the third person effect that their peers were impacted by sex related television programming, and based on that belief, they altered their attitudes in order to more close ly reflect their peers (p. 600). Results from Wisdom and Agnors (2007) study showed that peers influence how individuals view symptoms of depressions and options for treating that depression. The researchers conducted in depth, 90minute interviews with 1 5 teenagers (ages 14 19) who had been diagnosed with depression or dysthymia (a mild, chronic mood disorder) (p. 337). Those who indicated that their peers helped guide them through the diagnosis and seekinghelp process also reported that these peers were very important in the recovery process by normalizing the experience of depression (p. 340, 342). Peers who considered the depressed teens thoughts and ideas about their condition important and took them seriously had a strong impact (p. 342). Depr essed individuals were, however, more likely to deny to peers their symptoms of depression or the treatment they were receiving, if they knew the peers had negative ideas about depression (p. 342). Paek and Gunther (2007) found that peers influenced the attitudes that adolescents age 1113 held toward smoking; more specifically, peers influenced these adolescents attitudes
50 about whether or not to initiate smoking (p. 414, 424). Paek and Gunther (2007) conducted their survey of 1,687 middle school students in Wisconsin (p. 414). The findings showed that although respondents were not motivated against smoking by antismoking media campaigns, they were motivated not to start smoking if they believed that their proximate peers were against smoking (p. 424) When they believed that their peers held antismoking views, the likelihood that respondents would initiate tobacco use diminished and positive thoughts, which they may have held about smoking, decreased (p. 424). Interestingly though, the study found tha t only proximal peers beliefs and attitudes were of concern to, and impacted, survey respondents (p. 424). Peer Influence on Attitudes Toward Media Credibility Kiousis (2001) found that peer interpersonal discussions tended to cause respondents to view c able news networks as less credible, and the more frequently peers discussed the news, the lower their credibility ratings became (p. 395). However, this negative association had only a modest impact on media credibility ratings (p. 395). Resistance t o Peer Influence with Age In order to gauge the age at which individuals become less influenced by their peers, Walker and Andrade (1996) conducted an experiment using 154 males, ages 317 (p. 369). The experiment consisted of a participant in a room with three confederates, during which the confederates would all give an obviously incorrect answer. This set up allowed the researchers to test whether or not the confederates answers influenced the participants responses (p. 370). The studys findings demonstrated that the pressure to conform decreases with age (p. 371). In 2007, Steinberg and Monahans study, examining resistance to conformity, revealed that resistance to peer influence waxes strongest between the ages of 14 18 (p. 1538). Data from more than 3,600 males and females between the ages of 1030 showed that resistance to peer influence stagnates in early adulthood, after which there is little evidence for growth in
51 resisting peer pressure (p. 1538). According to the researchers, this patter n is identical among males and females (p. 1533, 1538). A similar study, conducted by Sumter, Bokhorst, Steinburg, and Westenberg (2009), surveyed 464 children and young adults, ages 10 18 (p. 1012). Sumter et al.s results indicate that older adolescents have greater resistance to peer influence than younger children (p. 1016). The authors concluded that findings demonstrate a steady increase in resistance to general peer influence with age (p. 1016). Theories Used In Peer Influence Studies Numerous theories inform a vast body of literature on the effects of peer influence. It is necessary to examine these theoretical frameworks, as they guide research questions and hypotheses, to reveal scholars assumptions about how individuals attitudes and behaviors are shaped by external factors, and potentially impact study designs. When Mirande (1968) examined adolescents premarital sexual behaviors, he relied on reference group theory which proposes that the behavior of a person placed under such cross press ure will be consistent with the expectations of the group which serves as a reference point at the time (p. 572). In their research concerning the modification of consumer behavior via peer influence, Mangleburg, Doney, and Bristol (2004) drew upon soci al comparison theory. According to the authors, social comparison theory posits that people have a basic need to evaluate themselves [and] in the absence of objective standards by which to evaluate ones own attitudes and behaviors, people will use social bases of comparison as evaluative standards (Mangleburg, Doney & Bristol, p. 104). In a related study on peer pressure and how it mediates the medias influence on individuals, Chia (2006) relied on the theory of reasoned action (TRA) where a persons a ttitudes toward a behavior, jointly with subjective norms (e.g., perception of peer
52 norms) about performing the behavior, would predict that persons intention to engage in the behavior (p. 590). A study of peer influence on risk taking, conducted by Ga rdner and Steinberg (2005), used group polarization theory where group effects on risk taking depends on the risktaking tendencies of the group members (p. 626). The authors also noted that relatively conservative individuals should become even more conservative when grouped together, whereas individuals who are inclined to take risks should make even more risky choices (p. 626). Neuwirth and Frederick (2004) used the theory of planned behavior as their conceptual orientation when studying peer and soc ial influence on communication of information about drinking behaviors among college students. In a similar study of peer influence on adolescent alcohol consumption, Gaughan (2006) employed status characteristics theory, which posits that cultural belief s about an ascribed status such as sex organize the pattern of interaction in goal oriented settings (p. 47). The hostile media effect holds that media consumers believe the press to be biased against their viewpoints and has also guided peer influence research. Eveland and Shah (2003) used the hostile media effect as their organizing framework when analyzing how friends and social networks impact individuals perceptions of media bias. It is not uncommon for researchers to draw upon more than one theory within a given study. For example, Paek and Gunther (2007) utilized cognitive development theory, reference group theory, the theory of reasoned action, and the theory of planned behavior to create a conceptual schema which served to guide their research c oncerning the impact of antismoking advertisements on middle school students attitudes toward tobacco use. The study found that peer influence was only one of many factors that affected attitudes toward smoking and intentions to begin smoking.
53 Biddle, Ba nk, and Marlin (1980) also employed multiple theories in their study of peers influence on high school students alcohol consumption and academic achievement. Specifically, the authors viewed the issue of peer influence through the lens of instrumental/ro le conflict theory and socialization theory. According to an instrumental/role conflict perspective, adolescents are seen as likely to conform to parental or peer pressures when those others may observe the adolescent's behavior and are deemed likely to a pply sanctions to encourage conformity. Socialization theory speaks to the mechanism by which this occurs, as it holds that the adolescent is influenced primarily because he or she internalized the other's pressure; thus what was once the parent's (or pe er's) pressure has now become an expectation that is accepted by the adolescent for his or her own conduct (p. 1059). Social Cognitive Theory Although myriad extant theories could serve to inform research on peer influence, this study is mainly guided by social cognitive theory (SCT). This perspective was first posited by psychologist Albert Bandura and introduced to the academic community in his seminal work titled Social foundations of thought and action: A social cognitive theory (1986) This theory built upon social learning theory (SLT) a perspective Bandura outlined decades earlier (Tragesser, AloiseYoung, Swaim, 2006). In 1963, Bandura coauthored Social Learning and Personality Development with Richard Walters. In this book, Bandura and Walters (1963) enumerated a set of social learning principles that emphasize the role of social variables to a greater extent than existing learning theories (p. vii). Because social cognitive theory merely finetuned certain elements established by it s predecessor, the transition from social learning theory to social cognitive theory does not represent much of a conceptual departure. Bandura (1985) referred to SLT in terms of triadic reciprocality the theorist continued to use this terminology when discussing social cognitive
54 theory (p. 27). The three factors which comprise the triad include: the individuals behavior, his or her cognition, and the persons environment. Bandura (1985) also forwarded the notion of reciprocal determinism (p. 27). He maintained that In this model of reciprocal determinism, behavior, cognitive and other personal factors, and environmental influences all operate as interlocking determinants which affect each other bidirectionally (p. 27). Bandura (1985) asserted that behavior, cognition, and environment interact to create the sum of human learning. Each element of the triad, however, can be expected to affect individuals differently. In describing their differing levels of salience, Bandura (1985) noted that the rela tive influence exerted by the three sources of interlocking determinants will vary for different activities, different individuals, and different circumstances (p. 27). This variance can be attributed to dissimilarities in peer group composition, place of residence, educational attainment, socioeconomic status, and the nature of individuals social interactions. Bandura (1985) conceptualized the process of reciprocal determinism as occurring when pairs of these factors act on each other simultaneously (p. 26). In social cognitive theory, three dyads exist; these include the behavior cognitive, behavior environment, and cognitive environment pairings (p. 26). Bandura (1985) stated that each of these dyadic relationships is essential to the reciprocal model since through their interactions people influence situations, which in turn affect their thoughts, emotional reactions, and behavior. Behavior is an interacting determinant, not a detached by product that plays no role in transactions between persons and situations (p. 27). Behavior affects cognitive qualities in that, once a person has acted, he or she then has new information about (and, subsequently, a new appreciation of) a given situation. Additionally, behavior influences the environment sinc e individuals actions change the course of future
55 events. Thus, all three determinants are indispensable to SCT because, inexorably, a change in one determinant results in changes to the others. When Bandura (1986) created SCT, he relied heavily on social learning theorys foundations. As was the case with social learning theory, SCT is couched in terms of triadic reciprocal causation, which is used to elucidate the relationships among actions, attitudes/thoughts, and the surrounding environment (p. 12). The combination of these factors allows for human learning to occur. Under the new theory, Bandura (1986) continued to emphasize that people engage in transactional interactions at both the individual and societal levels; he elaborated on SLT by identifyi ng personal determinants, behavioral determinants, and environmental determinants as the three factors which govern these interactions (p. 12). Of the three determinants, cognitive, was the only element to change significantly as social learning theory e volved into social cognitive theory; it was renamed personal. Under SCT, the personal determinant is represented by internal personal factors in the form of cognitive, affective, and biological events (Bandura, 2001, p. 14). The environmental det erm inant refers to surroundings peers, family, society, climate, geography, etc. Bandura (2001) stated that environment is not a monolithic entity (p. 15). It is made up of three subsets: imposed environment, selected environment, and constructed environment (p. 15). The third (and final) determinant, behavior, is a composite of actions encompassed by choices and instinctive needs. In summary, Bandura (1986) based social cognitive theory on three causal relationships, which all operate as interacting de terminants that influence one another bidirectionally and thereby account for changes in individuals behaviors, environments, and personal attitudes/opinions (p. 15). Like any valuable theoretical orientation, social cognitive theory clarifies a complex phenomenon. Its parsimonious nature makes it a particularly
56 appealing framework for understanding the process of peer influence: the environmental determinant accounts for peer influence, the personal determinant reveals the cognitive effects of peer influ ence, and the behavioral determinant demonstrates individuals responses to peer influence. Another component of SCT is personal agency. According to SCT, personal agency and social structure operate as codeterminants in an integrated causal structure r ather than as a disembodied duality (Bandura, 2001, p. 266). This means social cognitive theory predicts that individuals through their interactions with each other both influence society and are influenced by society. Bandura (1989) asserted that be cause of the bidirectionality of influence, people are both products and producers of their environment (p. 362) It is also important to note that in SCT none of the three determinants is superior in terms of strength or influence (Bandura, 1986). Bandura (1986) was emphatic that factor salience varies among individuals according to their previous and current experiences. Therefore, at any one time, an individual may be influenced more strongly by environmental determinants than by personal or behavioral determinants, while at another time he or she may be more influenced by personal determinants: this varies largely by person and situation (Bandura, 1986). Also, under Banduras SCT one persons behavior could influence the social environment, which in t urn can influence the behavior of others in that environment (Ward, 2009, p. 7). Additionally, SCT posits that human beings are constantly analyzing their decisions, their beliefs, and their views to determine their level of accuracy (Bandura, 2001, p. 271). However, individuals are often forced to evaluate the validity of their beliefs in the absence of objective standards. Therefore, social verification is frequently employed, wherein people assess the soundness of their views by checking them against what others believe (Bandura, 2001, p. 269).
