Docstoc

The Media_ Public Opinion and Iraq

Document Sample
The Media_ Public Opinion and Iraq Powered By Docstoc
					INTRODUCTION Americans rely on the national news media for information about public affairs, current events and political actors. Scholars have examined the news media’s ability to mobilize or shape public opinion, and its effects on social behavior (Bennet 1994, Mondak 1995, West 1991, Volgy and Schwarz 1980). Over the years, scholars have offered theories about the

consequences of news media influence, while many political actors have embraced the news media in order to affect public opinion. If Americans’ views and opinions can be shaped, what potential problems may result if public opinion is shaped by false or inaccurate information? A recent study by Kull, Ramsay, and Lewis (2003) demonstrates that although it is now widely stated that the United States has not found evidence of Iraqi weapons of mass destruction (WMD), and that evidence about preexisting links between Iraq and al qaeda are mixed, and that world public opinion did not support the US-led invasion of Iraq, many Americans hold misperceptions about this evidence (as high as 68% for some misperceptions). The survey data analysis completed by Kull, Ramsey and Lewis finds that while controlling for other variables, including gender, race and age, the respondents’ primary media source is the strongest indicator of misperceptions.1 These startling findings are important because they challenge the assumption of accurate, unbiased, objective reporting in journalism. These findings also suggest greater understanding is needed about the role of the news media in shaping public opinion. The work of Kull, Ramsey and Lewis leads to our question of whether news reports from different media sources have systematically varied in content during the Iraqi War, thus shaping public misperceptions of the situation. We focus our analysis on the influence of an individual’s choice of news source on 1) public perceptions about the existence of weapons of mass

Interestingly, this research finds that respondents reporting FOX news as their primary source of news were most likely to hold these misperceptions.

1

destruction in Iraq, and 2) a pre-existing link between Iraq and al qaeda, through a content analysis of television transcripts and newspaper reports from December 2002 through September 2003. Specifically, we ask: Do different news sources present different information to the public? Do public perceptions of the news depend on the source of news or the individual viewer? Does the amount of coverage given to an issue affect public perceptions? Our findings lend understanding to the significance of news choice in the development of public perceptions and whether different news sources present similar information to the public. In addition, our findings lend to understanding to the question of whether the amount of news coverage given to an issue influences public perceptions. We will first offer a definition of both public opinion and perception, followed by a discussion of the literature associated with the media’s relationship to public perception and the significance of the choice of news source. In the second section of the paper we identify our hypotheses, present our data and operationalize the variables to be included in our model. We discuss our research design and report our findings, then conclude with observations and suggestions for further research. PUBLIC OPINION AND PERCEPTION DEFINED Political scientists offer a plethora of definitions of public opinion, all too numerous to offer here. For the purposes of our study we follow Glynn, Herbst, O’Keefe and Shapiro (1999) who define public opinion as “the