57 Social cognitive theory speaks to this process of social verification and has important implications for research on how peer influence might alter perceptions of media credibility. For example, news media consumers who feel unequipped to properly judge the credibility of cable news networks are apt to rely on their family and friends opinions on this issue. The current study seeks to address the role of peer influence in such assessments through the lens of social cognitive theory. Obviously, peers make up a significant portion of the environmental determinant of the triad, but the valuable aspect of social cognitive theory is that it factors in the knowledge and awareness that a person has already been equi pped with through his or her personal determinants. Additionally, any past behavior initiated by an individual regarding an issue has resulted in learning based on personal determinants (how the person thinks and feels after having acted on that issue) and environmental determinants (how the environment surrounding the issue has changed since he or she acted). Vicarious learning is another one of SCTs tenets. Bandura (2001) stated that virtually all behavioral, cognitive, and affective learning from dire ct experience can be achieved vicariously by observing peoples actions and its consequences for them. Much social learning occurs either designedly or unintentionally from models in ones immediate environment (p. 270271). However, this is not to imply that all learning can be attributed to observation of environmental norms. Bandura (2001a) was adamant that recognition of direct experience is also vitally important to our understanding of human learning. Bandura (2001) maintained that individuals pers onal determinants play an essential role in the learning process via agentic features (p. 6). The term agentic refers to all personal and behavioral determinants which enable people to play a part in their self development, adaptation, and self renewa l with changing times (Bandura, 2001, p. 2). These agentic features
58 include intentionality, forethought, self reactiveness, and self reflectiveness (Bandura, 2001, p. 610). Bandura (2001a) does not discount a human beings ability to analyze info rmation based on past behavioral choices and current environmental information, as social cognitive theory accounts for the formation of rational attitudes and opinions. Additionally, SCT explicates that a rational, introspective process occurs when person al and behavioral determinants interact, which allows for learning to take place. Social Cognitive Theory in Peer Influence Studies Past research provides ample justification for use of social cognitive theory (and its predecessor: social learning theory) in studies of peer influence. Wouters, Larsen, Kremers, Dagnelie, and Geenen (2010) utilized social learning theory when conducting their research on peer influences impact on snack and soft drink consumption. Their research was informed by a Bandura stu dy published in 1986 which asserted that peers may influence each other when they observe, model, and imitate the behavior of persons whom they deem important in some way. Tragesser, Aloise Young, and Swaim (2006) drew upon SCT when they examined peer in fluence and pre adolescents attitudes toward smokers. The study queried 292 fourth through sixth graders about their susceptibility to peer influence and personal images of smokers. The researchers utilized SCT, which posits that peers offer standards of comparison and a model of values and appropriate behaviors. Peers can intentionally or unintentionally shape individuals beliefs and expectations about participating in certain behaviors through the process of behavior modeling and observational learning (Bandura 1992, 2004). Tragesser, Aloise Young, and Swaim (2006) found it appropriate to use SCT because peers are highly influential models in early and preadolescence (p. 312). In a related study using SCT, Kinard and Webster (2010) researched the eff ects of advertising, peer and parental influence, and self efficacy on adolescents alcohol consumption
59 and tobacco use. The authors generated two separate questionnaires to survey 190 respondents with an average age of sixteen. In describing why they chos e SCT as a theoretical framework, Kinard and Webster (2010) stated that this theory implies that research aimed at understanding any type of behavior should view possible predictors as interwoven, rather than independent, sources of influence (p. 25). In this study, SCT dictated that advertising and peer/parental influence be considered environmental determinants, while self efficacy was labeled a personal determinant. The results of Kinard and Websters (2010) study revealed peer influence to be the str ongest predictor of adolescent tobacco use and alcohol consumption (p.39). Denham (2008) applied SCT to research on peer influence among high school seniors (N=2542) and their perceptions of the risks associated with anabolic androgenic steroids use (p. 9). Denham (2008) justified his use of SCT because personal determinants, behavioral determinants, and environmental determinants all impact the learning involved in adolescents risk perceptions on anabolic androgenic steroids. The study found that soci alizing with friends showed significance as a determinant of AAS risk perceptions, with those who spent more time with their peers estimating lower levels of risk (p. 22). Denham concluded that peer groups are important to the formation of risk pe rceptions. (p. 22). Ward (2009) relied on SCT when studying how an adolescents environment, particularly peer influence, and an adolescents knowledge and attitudes about energy drinks affected his or her behavior of energy drink consumption (p. 7). W ard (2009) defended the use of SCT as the theoretical basis for a peer influence study because it recognizes that behavior is multidimensional and therefore can be used to understand the influence of an individuals knowledge and attitudes as well as th eir social environment on their behavior (p. 7)
60 Justification for Using SCT Bandura (2001) stated: Much social learning occurs either designedly or unintentionally from models in ones immediate environment (p. 271). Therefore, unlike learning by doing, which requires altering the actions of each individual through repeated trial and error experiences, in observational learning a single model can transmit new ways of thinking and behaving simultaneously to countless people in widely dispersed locales ( Bandura, 2001, p. 271). In essence, a large portion of human learning involves observing and modeling socially acceptable attitudes and behaviors since the proven skills and established customs of a culture may be adopted in essentially the same form as t hey are exemplified because of their high functional value (Bandura, 2001, p. 275). Social cognitive theory recognizes the importance of social/peer pressures to conform: People share information, give meaning by mutual feedback to the information they exchange, gain understanding of each others views, and influence each other (Bandura, 2001, p. 291). In this process, SCT accounts for individuals tendency to analyze their thoughts and beliefs against those of others to determine their validity and fu nctional value (p. 269). This is especially the case when individuals do not have enough personal experience to confirm the validity of their positions. They will then use social verification to evaluate the soundness of their views by checking them agai nst what others believe (Bandura, 2001, p. 269). Of course, social verification does not preclude other avenues by which individuals may confirm validity, as people will often substantiate personal opinions/attitudes via observation and/or logical reasoni ng (Bandura, 2001, p. 269). In sum, social cognitive theorys structural framework and its recognition of peer influence situate it as the most appropriate and applicable theory for th is study regarding peer influences impact on individuals perceptions of cable news channels credibility. It wa s
61 expected that nearly all study respondents (even those who have had limited exposure to one or more of the cable news outlets) w ould report having opinions about the credibility of the three cable news channels. Social cognitive theory predict ed that, in the absence of personal experience, individuals w ould base their credibility ratings on how believable they think their peers find the channels in question. Furthermore, SCT also account ed for those who seek conf irmation regarding the validity of their existing beliefs without explicitly discussing the matter with their peers. This process of vicarious verification occurs when an individual observes his or her peers opinions on and reactions to various news out lets and then allows such observation to influence his or her own cable news credibility ratings. Furthermore, the literature shows that though media as an aggregate has seen its credibility fall in the last 25 years, cable news has emerged in that time with CNN being consistently rated as the most credible (Pew Research Center, 2004 ; Pew Research Center, 2008 ) During this timeframe, MSNBC and Fox News came of age. Barring any dramatic changes in content or philosophy, it seems likely that CNN will conti nue to be rated as the most credible cable news channel. Research Questions The purpose of this study wa s to explore traditionalaged, undergraduate students attitudes regarding media credibility. In particular, the role of peer influence on responden ts credibility ratings w as examined. Because academic investigation into this specific topic wa s lacking, an analysis of resultant data w as used to address a significant gap in the literature. The following research questions served as a guide: RQ1: Whic h cable news channel (CNN, Fox News, or MSNBC) emerges as being the most/leastcredible? RQ2: According to respondents, w hich of these cable news channels do their peers findmost/least credible?
62 RQ3: What is the relationship between respondents perceptions of peers credibility ratings and respondents own credibility ratings? Based on a review of the litera ture, the following hypothesis wa s proposed: H1 : There will be a significant, positive relationship between respondents perceptions of peers credibili ty ratings and respondents own credibility ratings.
63 CHAPTER 4 RESEARCH DESIGN AND METHODS Data Collection Data for this crosssectional study w as obtained via self administered, closedended questionnaires, distributed to UF undergraduates These stude nts were recruited from eight classes covering six disciplines; each class averaged 70 students. The classes were: Architectural History (ARC1701), History of Construction (BCN3012), a philosophy course titled: What is the Good Life? (IDS2935), Vertebrate Zoology (ZOO2303C), Integrated Principles of Biology (BSC2010), Biology for Engineers (ABE2062), Principles of Sociology (SYG2000), and Social Problems (SYG2010). Although the nonprobabilistic sample obtained for this study cannot represent all UF undergraduate students in the way a random sample would, an effort was made to survey from a wide range of disciplines to minimize selection bias. The researcher request ed that professors allow the questionnaire to be distributed during class time in the first w eek of the fall semester (August 2327, 2010) and all data was collected during the first two weeks of the semester. Conducting the study early in the term wa s preferable, as research has shown that college students absenteeism increases and their willin gness to participate in nongraded classroom activities decreases as the semester progresses (Marburger, 2001; Sharma, Mendez, OByrne, 2005). As a gesture of appreciation for completing the survey, the researcher offered to all professors and arranged w ith one for students to receive a small amount of extra credit ( one point added to their final course grades). In accordance with Institutional Review Board (IRB) guideline s, an alternative assignment was made available for those who prefer red not to parti cipa te in the study. When surveys wer e distributed ( at the beginning and end of class), the
64 researcher notif ied the students that participation wa s completely voluntary. Informat ion regarding extra credit was announced at the time the survey was administe r ed (and was also explained in the informed consent form). Recruitment continue d until eight classes were surveyed, yielding 562 completed questionnaires. This sample size wa s large enough to conduct all appropriate statistical analyses. Additionally, pa st mass media studies, which serve as a guide to inform this research, frequently rely on N=400 as a baseline minimum when conducting survey research (Miller, Goldenberg, & Erbring, 1979; Heong, Escalada, Huan, & Mai, 1998; Allen, Richard L., & Clarke, David E., 1980). Thus, a sufficient quantity of surveys was obtained. Although there is no perfect data collection method, research with undergraduate student samples is well established in academic literature (Bernard, 2000, p. 278). Furthermore, Bernard ( 2000) maintains that the use of self administered questionnaires is ideal for studying certain populations and researchers can expect a response rate of at least 70% when surveying literate respondents (p. 278). Using a 70% response rate as an estimate, the researcher planned to contact at least 572 students in order to procure 400 surveys Ob viously, higher response rates we re preferable (and were achieved) as they minimize d the potential for nonresponse bias (Bailey, 2007). The problem with low response rates is that subjects who agree to participate may differ substantially from those who do not (Bryman, 1989, p. 112). Unfortunately, because it is impossible to know how potential study participants would have answered if they had been recruited, re searchers cannot measure non response bias directly. One can, however, compare respondent demographics to existing information about the target population as a whole. In this case, it will be possible to assess whether the sample differs from UF undergradu ates in general on certain demographic factors (e.g. race, gender, ethnicity) using data provided by the
65 University of Floridas Office of Institutional Planning and Research. If undergraduates overall and those who participate in this study are comparable on several different variables, this may indicate that the data gathered is at least somewhat representative of UFs undergraduate population at large. Although it is unnecessary to calculate a minimum sample size for studies that involve nonrandom recr uitment, such calculations can still help to inform a researchers target sample size. In this case, 380 completed surveys would have been needed to represent all 31,415 UF undergraduate students between the ages of 1825 with a 95% confidence level and a margin of error equaling plus or minus 5% (Office of Institutional Planning and Research, 2010). As stated previously, the goal for the study wa s to obtain at least 400 completed questionnaires and thereby ensure an adequate sample for statistical analyses. This figure also approximate d the number that would have been needed if a random sampling technique had been utilized for the purpose of generalizing findings to the entire undergraduate student body. Instrument and Measures Independent Variables T he key independent variables in this study are those which assess respondents perceptions of how peers rate cable news channels credibility (CNN, Fox News, MSNBC). Throughout the literature, researchers have mainly defined the term peer in one of two w ays. Some references to peers are rather broad and include co workers, friends, neighbors, classmates, affiliates from professional organizations and members of business/social clubs, while the narrowest conception of the term refers only to best friend s of the same sex ( Childers & Rao, 1992; Billy, Rodgers, & Udry, 1984, p. 659). For the purpose of this study, peers are defined generally as friends. Some previous research has differentiated among different types of friendship (e.g. best friends, clos e friends, and casual friends), but factoring these relationship
66 nuances into the current projects questionnaire would have ma de the instrument overly long and unwieldy (Antonucci 2001; La Gaipa 1977). On t he survey itself (see Appendix A ), instructions c larif ied that the term friends refers to best, close, and casual friends. Key independent variables mirror ed the dependent variables designed to assess media cr edibility except that they queried respondents about their friends opinions. For these items, respondents indicate d their perceptions of their friends beliefs regarding the major cable news channels (i.e. CNN, MSNBC, and Fox News credibi lity ratings). Respondents conveyed the extent to which they agree d with statements nearly identical to those that assessed their own views on media credibility, except in these items the word I was replaced by my friends. The range of answer choices span ned the same 5 point range in which 1=Strongly Disagree, 2=Disagree, 3=Neutral, 4=Agree, and 5=Strongly Agree As with the outcome variables, credibility composite s were also developed after data collection. In addition to the key independent variables which measure d how respondents rate their peers judgments about cable news channels credibility, the s urvey include d several demographic questions. Although demographic characteristics we re not of primary importance for th is research project, studies have show n that news consumers backgrounds can influence their media credibility ratings (Gun ther, 1992; K iousis, 2001). It wa s, therefore, important to hold these factors constant by controlling for them in linear regressions. These demographic variables include d gender, age, race, and ethnicity. Additionally, the researcher account ed for: a respondents poli tical party affiliation; the frequency with which he/she discussed news/current events with friends; the respondents self assessment of how well informed he/she wa s about current events; the frequency with which the respondent watched the three cable news channels
67 (CNN, MSNBC, Fox News ), and the level of value the respondent ascribe d to his/her friends opinions. Outcome Variables Dependent variables in this study w ere composites of multiple items designed to measure respondents perceptions of credibility for the three major cable news channels (i.e CNN, Fox News, and MSNBC). Only CNN, MSNBC, and Fox News we re examined in the study because they dominate cable news viewership. Stroud (2007) found that 92% of those who claimed to watch cable news watched CNN, MSNBC, and Fox News most frequently (p. 16, 17). The data, released in 2004, revealed that 34% [of respondents] watched Fox News, 45% watched CNN, and 12% watched MSNBC (p. 17). By examining CNN, Fox News, and MSNBC, the research scope was limited to a manageable number of outlets o n which statistical analyses could be run. The following statements appear on the survey questionnaire: 1.) I believe all or most of the information presented on CNN/Fox News/MSNBC. 2.) There is often biased reporting on CNN/Fox News/MSNBC. 3.) An effort is made to present all sides of the issues on CNN/Fox News/MSNBC. 4.) I distrust the reporting on CNN/Fox News/MSNBC. Respondents were instructed to indicate the extent to which they agree or disagree with these statements. A nswer choices we re measured on a 5 point Likert scale where 1=Strongly Disagree, 2=Disagree, 3=Neutral, 4=Agree, and 5=Strongly Agree. The two negative i tems (questions 2 and 4) were reverse coded before data analysis such that higher numbered scores (e.g. 5=Strongly A gree) indicate d that respondents perceived these news outlets to be less credible. Before the outcome variables we re finalized, Cronbachs alpha scores were determined for these items, and factor analyses were run to confirm that they loaded on only one component. A ll had an alpha greater than .70 which was considered adequate because the factor analyses also support ed the idea that these items measured a single concept in this instance credibility (Nunnally, 1978). These
68 statistics provided sufficient information to determine that all survey questions merit ed inclusion. Because three separate linear regressions were run in SPSS, credibility composite s for each of the three channels w ere made to be identical so that comparisons c ould be made. As mentioned previously, four types of questions were included in the final credibility composite s. Each of these questions appear ed multiple times, as respondents were queried separately about CNN, MSNBC, and Fox News The item that read I believe al l or most of the information pr esented on [cable channel X] was one of two statements which gauge d respondents trust in these cable news channels. The other item related to trustworthiness read I distrust the reporting on [cable channel X]. The two rem aining questionnaire statements assessed respondents perceptions of cable news channels objectivity. As was the case with the trust questions, one objectivity item wa s worded positively, while the other wa s worded negatively. These statements read as fol lows: I find that an effort is made to present all sides of the issues reported on by [cable channel X] and I find that there is often biased reporting on [cable channel X]. In summary, the three dependent variables are composites of the four questions designed to measure respondents views on media credibility for the three pr imary cable news channels (CNN, MSNBC, Fox News). Each of the items included in the final composite s was weighted equally. Because all potential items were included, the final c omposite s are averages of four questions that relate to trustworthiness and objectivity Although researchers have proposed that the concept of media credibility involves various elements including competence, knowledgeability, professionalism, believabi lity, accuracy, expertness, etc. trustworthiness and objectivity carry special importance, as they are mentioned with the greatest frequency (Hovland, Janis, & Kelley, 1953; Westley & Severin, 1964; Whitehead, 1968; Kiousis, 2001).
69 Data Analysis All an alyses were conducted using SPSS (Statistical Package for the Social Sciences), a computer program used to analyze and graphically represent data. Three nested linear regression models were used to test the effect of the key independent variables (i.e. those measuring respondents assessments of their peers judgments about cable news channels credibility) on the dependent variables (respondents own credibility ratings) while controlling for a host of background characteristics.
70 CHAPTER 4 RESULTS Descri ptive Statistics Data for the current study was obtained by distributing surveys in eight undergraduate level classes, yielding a total of 562 completed questionnaires. Humanities courses, natural science courses and social science courses were surveyed. Specifically, questionnaires were administered in Architectural History (ARC1701; N=94), History of Construction (BCN3012; N=29), a philosophy course titled What is the Good Life? (IDS2935; N=144), Vertebrate Zoology (ZOO2303C; N=57), Integrated Principles of Biology (BSC2010; N=118), Biology for Engineers (ABE2062; N=39), Principles of Sociology (SYG2000; N=43), and Social Problems (SYG2010; N=38). The samples gender distribution (58% female, 42% male) is similar to university generated s tatistics for total undergraduate enrollment in 2009 (55% female, 45% male) (Office of Institutional Planning and Research, 2010) In this sample, 347 respondents (62%) labeled themselves as White/European, 66 respondents (12%) identified as Black/AfricanAmerican, 54 respondents (9%) reported being Asian/ Pacific Islander and 91 respondents (16%) indicated belonging to the Other" category which includes Hispanic, American Indian, Alaskan Native, and mixed race. Four surveytakers did not answer the race question. These numbers align closely with the schools race statistics for 2009: 60% White/European, 10% Black/AfricanAmerican, 9% Asian, and 18% Other (including Hispanic and American Indian), with 3% choosing not to report their racial backgrounds (Office of Institutional Planning and Research, 2010) Slight incongruities between this studys sample and existing data for the schools entire undergraduate population may be due to changes over time as the current survey was conducted in 2010, not 2009 and/or the nonrandom selection of study participants.
71 With regards to political affiliation, 139 (25%) labeled themselves Republican, 181 (32%) labeled themselves Democrat, 229 (41%) claimed No Political Affiliation, 9 (2%) chose Other, and 4 st udents (.7%) did not respond. A total of 559 participants (99.5%) answered the age question, which yielded a mean of 19.03 years old. The youngest respondent was 18 and the oldest was 30. Eighteenyear olds accounted for 51.42% of those surveyed, and the m ajority of respondents (73%) were nineteen or younger. When study participants were asked about the frequency with which they watched the three cable news channels, 154 (27%) claimed never to have watched CNN, 248 (44%) claimed never to have watched MSNBC, and 199 (35%) claimed never to have watched Fox News. Research Question 1 The first research question asked which cable news channel (CNN, MSNBC or Fox News ) would emerge as being the most/least credible. In order, to assess survey respondents perceptio ns of CNN, MSNBC, and Fox News credibility, three separate composite s were created. Each composite is composed of the four survey questions that measure respondents opinions about the channels trustworthiness and objectivity. These same four questions w ere asked about CNN, MSNBC, and Fox News so that comparisons could be made across the various channels. Factor analyses confirmed that all four questions loaded on the same component for each of the three composite s, and Cronbachs alpha scores were all gr eater than .70 the pre established cutoff. It was necessary to reverse code two items before composite construction. In the final variables, higher scores indicate that respondents view a given channel as more credible. R espondents credibility ratings f or CNN ( see Figure 5 1) revealed a composite mean of 3.13 (and SD=.647). For MSNBC (see Figure 5 2), the composite mean was 3.03 (and SD=.498). For Fox News (see Figure 5 3), the composite mean equaled 2.66 (and SD=.858). Therefore,
72 respondents rated CNN t he most credible (mean=3.13), while Fox News was viewed as the least credible (mean=2.66). All three histograms (see Figures 5 1, 5 2, and 53) approximate normally distributed bell curves, though each exhibits a higher than expected peak in the number of neutral responses (i.e. the Neither Agree Nor Disagree category). According to this global measure of credibility, 151 respondents (27%) reported neither a positive nor a negative attitude towards CNN. A similar number of respondents (147 or 26.4%) indic ated a neutral opinion of Fox News. An even larger proportion of study participants espoused dispassionate attitudes toward MSNBCs credibility, as 262 students (47.2% of the sample) answers reflected that they Neither Agree Nor Disagree that the ch annel is objective and trustworthy. Standard deviations for the three channels credibility scores (MSNBC = 0.498; CNN = 0.647; Fox News =0.858) indicate that the distribution of answers remains closest to the mean (i.e. neutrality) for MSNBC and that the data is spread out over the largest range of values for Fox News, suggesting that study participants feel less strongly about the former compared to the latter. Research Question 2 The second research question was concerned with study participants pee rs relative credibility ratings for the three channels (as reported by the respondents). The questionnaire asked students to indicate how credible their peers find CNN, MSNBC, and Fox News. As was the case with respondents personal credibili ty ratings, t hree identical composit es were constructed to assess respondents perceptions of peers credibility ratings for each channel. Items composing these final variables exactly mat ched those in the original composite s except these questions instructed responden ts to analyze their peers attitudes toward all three channels. Once again, Cronbachs alpha scores and factor analyses confirmed the appropriateness of using all four trustworthiness and objectivity questions in the final composites Descriptive statistics run in SPSS revealed that CNNs mean credibility score was 3.11 with a standard deviation of
73 .504 (see Figure 54), MSNBCs mean equaled 3.04 with a standard deviation of .405 (see Figure 55), and Fox News mean was 2.77 with a standard deviation of .668 (see Figure 5 6). Therefore, respondents perceived that their peers find CNN and Fox News to be the most and least credible, respectively. As was the case with respondents personal credibility ratings, each of the histograms (see Figures 5 1, 5 2, and 53) appear normally distributed. Again, more responses fell into the Neither Agree Nor Disagree category than would be expected. In fact, an even greater percentage of respondents indicated a three (i.e. neutrality) when it came to their assessments of p eers credibility ratings. A total of 272 respondents (49% of the sample) thought their friends had neither positive nor negative attitudes toward CNN, 237 (42.9%) indicated this for Fox News, and 362 (65.6%) of study participants reported that peers were completely neutral about MSNBCs trustworthiness and objectivity. A comparison of standard deviations shows that the range of responses varies most from the mean for Fox News and least for MSNBC. This suggests that respondents think their peers have the s trongest attitudes (positive or negative) toward Fox News and that their opinions toward MSNBC are less extreme. Research Question 3 and Hypothesis 1 The hypothesis and final research question pertain to the relationship between respondents persona l credibility ratings and their assessments of peers credibility ratings for each of the three channels. It was predicted that the data would reveal a significant, positive relationship between respondents perceptions of peers credibility ratings and re spondents own credibility ratings Three linear regression tables were created to address this facet of the research project. In the initial models, before controlling for other factors, there is positive association between the key independent variable ( i.e. respondents perceptions of peers credibility ratings) and the dependent variable (i.e. respondents own credibility ratings) for all three cable news channels (CNN, MSNBC, and Fox News), which in all cases is significant at
74 the .001 level. This significant relationship is maintained across all five models in each of the three channels (see Table 5 1, Table 5 2, and Table 53). The hypothesis is, therefore, supported, leading to the rejection of the null hypothesis (i.e. that no relationship exist s between the dependent and the key independent variable). The following sections describe how controlling for demographic factors, media engagement, susceptibility to peer influence, and political affiliation affects the relationship between peer and personal credibility ratings. Linear Regression for CNN The first model of CNNs linear regression (see Table 5 1) shows a strong significant relationship between respondents credibility ratings for CNN and perceptions of their peers ratings for CNN (p value <.001). Peers credibility rating alone explains 39% of the variation in the dependent variable (adjusted R2 = .39). The standardized coefficient is .63 higher than the standardized coefficients for MSNBC and Fox News indicating that the key i ndependent variable is a better predictor of respondents personal assessments of CNNs credibility than it is of Fox News and MSNBCs credibility. When demographic information is taken into account in the second model the only significant relationship to emerge is a difference between White/Europeans ratings for CNN and Asian respondents credibility ratings. Specifically, Asians rate CNN as being more credible than the White/European reference group (p<.05). The relationship between respondents credibility rating and perception of peers credibility rating remains strongly significant at the .001 level and the coefficient is relatively unchanged. The introduction of age, race, and gender explains only a small amount of the total variance. Taking demographic information into account only increases adjusted R2 up from .39 to .40 and lowers the key independent variables standardized coefficient from .63 to .62.
75 In Model 3, media engagement variables are introduced. There is a positive, significant relationship between watching CNN and finding it credible (p<.001 with a standardized coefficient of .20). Also, after controlling for media engagement, a significant relationship (p<.05) between age and credibility is uncovered. Its negative standardized coefficient ( .07) demonstrates that older respondents find CNN to be less credible. This is the only regression among the three channels in which age is statistically significant. The peers credibility rating variable for CNN remained significant at th e .001 level with a standardized coefficient of .57. Overall, adding in the media engagement variables reveals only a small amount of variation in respondents personal credibility ratings for CNN (adjusted R2 increases to .43 in model 3 compared to .40 in the previous model). Model 4 controls for respondents susceptibility to peer influence, which includes two variables: frequency of current event discussions with friends and value placed on friends opinions. Neither of the variables in model 4 a re statistically significant. The variable peers credibility rating remains significant (p<.001) in model 4 with an unchanged standardized coefficient (.57). The age variable also retains significance with its standardized coefficient remaining at .07. Adjusted R2 does not increase with the introduction of media engagement variables in model 4. Finally, model 5 controls for respondents political affiliation. Because Democrats comprised the largest politically affiliated segment, they were chosen as the reference group. Compared to the Democrat reference group, self identified Republicans rated CNN as being significantly less credible (p<.001). Those indicating that they belong to some other political group or have no political affiliation also rated CNN as being significantly less credible than did self identified Democrats (p<.01). The significance and strength of age remains unchanged in
76 the final model. Peers credibility rating retains significance in model 5 (p<.001) with a .55 standardized coefficient. When comparing the key independent variables standardized coefficient to those of other independent variables, it becomes apparent that peers credibility ratings for CNN is byfar the best single predictor of respondents personal credibility ratings. The introduction of political affiliation increased adjusted R2 from .43 to .46. Even after controlling for a host of other variables Table 5 1 shows a consistent, positive relationship between respondents perceptions of peers credibility ra tings for CNN and respondents own credibility ratings for CNN. Data do not allow for the rejection of the Hypothesis 1 as it pertains to CNN, and it is likely that that the positive association between these variables did not occur by chance. Linea r Regression for MSNBC Table 5 2 addresses the same relationships for MSNBC as Table 5 1 did for CNN. As was the case for linear regressions predicting respondents credibility ratings for CNN, peers credibility ratings for MSNBC are significant at the 001 level in all five models. In the first model, when respondents perceptions of peers credibility rating for MSNBC is entered by itself, 30% of the variation in respondents personal credibility ratings can be explained (standardized coefficient = .5 5; adjusted R2 = .30). Model 2 controls for demographic factors age, gender, and race. Gender was significant (at the .05 level) with a standardized coefficient of .08 and showed that female study participants rated MSNBC as more credible than their male counterparts. Once again, Asian was significant (p<.05), such that they assigned MSNBC higher credibility than White respondents (the reference group). The previous table showed no difference between Black and White respondents rating for CNN, but Tabl e 5 2 indicates that Blacks find MSNBC to be less credible than Whites (p<.05). Peers credibility rating remained significant (at the .001 level) with an unchanged standardized coefficient of .55. Once again, the
77 introduction of demographic information in the second model explains very little of the variation in respondents personal credibility ratings (adjusted R2 increases only slightly from .30 to .31). Variables relating to media engagement are controlled for in the third model. Frequency of watc hing MSNBC proved to be highly significant (at the .001 level) with a standardized coefficient of .23. Additionally, controlling for media engagement variables causes differences between Blacks and Whites credibility ratings for MSNBC to disappear. Both female and Asian maintained significance at the .05 level, though Asians standardized coefficient drops from .09 to .07. The key independent variable maintains significance at the .001 level, but its standardized coefficient falls from .55 in model 2 to .50 in model 3. Overall, the third model explains 36% of the variation in respondents credibility ratings for MSNBC (up from 31% in the previous model). In Model 4, respondents susceptibility to peer influence is taken into account. Neither of its variables discussion of current events with friends or value placed on friends opinions were significant. All previously significant variables in the regression remain unchanged after these new variables were introduced. Peers credibility rating remained significant (at the .001 level), keeping its .50 standardized coefficient from model 3. Both female and Asian remained significant (at the .05 level), and their standardized coefficients were unaltered. Frequency of watching MSNBC retained significance at the .001 level, while adjusted R2 was static at .36. The fifth and final model controls for political affiliation. Republicans, compared to Democrats (the reference group) rated MSNBC as being significantly less credible (p<.001). The var iable no political affiliation/other was also significant (at a .001 level), indicating that respondents in this group assigned MSNBC lower credibility ratings than those assigned by
78 Democrats. This model explains 2% more variation in the dependent varia ble than could be explained by the previous model R2 increased from .36 to .38. Peers credibility rating remained significant at the .001 level, and its standardized coefficient dropped only .01 between Model 4 and Model 5, from .50 to .49. Also, with the introduction of party affiliation, female, which was significant since its introduction into the regression in Model 2, lost its significance in Model 5. However, black, which was significant when first introduced in Model 2 but lost its significa nce in Models 3 and 4, again becomes significant in Model 5 (p<.05) with a standardized coefficient of .09. Differences between Asians and Whites remained significant in Model 5 (p<.05). Finally, the variable frequency of watching MSNBC maintained signi ficance (at a level of .001). Table 5 2 shows a consistent, positive relationship between respondents perceptions of peers credibility ratings for MSNBC and respondents own credibility ratings for MSNBC even after controlling for several independent variables. Data do not allow for the rejection of the Hypothesis 1 as it pertains to MSNBC, and it is likely that that the positive association between these variables did not occur by chance. Linear Regression for Fox News Results for Fox News linear regression are shown in Table 5 3. Data reveal a highly significant, positive relationship between peers credibility rating for Fox News and respondents own credibility ratings (p<.001 in all five models). In the first model, the key independent vari able by itself explains 35% of the variation in the dependent variable; its standardized coefficient is .59. Model 2 introduces the demographic variables age, gender, and race but none of these variables are significant. In fact, the variable peers c redibility rating actually sees an increase in its standardized coefficient, from .59 to .60. The adjusted R2 increases only marginally from .35 to .36.
79 The third model controls for two variables measuring media engagement. Both variables are significant at the .001 level. Frequency of watching Fox News has a positive coefficient, showing that those who watch Fox News more often rate it as being more credible. The variable informed about current events has a negative coefficient, meaning that those wh o consider themselves highly informed about current events rate Fox News credibility lower than respondents who report being less informed about current events. This is the only regression table out of the three in which this variable achieves statistical significance. Self reports concerning knowledge about current events are unrelated to respondents credibility ratings for CNN and MSNBC, but the extent which respondents think they are informed about current events is related to their credibility ratings for Fox News. Peers credibility rating maintains significance at a .001 level in this model, though its standardized coefficient falls from .60 to .54. Although no demographic variables were significant in model 2, other/mixed race becomes significant (p<.05) in the third model. Its standardized coefficient is positive demonstrating that Other and Mixed Race respondents find Fox News more credible than the Whites (the reference group). Out of the three regressions, other/mixed race is only signif icant with Fox News. Adjusted R2 is .45 in Model 3 up from .36 in the Model 2. This is the largest increase in the adjusted R2 between sequential models in any of the three regressions. In Model 4, variables measuring susceptibility to peer influence ar e introduced. Frequency of discussing current events with friends, is significant at the .05 level. Its negative coefficient reveals that respondents who frequently discuss current events with their friends rate Fox News as less credible. The second vari able, value placed on friends opinions, is also significant (p<.05). It has a positive coefficient, which indicates that those who value their friends opinions rate Fox News as being more credible. Out of the three channels regression
80 tables, this is only one in which either of the susceptibility to peer influence variables attain significance. This is not to say, however, that these variables explain much of the variation in respondents credibility ratings, as adjusted R2 increases only slightly from .45 in Model 3 to .46 in Model 4. The standardized coefficient for peers credibility rating remains unchanged at .54 and the standardized coefficient for other/mixed race is also unaltered at .08; both variables maintain their significance levels of .001 and .05, respectively. Frequency of watching Fox also maintains its significance (at the .001 level), its standardized coefficient increasing from .24 to .26. The variable informed about current events saw its standardized coefficient change fr om .17 in Model 3 to .14 in Model 4. In the final linear regression for Fox News, political affiliation is introduced. When comparing self identified Republicans to Democrats, the former rate Fox News as being significantly more credible (p<.001). Study respondents who indicate no political affiliation or being part of a political party in the other category do not differ from Democrats in terms of their credibility ratings for Fox News. Peers credibility rating retains its significance at the .001 level in Model 5, but its standardized coefficient falls from .54 to .52. Although the demographic variables Black and Asian, were not significant in any of the previous Fox News regressions, they attain significance in the final model (both with p values <.05). Other/Mixed Race gains even greater significance changing from a pvalue of less than .05 in models three and four to a pvalue of less than .01 in Model 5. The variable frequency of watching maintains its significance (at the .001 level ), but its standardize coefficient falls from .26 in Model 4 to .20 in Model 5. The variable informed about current events is not significant in regression for CNN or MSNBC, but it achieves significance for each Fox News model in which it is present. In fact, the
81 variable maintains its significance at the .001 level throughout. Its standardized coefficient does fall slightly from .14 in model 4 to .13 in model 5. The variables frequency of current event discussion with friends and value placed on fr iends opinions remain significant (at the .05 level) in the final model. As stated previously, Fox News regressions are the only ones in which being informed about current events, frequency of current event discussion with friends, and value placed on friends opinions are significantly related to respondents personal credibility ratings. On the whole, Table 53 demonstrates a consistent, positive relationship between respondents perceptions of peers credibility ratings for Fox News and respo ndents own credibility ratings for Fox News even after controlling for several independent variables. Data do not allow for the rejection of the Hypothesis 1 as it pertains to Fox News, and it is likely that that the positive association between these va riables did not occur by chance. In conclusion, the study revealed that respondents found CNN to be the most credible, while Fox News was rated least credible. In terms of respondents assessments of peers credibility ratings CNN emerged as most credible, and again Fox News was shown as least credible. T he linear regression s for al l three cable news channels demonstrated that peers credibility ratings w ere the best predictor of respondents own credibility ratings. Therefore, Hypothesis 1 was supported.
82 Figure 51. Distribution of Respondents Credibility Ratings for CNN Figure 52. Distribution of Respondents Credibility Ratings for MSNBC
83 Figure 53. Distribution of Respondents Credibility Ratings for Fox News Figure 54. Distribution of Res pondents Perceptions of Peers Credibility Ratings for CNN
84 Figure 55. Distribution of Respondents Perceptions of Peers Credibility Ratings for MSNBC Figure 56. Distribution of Respondents Perceptions of Peers Credibility Ratings for Fox News
85 Table 5 1. Results of Linear Regressions Predicting Effect of Independent Variables on Respondents Credibility Ratings for CNN Model 1 Model 2 Model 3 Model 4 Model 5 Peers Credibility Rating for CNN Demographics Age Female Blacka Asiana Other/Mixe d Racea Media Engagement Frequency of watching CNN Informed about current events Susceptibility to Peer Influence Frequency of current event discussion with friends Value placed on friends opinions Political Affiliation Republicanb No political af filiation/Otherb Adjusted R2 F .81(.63)*** .39 359.58*** .81(.62) *** .02( -.04) .03(.02) .04(.02) .19(.09) .02(.01) .40 61.92*** .74(.57) *** -.03( -.07) .05(.04) .01(.01) .14(.06) .02(.01) .10(.20)*** -.00( -.00) .43 52.37*** .74(.57)*** -.03( -.07) .05(.04) .01(.00) .14(.06) .02(.01) .10(.21)*** .01(.01) -.02( -.04) .00(.01) .43 41.81*** .71(.55)*** -.03( -.07) .02(.02) -.06( -.03) .13(.06) .02(.01) .09( .20)*** -.00( -.00) -.02( -.05) .00(.01) -.29( -.20)*** -.16( -.12)** .46 38.5 8 *** Notes: ***p<.001; **p<.01; *p<.05 N= 562 (total sample); Standardized coefficients in parentheses a. White/European is the reference group b. Democrat is the reference group
86 Table 5 2. Results of Linear Regressions Predicting Effect of Independent Variables on Respondents Credibility Ratings for MSNBC Model 1 Model 2 Model 3 Model 4 Model 5 Peers Credibility Rating for MSNBC Demographics Age Female Bla cka Asiana Other/Mixed Racea Media Engagement Frequency of watching MSNBC Informed about current events Susceptibility to Peer Influence Frequency of current event discussion with friends Value placed on friends opinions Political Affiliation Republicanb No political affiliation/Otherb Adjusted R2 F .67(.55)*** .30 233.70*** .67(.55)*** .00( -.01) .08(.08)* -.11( -.07)* .15(.09)* .04(.03) .31 41.89*** .61(.50)*** -.00( -.01) .08(.08)* -.0 9( -.06) .13(.07)* .05(.04) .10(.23)*** -.03( -.04) .36 38.20*** .61(.50)*** .00(.00) .08(.08)* -.09( -.06) .13(.07)* .05(.04) .10(.24)*** -.01( -.01) -.03( -.07) .01(.04) .36 30.99*** .59(.49)*** -.00( -.00) .07(.07) -.13( -.09)* .13(.08)* .05(.04) .10(.24)*** -.03( -.03) -.03( -.07) .01(.04) -.18( -.16)*** -.14( -.14)*** .38 28.12*** Notes: ***p<.001; **p<.01; *p<.05 N= 562 (total sample); Standardized coefficients in parentheses a. White/European is the referenc e group b. Democrat is the reference group
87 Table 5 3. Results of Linear Regressions Predicting Effect of Independent Variables on Respondents Credibility Ratings for Fox News Model 1 Model 2 Model 3 Model 4 Model 5 Peers Credibility Rating for Fox News Demographics Age Female Blacka Asiana Other/Mixed Racea Media Engagement Frequency of watching Fox Informed about current events Susceptibility to Peer Influence Frequency of current event discussion with friends Value placed on friends opinions Political Affiliation Republicanb No political affiliation/Otherb Adjusted R2 F .76(.59)*** .35 299.89*** .76(.60)*** -.03( -.05) .03(.02) .07(.03) .13(.05) .15(.06) .36 53.08*** .70(.54)*** -.03( -.05) .03(.02) .10(.04) .14(.05) .18(.08)* .16(.24)*** -.20( -.17)*** .45 55.19*** .69(.54)*** -.02( -.05) .02(.01) .10(.04) .14(.05) .18(.08)* .17(.26)*** -.17( -.14)*** -.03( -.08)* .03(.07)* .46 46.12*** .67(.52)* ** -.02( -.04) .03(.02) .18(.07)* .19(.07)* .21(.09)** .13(.20)*** -.16( -.13)*** -.05( -.07)* .03(.08)* .36(.18)*** .00(.00) .48 42.62*** Notes: ***p<.001; **p<.01; *p<.05 N= 562 (total sample); Standardized coefficients in parentheses a White/European is the reference group b. Democrat is the reference group
88 CHAPTER 5 DISCUSSION AND CONCL USION Summary One of this studys most salient findings is that peers credibility rating is highly significant (p<.001) in all models for all three regressions. In fact, respondents perceptions of peers credibility ratings are, by far, the single best predictor of respondents own ratings. Because of this, we fail to reject Hypothesis 1. As was previously stated in the literature review and was bor n out in the study, Banduras social cognitive theory informed that there was likely a link between an individuals environment and their cognitive decisions be they on cable news credibility or any other choice. Although causality cannot be established in cross sectional studies, we do learn that there is a correlation between how respondents rate cable news channels credibility and how they perceive their peers to rate these same cable news channels credibility. In each of the linear regressions tabl es, this independent variable, peers credibility rating, had the highest standardized coefficient, never dropping below .49 more than double the standardized coefficients of all other independent variables. Another noteworthy finding is the relative ly strong relationship between frequency of watching and respondents credibility rating, which indicated that watching a channel more frequently is associated with higher credibility ratings. This variable attained significance at the .001 level for a ll models in CNN, MSNBC and Fox News regressions. When comparing across all the final models, frequency of watching was the second best predictor of respondents credibility ratings after peers credibility rating. Specifically, standardized coefficients for this variable are as follows: CNN is .20, MSNBC is .24, and Fox News is .20. Political affiliation also had an effect. Significant differences between Republicans and Democrats were evident in all three regressions (p<.001). CNN received a .20 st andardized
89 coefficient indicating that Republicans rate CNN as less credible than Democrats. This is expected based on numerous studies which show that Republicans rate CNN as less credible than Fox News (Pew Research Center, 2004; Pew Research Center, 200 6; Pew Research Center, 2009; Public Policy Polling, 2009; Public Policy Polling, 2010). Interestingly, CNN had a similar disparity between the no political affiliation/other variable and the Democrat reference group (at .01 significance level) with a .12 standardized coefficient. Thus, those with no political affiliation also rated CNN less credible than Democrats. MSNBCs results for political affiliation nearly mirrored those of CNN. The variable Republican registered a .001 significance level with a .16 standardized coefficient, and no political affiliation/other had a .001 significance level and a .14 standardized coefficient. Thus, similar to CNN, Republicans and those with no political affiliation rated MSNBC less credible than Democrats (the reference group). However, though Fox News, too, had a .001 significance level for Republican, with a .18 standardized coefficient, no political affiliation/other was not significant at any level. The Republican standardized coefficient was positive o nly in the Fox News regressions, indicating that Republicans finds Fox News to be significantly more credible than Democrats (the reference group). The reverse is true for the other two channels; Democrats rate CNN and MSNBCs credibility higher than Republicans do. This supports previous studies in which Republicans rated Fox News as more credible than both CNN and MSNBC (Pew Research Center, 2004; Pew Research Center, 2006; Pew Research Center, 2009; Public Policy Polling, 2009; Public Policy Polling, 201 0). Another interesting result of this study is that Blacks rated Fox News higher in terms of credibility than the White reference group. Additionally, Blacks rated MSNBC as less credible compared to Whites. Both of these racial differences remained even after controlling for political
90 affiliation. The literature shows that MSNBC is consistently rated by viewers to be more frequently watched and trusted by Democrats (Pew Research Center, 2004; Pew Research Center, 2006; Pew Research Center, 2009; Public Policy Polling, 2009; Public Policy Polling, 2010). African Americans have traditionally voted for Democrats in much higher numbers than for Republicans, including the 2008 presidential election where 95% of blacks voted Democratic ( ElectionCenter, 2008). It is perhaps surprising then that after taking political party into account Blacks tended to rate Fox News as more credible and MSNBC as less credible compared to Whites. Perhaps the most interesting and telling statistic in the entire study was the result for the variable informed on current events, which revealed that those who considered themselves highly informed about current events rated Fox News as significantly less credible (p<.001). The variables standardized coefficient for Fox News was .13. Also of interest is that CNN emerged as most credible, while Fox News was considered least credible, according to both respondents personal assessments and their perceptions of peers opinions regarding the channels trustworthiness and objectivity. This fo llows past research which shows CNN to be the most credible of the three channels (Pew Research Center 2004; Pew Research Center 2008) Specifically, the average personal credibility ratings for these news channels are as follows: CNN= 3.13, MSNBC= 3.03, and Fox News= 2.66. Respondents assessments of peers credibility ratings for the three news channels were similar (CNN= 3.11, MSNBC= 3.04, and Fox News= 2.77). These relative credibility ratings confirm some previous studies (Pew Research Center, 2004; Pew Research Center, 2006; Pew Research Center, 2008), but contradict others (Public Policy Polling, 2009; Public Policy Polling, 2010)
91 Explanation Certainly, the most important finding in this study was that Hypothesis 1 was not rejected and, in fact, a chieved significance at the .001 level. The presence of a correlation between how respondents rated the three cable news channels and how they perceived their peers would rate the three channels suggests that peers may, in fact, influence how individuals j udge media credibility. Based on Banduras (1985) social cognitive theory, these respondents cognition (how they view and analyze the media) and their behavior (how they filled out the survey) are related to their environment (interaction with peers). Thu s, according to Banduras theory, this peer interaction may have influenced respondents opinions regarding cable news channels trustworthiness and objectivity. Causality, however, can only be ascertained in longitudinal studies, and, therefore, we cannot know whether respondents choose to interact with those who have similar opinions regarding media credibility or if their own opinions change over time to more closely resemble those of their peer group. Perhaps both factors are at play: respondents might be drawn to their like minded peers, but may also alter their attitudes to align even more closely with their friends attitudes. Though past studies show that individuals seek out those who share similar world views (Shimahara, 1983), there is also reason to believe that young adults are influenced by their friends. When considering Fox News regression (Table 53), frequency of current event discussion with friends is significant in both models 4 and 5. The coefficients have negative signs, indicating that those who discuss current events with friends more frequently rate Fox News as less credible. Respondents who often engage in conversations about world events have comparatively negative views regarding Fox News credibility. Conversely, those who do not usually discuss current events with their friends rate Fox News credibility higher. This suggests that these higher credibility ratings are due to a lack of discussion with their peers during which
92 their opinions on Fox News credibility could have been modified. This variable failed to achieve significance in either the CNN or MSNBC regressions. T hese two ch annels were rated as more credible than Fox News. Avid Fox News viewers, then, would seem to be less aware of their friends opinions regarding t his channels credibility, as individuals who find Fox News highly credible are less likely to discuss current events, and, therefore, Foxs reputation with peers. In other words, they may not have heard their peers opinions regarding cable news channels, which, thus, would negate the possibility of peer influence. Also, this study showed Fox News to be very polarizing. For example, Fox News has larger standard deviations than either CNN or MSNBC for both per sonal and peer credibility composite s (see Fig ures 5 1 through 56). Additionally, 49 study participants assigned the lowest possible credibility rating for all four of the questions comprising the respondents credibility composite for the channel (see Figure 5 3), while only eight respondents did this for CNN and MSNBC, combined (see Figure 5 1 and Figure 52). Furthermore, when answering questions about how their friends assess a given channels credibility, only three respondents, for CNN and MSNBC (see Figure 5 4 and Figure 55) combined, gave the channels the lowest possible rating on all four questions. Twentyfive respondents, however, indicated their friends found Fox News (see Figure 5 6) to be completely devoid of credibility. Because there are such negative sentiments against Fox News, it seems likely that peers are strongly advocating against the channel, which may partially explain its comparatively low personal credibility rating. Data indicate that greater media exposure is related to higher credibility assessments. For all three cable news channels, there was a positive correlation between channel viewing and credibility ratings (p<.001 in all cases). This association parallels results from several other studies (e.g. Greenberg, 1966; Whitney, 1986; Johnson & Kaye, 1998; Kiousis, 2001).
93 Limitations The current research project, like all studies, has limitations. The fact that a non random sample of undergraduate students was recruited from a single university does not allow for the generalization of findings to circumstances beyond the studys context. For example, conducting the research at UF, a Southeastern university likely resulted in an overrepresentation of students with conservative political leanings compared to students attending colleges in the Northeastern United States. Thus, study findings could overestimate credibility ratings for Fox News, as a nationally representative sample of college students might very well have included fewer Republicans. Also, the UF Office of Institutional Planning and Research reports that 1.4% o f undergraduate students are NonResident Aliens; persons born outside the United States may express unique attitudes toward cable news channels credibility that are hidden by their inclusion in this studys broad racial categories (2010) Additio nally, if one is interested in peer influence throughout the life course, the use of a college student sample is particularly problematic. First, college students are more educated or at least on their way to being more educated than the majority of Am ericans (U.S. Census Bureau, 2000). Second, college students tend to be more politically liberal and are more likely to identify as Democrats compared to the population as a whole. Self selection bias can also sometimes present a problem in social science research because its voluntary nature may allow for dissimilarity between survey takers and those who decline participation. In this instance, however, very few students chose not to complete the survey. Some respondents claimed never to have viewed CNN, M SNBC, and Fox News. The ir answers tended toward neutral ity, which pulled the mean closer to In the future, perhaps these non viewers should either be analyzed separately or eliminated from the study. However, because these study participants may still
94 have opinions about the channels, even without first hand exposure, both options have drawbacks. It is also somewhat disconcerting that a large number of students chose the neutral or Neither Agree Nor Disagree option when rating the three cable news channels. Although social science research demonstrates that respondents nearly always gravitate toward the center of any Likert scale (Garland, 1991) and it is quite possible that a large number of respondents have relatively neutral attitudes rega rding cable new channels trustworthiness and objectivity, a high concentration of response around the number 3 could also indicate that study participants did not take the time to carefully consider each and every question. It is, after all, easier to a nswer Neither Agree Nor Disagree than to expend the effort necessary to contemplate every item on a three page survey. There is no way to know for certain if a respondent actually espoused rather neutral views regarding the three channels credibility or they simply tired of answering questions. Perhaps, in the future a different Likert scale could be used. A four point scale which eliminates the Neither Agree Nor Disagree answer choice is a potential option because there will not be a neutral response. Also, sevenpoint and nine point Likert scales could be advantageous, as they allow respondents to express opinions that vary only slightly from neutrality. Therefore, respondents may be less inclined to report neutral opinions. Suggestions for Future Res earch This study lays the groundwork for a plethora of future research. Because media credibility has not been viewed in terms of how peers impact individuals opinions on media credibility, there are many areas in which future research might expand on this concept. Using a similar format to that used in this study, research may be conducted on medium credibility, comparing television, radio, Internet, and print. Additionally, studies could be conducted on how peers affect respondents credibility assessment for different network news broadcasts or their
95 attitudes toward other mediums (e.g. newspapers or websites). Ideally, these future studies would employ a longitudinal design and random, nationally representative samples, but time and cost constraints obviously make this less feasible. Finally, experiments would be useful to demonstrate the extent to which peers influence individuals assessments of media credibility. Analyzing how a confederates outburst against CNN, MSNBC or Fox News prior to survey ad ministration might prove particularly illuminating. Using a scenario which allows respondents to immediately echo or counter the opinion of one or more confederates would be helpful in understanding the role immediacy plays in peer influence. Of course adequate control groups would need to be put in place (perhaps by varying the amount of time between exposure to a peers opinion and survey administration). Closing The current research project makes a significant contribution to the literature. Findings point to a link between individuals assessments of media credibility and their perceptions of peers assessments. The existence of a correlation between respondents and peers credibility ratings for various television news sources has widereaching i mplications. For example, the cable channels in question (CNN, MSNBC, and Fox News) would perhaps benefit from knowing that viewers, or potential viewers, will judge their trustworthiness and objectivity based not only on their own viewing experience but a lso on their friends opinions about the channel. The study also suggests an explanation for the correlation between individuals and peers opinions. It reveals that those who discuss current events with their friends rate Fox News credibility very low which indicates that peers alter one anothers credibility opinions through discourse. Conversely, those who do not discuss current events rate Fox News as being more credible, which indicates that if they were to engage in such discourse, their opinions might be
96 modified to converge closer toward the majority opinion (Fox News mean = 2.66). Those who engage in no such discourse, thus, remain steadfast in their minority viewpoint. The most important accomplishment of the current research project, however, is perhaps the fact that it unites two disparate bodies of literature. Specifically, media credibility studies and peer influence research intersect as the link between individuals and peers credibility ratings. Furthermore, results demonstrate that it is vitally important for media companies to convince peer group leaders that their channels are credible perhaps this is even more essential than broad appeals aimed at the entire audience, as these opinion leaders will likely affect others attitudes. B efore marketing campaigns can prove truly effective, however, news outlets must ensure that they are providing audiences with consistently credible journalism. In doing so, the media may be able to regain some of the publics faith and trust it has lost over the last quarter century.
97 APPENDIX A SURVEY QUESTIONNAIRE A Thank you for agreeing to participate in this research project Th e following survey includes questions about your background, cable news viewing habits, and attitudes toward cable news. There are no right or wrong answers. Simply mark the choices that best describe you. PLEASE NOTE: For this survey, the term friends includes best friends, close friends, and casual friends 1. How often do you discuss the news/current events with your friends? DISCUSS 1 [ ] Never 2 [ ] Once a month or less frequently 3 [ ] Two or three times a month 4 [ ] Once a week 5 [ ] Two or three times a week 6 [ ] Daily 2. To what extent are you informed about current events? INFRMD 1 [ ] Not at all Inform ed 2 [ ] Not Informed 3 [ ] Somewhat Informed 4 [ ] Informed 5 [ ] Very Informed 3. What value do you place on your friends opinions when making decisions? FR_DEC 1 [ ] Very Low Value 2 [ ] Low Value 3 [ ] Some Value 4 [ ] High Value 5 [ ] Very High Value 4. How frequently do you watch CNN WAT_CNN 1 [ ] Never 2 [ ] Once a month or less frequently 3 [ ] Two or three times a month 4 [ ] Once a week 5 [ ] Two or three times a week 6 [ ] Daily
98 Strongly Disagree 1 Disagree 2 Neither Agree Nor Disagree 3 Agree 4 Strongly Agree 5 5. I believe all or most of the information presented on C NN CRED_01 1 2 3 4 5 6. A n effort is made to present all sides of the issues on CNN CRED_0 2 1 2 3 4 5 7. There is often b iased reporting on CNN CRED_0 3 1 2 3 4 5 8. I often distrust the reporting on CNN CRED_ 04 1 2 3 4 5 9. My friends believe all or most of the information presented on CNN CRED_ 05 1 2 3 4 5 10. My friends think a n effort is made to present all sides of the issues on CNN CRED_ 06 1 2 3 4 5 11. My friends think there is often biased reporting on CNN CRED_ 07 1 2 3 4 5 12. My friends often distrust the reporting on CNN CRED_ 08 1 2 3 4 5 13. How frequently do you watch MSNBC WAT_NBC 1 [ ] Never 2 [ ] Once a month or less frequently 3 [ ] Two or three times a month 4 [ ] Once a week 5 [ ] Two or three times a week 6 [ ] Daily
99 Strongly Disagree 1 Disagree 2 Neither Agree Nor Disagree 3 Agree 4 Strongly Agree 5 14. I believe all or most of the information presented on MSNBC CRED_ 09 1 2 3 4 5 15. A n effort is made to present all sides of the issues on MSNBC CRED_ 10 1 2 3 4 5 16. T here is often biased reporting on MSNBC CRE D_ 11 1 2 3 4 5 17. I often distr ust the reporting on MSNBC CRED_ 12 1 2 3 4 5 18. My friends believe all or most of the information presented on MSNBC CRED_ 13 1 2 3 4 5 19. My friends think a n effort is made to present all sides of the issues on MSNBC CRED_ 14 1 2 3 4 5 20. My friends think there is often biased reporting on MSNBC CRED_ 15 1 2 3 4 5 21 My friends often distrust the reporting on MSNBC CRED _16 1 2 3 4 5 22. How fre quently do you watch Fox News WAT_FX 1 [ ] Never 2 [ ] Once a month or less frequently 3 [ ] Two or three times a month 4 [ ] Once a week 5 [ ] Two or three times a week 6 [ ] Daily
100 Strongly Disagree 1 Disagree 2 Neither Agree Nor Disagree 3 Agree 4 Strongly Agree 5 23. I believe all or most of the information presented on Fox News CRED_ 17 1 2 3 4 5 24. A n effort is made to present all sides of the issues on Fox News CRED_ 18 1 2 3 4 5 25. T here is often biased reporting on Fox News CRED_ 19 1 2 3 4 5 26. I often distrus t the reporting on Fox News CRED_ 20 1 2 3 4 5 27. My friends believe all or most of the information presented on Fox News CRED_ 2 1 1 2 3 4 5 28. M y friends think a n effort is made to present all sides of the issues on Fox News C RED_ 22 1 2 3 4 5 29. My friends think there is often biased reporting on Fox News CRED_ 23 1 2 3 4 5 30 My friends often distrust the reporting on Fox News C RED_ 24 1 2 3 4 5
101 31. Gender: GENDER 1 [ ] Fem ale 2 [ ] Male 32. What is your age? (please enter it in the blank) AGE _______________ 33. Which of the following races do you consider yourself? You may select more than one and/or specify another racial group. RACE 1 [ ] White/European 2 [ ] Black/African -American 3 [ ] Asian or Pacific Islander 4 [ ] American Indian or Alaskan Native 5 [ ] Other: _______________________ 34. What is your political affiliation? POLITIC 1 [ ] Republican 2 [ ] Democrat 3 [ ] No political affiliation 4 [ ] Other: _______________________ Thank you very much fo r participating in this study. I really appreciate your help!
APPENDIX B IRB DOCUMENTATION Matthew Beaton P.O. Box 118400 Department of Journalism Gainesville, FL 32611 7330 Phone: (352 ) 3920466 Email: firstname.lastname@example.org INFORMED CONSENT FORM You are cordially invited to participate in a study on cable news viewing habits and attitudes toward cable news conducted by Matthew Beaton, a graduate student in the College of Journalism and Communications, at the University of Florida. In addition to questions about cable news, this questionnaire will ask for a variety of demographic information such as age, gender, race, ethnicity, parents household income, and political affiliation. The survey will require approximately ten to fifteen minutes of your time. There are no anticipated risks associated with this research project. Your participation in this study is completely voluntary and you are not required to answer any questions that you do not wish to answer. As a thank you for participating, you will receive one extra credit point (above the maximum of 100 credit points or its equivalent) added to your final course grade. If you do not wish to complete the survey, there is an optional activity which you may perform in lieu of the survey which will earn you one extra credit point. Please notify the researcher if you would prefer performing the activity instead of completing the survey. After the surveys completion, your professor may direct you to add your name and/or UF ID number to a list of students who have participated. If such a list exists, it would merely serve as a way to ensure extra credit points are assi gned correctly and would never be used to connect your name to your survey responses. Any data collected in conjunction with the study will be kept confidential to the extent provided by law. In order to protect your identity questionnaires will contain only random identification numbers. By signing this form below, you indicate that you have chosen to participate in this research and that you have read and understand the information in this consent form. Your decision regarding whether or not to partici pate will not prejudice your relations with the College of Journalism and Communications or the University of Florida. If you decide to participate, you are completely free to withdraw consent and discontinue participation at any time without penalty. If you have any additional questions, you may contact Matthew Beaton at 706372 6707. Questions and concerns about your rights as a research participant can be directed to the University of Florida Institutional Review Board office, PO Box 112250, University of Florida, Gainesville, FL 32611 2250. Thank you very much! ______________________________________________________________________________ I have read the procedure described above. I voluntarily agree to participate in the procedure, and I have recei ved a copy of this description. Participants signature______________________________ Date____________ Principal investigator ______________________________ Date____________
103 Matthew Beaton P.O. Box 118400 Department of Journalism Gainesville, FL 32 6117330 Phone: (352) 3920466 Email: email@example.com INFORMED CONSENT FORM You are cordially invited to participate in a study on cable news viewing habits and attitudes toward c able news conducted by Matthew Beaton, a graduate student in the College of Journalism and Communications, at the University of Florida. In addition to questions about cable news, this questionnaire will ask for a variety of demographic information such as age, gender, race, ethnicity, parents household income, and political affiliation. The survey will require approximately ten to fifteen minutes of your time. There are no anticipated risks associated with this research project, but there are also no di rect benefits to you for participating in this study. Your participation in this study is completely voluntary, and you are not required to answer any questions that you do not wish to answer. After the surveys completion, your professor may direct you to add your name and/or UF ID number to a list of students who have participated. If such a list exists, it would merely serve as a way to ensure extra credit points are assigned correctly. Such extra credit would total no more than 1 extra percentage point in addition to the potential 100 percentage points a student may earn during the semester and would never be used to connect your name to your survey responses. If you do not wish to complete the assignment and extra credit is made available for completi ng it, an alternative assignment can be performed to earn the extra credit. Any data collected in conjunction with the study will be kept confidential to the extent provided by law. In order to protect your identity questionnaires will contain only random identification numbers. By signing this form below, you indicate that you have chosen to participate in this research and that you have read and understand the information in this consent form. Your decision regarding whether or not to participate will not prejudice your relations with the College of Journalism and Communications or the University of Florida. If you decide to participate, you are completely free to withdraw consent and discontinue participation at any time without penalty. If you have any additional questions, you may contact Matthew Beaton at 706372 6707. Questions and concerns about your rights as a research participant can be directed to the University of Florida Institutional Review Board office, PO Box 112250, University of Florida, Gainesville, FL 32611 2250. Thank you very much! If you have any further questions or concerns, feel free to direct them to my supervisor: Dr. David Ostroff, Ph.D. Department of Journalism E mail: firstname.lastname@example.org ____________________________________________________________________________________ I have read the procedure described above. I voluntarily agree to participate in the procedure, and I have rec eived a copy of this description. Participants signature______________________________ Date____________ Principal investigator ______________________________ Date____________
104 LIST OF REFERENCES Abdulla, R., Garrison, B., Salwen, M., Driscoll, P., & Case y, D. (2002). The credibility of newspapers, television news, and online news. P resented to the Mass Communication and Society Division, Association for Education in Journalism and Mass Communication, annual convention, Miami Beach, Fla. Abel, J. & Wirth M. (1977). Newspaper vs. TV credibility for local news. Journalism Quarterly, 54(2), 371375. Addington, D. (1971). The effect of vocal variations on ratings of source credibility. Speech Monographs, 38, 242247. Allen, R., & Clarke, D. (1980). Ethnicit y and mass media behavior: A study of blacks and Latinos. Journal of Broadcasting, 24 (1), 2334. Andsager, J. (1990). Perceptions of credibility of male and female syndicated political columnists. Journalism Quarterly, 67 (3), 485491. Antonucci, T. (2001). Social relations: An examination of social networks, social support, and sense of control. In J. E. Birren & K. W. Schaie (Eds.), Handbook of the psychology of aging (5th ed., pp. 427453). San Diego, CA: Academic Press. Bailey, K. (1994). Methods of social research (4th ed.) New York, NY: The Free Press. Bandura, A., &, Walters, R. (1963). Social learning and personality development. New York, NY: Holt, Rinehart, & Winston. Bandura, A. (1985). Model of causality in social learning theory. In Freeman, Arthur, Mahoney, Michael J., Devito, Paul, & Martin, Donna (Eds.), Cognition and psychotherapy (2544) New York, NY: Springer Publishing Company, Inc. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Clif fs, NJ: Prentice Hall. Bandura, A. (1989). Social cognitive theory of organizational management. The Academy of Management Review, 14 (3), 361384. Bandura, A. (1992). Social cognitive theory. In R. Vasta (Ed.), Six Theories of Child Development: Revised Fo rmulations and Current Issues (pp. 160). London, UK: Jessica Kingsley. Bandura, A. (1999). Social cognitive theory: An agentic perspective. Asian Journal of Social Psychology. 2, 2141. Bandura, A. (2001a). Social cognitive theory of mass communication. M edia Psychology, 3(3), 265299.
105 Bandura, A. (2001b). Social cognitive theory: an agentic perspective. Annual Review of Psychology 52, 126. Bandura, A. (2004). Health promoti on by social cognitive means. Health Education & Behavior 31, 143164. Bauman, K., & Ennett, S. (1994). Peer influence on adolescent drug use. American Psychologist, 49, 820822. Bearden, W., & Etzel, M. (1982). Reference group influence on product and br and purchase decisions. The Journal of Consumer Research, 9(2), 183194. Berlo, D., Lemert, J., & Mertz, R. (1970). Dimensions for evaluating the acceptability of message sources. Public Opinion Quarterly 33(4), 563576. Bernard, H. (2000). Social research methods: Qualitative and quantitative approaches Thousand Oaks, CA: Sage Publications, Inc. Biddle, B., Bank, B., & Marlin, M. (1980). Parental and peer influence on adolescents. Social Forces, 58(4), 10571079. Billy, J., Rodgers, J., & Udry, J.R. (1984). Adolescent sexual behavior and friendship choice. Social Forces 62(3), 653678. Breakwell, G., & Beardsell, S. (1992). Gender, parental, and peer influences upon science attitudes and activities. Public Understanding of Science, 1 (2), 183198. Bryman A. (1989). Research methods and organization studies. New York, NY: Rutledge. Bucy, E. (2003). Media credibility reconsidered: Synergy effects between onair and online news. Journalism and Mass Communication Quarterly 89(2), 247264. Cassidy, W. (2007) Online news credibility: An examination of the perceptions of newspaper journalists. Journal of Computer Mediated Communication, 12, 144164. Chia, S. (2006). How peers mediate media influence on adolescents sexual attitudes and sexual behavior. Journal of Communication, 56(3), 585606. Childers, T., & Rao, A. (1992). The influence of familial and peer based reference groups on consumer decisions. Journal of Consumer Research, 19, 198211. Denham, B. (2008, January). Determinants of anabolic androgenic s teroid risk perceptions in youth populations: A study in social cognitive theory. Paper presented at: 2008 Annual Conference of the International Communication Association, Health Communication Division, Montreal, May 2008. Dey, E. (1996). Undergraduate political attitudes: An examination of peer, faculty, and social influences. Research in Higher Education, 37 (5), 535554.
106 Dey, E. (1997). Undergraduate political attitudes: peer influence in changing social contexts. The Journal of Higher Education, 68(4), 398413. Diddi, A., & LaRose, R. (2006). Getting hooked on news: Uses and gratifications and the formation of news habits among college students in an internet environment. Journal of Broadcasting & Electronic Media, 50(2): 193210. Dohnt, H., & Tiggemann, M. (2006). The contribution of peer and media influences to the development of body satisfaction and self esteem in young girls: A prospective study. Developmental Psychology, 42 (5), 929936. Duncan, G., Boisjoly, J., Kremer, M., Levy, D., & Eccles, J. (2 005). Peer effects in drug use and sex among college students. Journal of Abnormal Child Psychology, 33(3), 375385. Edelstein, A., & Tefft, D. (1974). Media credibility and respondent credulity with respect to Watergate. Communication Research, 1(4), 426439. Eveland, W., Jr., & Shah, D. (2003). The impact of individual and interpersonal factors on perceived news media bias. Political Psychology, 24(1), 101117. ElectionCenter 2008 Exit Polls. (2008). CNN.com. Retrieved from http://www.cnn.com/ELECTION/2008/results/polls/#val=USP00p1 Fields, J. M., & Schuman, H. (1976). Public beliefs about the beliefs of the public. Public Opinion Quarterly, 40(4), 427448. Flanigan, A., & Metzge r, M. (2000). Perceptions of Internet information credibility. Journalism and Mass Media Quarterly, 77(3), 515 540. Galletta, D., Ahuja, M., Hartman, A., Peace, A. G., & Teo, T. (1994). An empirical study of peer influence on user attitudes, behavior, and performance Paper presented at: Proceedings of the fifteenth annual International Conference on Information Systems, Vancouver, December 1994. Gardner, M., & Steinberg, L. (2005). Peer influence on risk taking, risk preference, and risky decision making i n adolescence and adulthood: An experimental study. Developmental Psychology, 41(4), 625635. Garland, R. (1991). The midpoint on a rating scale: Is it desirable? Marketing Bulletin, 2, 6670. Gaughan, M. (2006). The gender structure of adolescent peer influence on drinking. Journal of Health and Social Behavior, 47(1), 4761. Gaziano, C., & McGrath, K. (1986). Measuring the concept of credibility. Journalism Quarterly, 3(3), 451462. Gaziano, C. (1987). News peoples ideology and the credibility debate. N ewspaper Research Journal, 9 (1), 118.
107 Greenberg, B. (1966). Media use and believability: Some multiple correlates. Journalism Quarterly, 43 (4) 665670. Gunther, A. C. (1988). Attitude extremity and trust in the media. Journalism Quarterly, 65(2) 279287. Gunther, A. C. (1992). Attitudes toward media coverage of social groups. Public Opinion Quarterly, 56 (2), 147167. Gunther, A. C., & Storey, J. D. (2003). The influence of presumed influence. Journal of Communication, 53(2), 199215. Hallinan, M., & Williams, R. (1990). Students characteristics and the peer influence process. Sociology of Education, 63(2), 122132. Henke, L. (1985). Perceptions and use of news media by college students. Journal of Broadcasting and Electronic Media, 29(4), 431436. Heon g, K. L., Escalada, M. M., Huan, N. H., & Mai, V. (1998). Use of communication media in changing rice farmers pest management in the Mekong Delta, Vietnam. Crop Protection, 17(5), 413425. Hopkins, D. (2009). The 2008 election and the political geography of the new democratic majority. Polity, 41(3), 368387. Howard, H., Blick, E., & Quarles, J. (1987). Media choices for specialized news. Journalism Quarterly, 64 (2), 620623. Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication effectiveness. The Public Opinion Quarterly, 15 (4), 635650. Hovland, C. I., Janis, I., & Kelley, H. (1953). Communication and persuasion. New Haven, CT: Yale University Press. Huff, R. (1996, July 19). Aiming to outfox CNN, MSNBC. The New York Da ily News. Ibelema, M., & Powell, L. (2001). Cable television news viewed as most credible. Newspaper Research Journal, 22(1), 4151. Johnson, T. J. (1993). Exploring media credibility: How media and nonmedia workers judged media performance in Iran/Contra Journalism Quarterly, 70 (1), 8797. Johnson, T., & Kaye, B. (1998). Cruising is believing?: Comparing Internet and traditional sources on media credibility measures. Journalism and Mass Communication Quarterly, 75(2), 325340.
108 Johnson, T., & Kaye, B. (2004). Wag the blog: How reliance on traditional media and the Internet influence credibility perceptions of weblogs among blog users. Journalism and Mass Communication Quarterly, 81(3), 622642. Jones, D. (2004). Why Americans dont trust the media: A prel iminary analysis. The Harvard International Journal of Press/Politics, 9(2), 60 75. Kelman, H., & Hovland, C. I. (1953). Reinstatement of the communicator in delayed measurement of opinion change. Journal of Abnormal Social Psychology, 48(3), 327335. Ki nard, B., & Webster, C. (2010). The effects of advertising, social influences, and self efficacy on adolescent tobacco use and alcohol consumption. The Journal of Consumer Affairs, 44(1), 2443. Kiousis, S. (2001). Public trust or m istrust ? Perceptions of media credibility in the information age. Mass Communication and Society, 4(4), 381 403. Kohring, M., & Matthes, J. (2007). Trust in news media: Development and validation of a multidimensional scale. Communication Research, 34 (2), 231252. La Gaipa, J. J. (1977). Testing a multidimensional approach to friendship. In S. Duck (Ed.), Theory and Practice in Interpersonal Attraction (pp. 249270) London: Academic Press. La Guardia, J. G., Tyan, R. M., Couchman, C. E., & Deci, E. L. (2000). Withinperson varia tion in security of attachment: A self determination theory perspective on attachment, need fulfillment, and wellbeing. Journal of Personality and Social Psychology 79, 367384. Larsen H. Engels R. Souren, P. Granic I., & Overbeek, G. (2010). Peer influence in a micro perspective: Imitation of alcoholic and nonalcoholic beverages. Addictive Behaviors, 35(1), 4952. Larson, L. (1972). The influence of parents and peers during adolescence: the situation hypothesis revisited. Journal of Marriage and Family, 34(1), 6774. Lee, R. (1978). Credibility of newspaper and TV news. Journalism Quarterly, 55(2), 282287. Lee, T. T. (2005). The liberal media myth revisited: an examination of factors influencing perceptions of media bias. Journal of Broadcasting & Electronic Media, 49 (1), 4364. Lublin, D. (2004). The Republican South. Princeton, NJ: Princeton University Press. Lucas, D. B., & Britt, S. H. (1950). Advertising psychology and research. New York, NY: McGraw Hill. MacKuen, M., & Brown, C. (1987). Poli tical context and attitude change. The American Political Science Review, 81 (2), 471490.
109 Maier, S. (2002) Getting it right? Not in 59 percent of stories. Newspaper Research Journal, 23(1), 1024. Major, A., & Atwood, L. E. (1997). Changes in media credibi lity when a predicted disaster doesnt happen. Journalism and Mass Communication Quarterly, 74(4), 797813. Mangleburg, T., Doney, P., & Bristol, T. (2004). Shopping with friends and teens susceptibility to peer influence. Journal of Retailing, 80, 101116. Marburger, D. R. (2001). Absenteeism and undergraduate exam performance. Journal of Economic Education, 32 99109. Markham, D. (1968). The dimensions of source credibility of television newscasters. Journal of Communication, 18, 5764. Meeus, W., & De kovic, M. (1995). Identity development, parental and peer support in adolescence: results of a national Dutch survey. Adolescence, 30(120), 931944. Mifflin, M. (1996, October 7). At the new Fox News Channel, the buzzword is fairness, separating news from bias. The New York Times. Miller, A. H., Goldenberg, E. N., & Erbring, L. (1979). Type set politics: Impact of newspapers on public confidence. The American Political Science Review 73(1), 6784. Mirande, A. M. (1968). Reference group theory and adolesce nt sexual behavior. Journal of Marriage and the Family 30(4), 572 577. Mitchelstein, E., & Boczkowski, P. (2010). Online news consumption research: An assessment of past work and an agenda for the future. New Media & Society XX(X): 1 18. Morris, J. (2005). The Fox factor. The Harvard International Journal Press/Politics, 10(3), 5679. Mulder, R. (1981). A loglinear analysis of media credibility. Journalism Quarterly, 58 (4), 635638. Mulder, R. (1980). Media credibility: A use gratifications approach. Journalism Quarterly, 57(3), 474477. Neuwirth, K., & Frederick, E. (2004) Peer and social influence on opinion expression: Combining the theories of planned behavior and the spiral of silence. Communication Research, 31 (6), 669703. Nunnally, C. ( 1978). P sychometric Theory, 2nd ed. New York: McGrawHill Book Co. Oliver, K., & Thelen, M. (1996). Childrens perceptions of peer influence on eating concerns. Behavior Therapy, 27(1), 2539.
110 Osgood, C., Suci, G., & Tannenbaum, P. (1957). The Measurement of Meani ng (8th ed.). Urbana, IL: University of Illinois Press. Oyedeji, T. (2007). The relation between the customer based brand equity of media outlets and their media channel credibility: an exploratory study. International Journal of Media Management, 9(3), 116 125. Paek, H. J., & Gunther, A. C. (2007). How peer proximity moderates indirect media influence on adolescent smoking. Communication Research, 34(4), 407432. Paletz, D., Short, J., Baker, H., Campbell, B., Cooper, R., & Oeslander, R. (1980). Polls in t he media: Content, credibility, and consequences. The Public Opinion Quarterly, 44(4), 495513. Parsons, T. (1963). On the concept of influence. The Public Opinion Quarterly, 27(1), 3762. Paek, H. J. (2009). Differential effects of different peers: Furthe r evidence of the peer proximity thesis in perceived peer influence on college students smoking. Journal of Communication, 59(3), 434455. Pew Research Center for People and the Press (2004). News audiences increasingly politicized. Retrieved April 8, 2010 from http://people press.org/report/?pageid=834 Pew Research Center for People and the Press (2005). The state of the news media, 2005: An annual report on American journalism. Retrieved April 8 2010 from http://www.stateofthemedia.org/2005/ Pew Research Center for People and the Press (2006). Online papers modestly boost newspaper readership: Maturing Internet news audience broader than deep. Retrieved July 14, 2010 from http://people press.org/report/282/online papers modestlyboost newspaper readership Pew Research Center for People and the Press (2008). Key news audiences now blend online and traditional sources. Retrieved April 10, 2010 from http://people press.org/report/444/news media Pew Research Center for People an d the Press (2009). Press accuracy rating hits two decade low. Retrieved April 8, 2010 from http://people press.org/report/543/ Pew Research Center for People and the Press (2010). The state of the news me dia, 2005: An annual report on American journalism. Retrieved April 8, 2010 from http://www.stateofthemedia.org/2010/ President map Election Results 2008 (2008, December 09). The New York Times. Retrieved from http://elections.nytimes.com/2008/results/president/map.html
111 Public Policy Polling (2009, March 25). Cable news networks mostly favorable among N.C. voters. Retrieved May 20, 2010 from http://www.publicpolicypolling.com/pdf/surveys/2009_Archives/PPP_Release_NC_325. pdf Public Policy Polling (2010, January 26). Fox the most trusted name in news? Retrieved May 20, 2010 from http://www.publicpolicypolling.com/pdf/PPP_Release_National_126.pdf Reiss, I. L. (1967). The social context of premarital sexual permissiveness New York: Holt, Rinehart and W inston. Retrieved from http://www2.huberlin.de/sexology/Reiss2/index.htm Riffe, D. (1980). Relative credibility revisited: How 18 unnamed sources are rated. Journalism Quarterly, 57 (4), 618623. Sumter, S., Bokhorst, C., Steinberg, L., & Westenberg, P. M. (2009). The developmental pattern of resistance to peer influence in adolescence: Will the teenager ever be able to resist? Journal of Adolescence, 32(4), 10091021. Schweitzer, D., & Ginsburg, G. (1966). Factors of communicator credibility. In C. W. Backman & P. F. Secord (Eds.), Problems in social psychology (pp. 94102). New York, NY: McGraw Hill Book Company. Sharma, M., Mende z, A., & O'Byrne, J. (2005). The relationship between attendance in student centred physics tutorials and performance in university examinations International Journal of Science Education, 27(11), 13751389. Shimahara, N. (1983). Socialization in an urban high school. Anthropology and Education Quarterly, 14 (2), 109130. Society of Professional Journalists (1996). SPJ Code of Ethics. Retrieved from http://www.spj.org/ethicscode.asp Steinberg, L., & Monahan, K. (2007). Age differences in resistance to peer influence. Developmental Psychology, 43 (6), 15311543. Stroud, N., & Lee, J. (2008, May). Politics and cable news credibility Paper presented at the International Communication Association Annual Conference, Montreal, Canada. Stroud, N. (2007, May). Revisiting the concept of selective exposure. Paper presented at the International Communication Association Annual Conference, San Francisco, California. Tedin, K. (1980). Peer and parent influence on adolesc ent political attitudes. American Journal of Political Science, 24 (1), 136154. Tehranian, M. (2002). Peace journalism: Negotiating global media ethics. The Harvard International Journal of Press/Politics 7, 5883.
112 Tragesser, S., Aloise Young, P., & Swaim R. (2006). Peer influence, images of smokers, and beliefs about smoking among preadolescent nonsmokers. Social Development, 15(2), 311325. Tsfati, Y., & Cappella, J. (2003). Do people watch what they do not trust? Exploring the association between news media skepticism and exposure. Communication Research, 30, 504529. Tsfati, Y., & Cappella, J. (2005). Why do people watch news they do not trust? The need for cognition as a moderator in the association between news media skepticism and exposure. Media Ps ychology, 7(3), 251271. TV Dimensions 2009 (27th ed.). (2009). Profile of weekly adult reach of selected basic cable program services. Nutley, NJ: Media Dynamics, Inc. United States Census Bureau (2000). USA quickfacts. Retrieved from http://quickfacts.census.gov/qfd/states/00000.html United States House of Representatives (2010). Member website listing (by state). Retrieved from http://www.house.gov/house/MemberWWW_by_State.shtml United States Senate (2010). U.S. Senate: senators home. Retrieved from http://senate.gov/general/contact_informat ion/senators_cfm.cfm University of Florida. (2009). UF Office of Institutional Planning and Research UF Factbook: Enrollment (Final Headcount Enrollment by Class Level, Gender and Ethnicity (1997 2009)). Retrieved from http://www.ir.ufl.edu/factbook/enroll.htm Valliant, P. (1995). Personality, peer influence, and use of alcohol and drugs by first year university students. Psychological Reports, 77, 401402. Walker, M., & Andrade, M. (1996) Confor mity in the Asch task as a function of age. The Journal of Social Psychology, 136(3), 376372. Wanta, W. & Hu, Y. W. (1994). The effects of credibility, reliance, and exposure on media agenda setting: A path analysis model. Journalism and Mass Communication Quarterly, 71(1), 9098. Ward, A. (2009). The impact of knowledge, attitude, and peer influence on adolescent energy drink consumption. Doctoral dissertation from Utah State University, from http://digitalcommons.usu.edu/etd/465 Watts, M., Domke, D., Shah, D., & Fan, D. (1999). Elite cues and media bias in presidential campaigns: Explaining public perceptions of a liberal press. Communication Research, 26(2), 144175. Westley, B. H., & Severin, W. J. (1964). Some correlates of media credibility. Journalism Quarterly, 41, 325335.
113 Whitney, D. C. (1986). The media and the people: Surroundings from two communities. New York: Columbia University, Gannett Center for Media Studies. Wisdom, J., & Agnor, C. (2007). Family heritage and depression guides: Family and peer views influence adolescent attitudes about depression. Journal of Adolescence, 30, 333346. Wouters, E., Larsen, J., Kremers, S., Dagnelie, P., & Geenen, R. (2010). Peer influence on snacking behavior in adolescence. (unpublished but accepted by Appetite, 17). 60 Minutes/Vanity Fair Poll (2010, May 2). Which of the following do you consider to be the most trustworthy source of daily news in the United States? Retrieved May 20, 2010 from http://www.cbsnews.com/stories/2010/04/28/60minutes/main6440363_page9.shtml?tag= contentMain;contentBody
114 BIOGRAPHICAL SKETCH Matthew Beaton was born and raised in Fernandina Beach, Florida. He earned a Bachelor of Arts in Business Administration from Florida Atlantic University, graduating magn a cum laude. He completed his Master of A rts in Mass Communication in December 2010. Having written for numerous publications while in graduate school, he plans to work as a full time journalist.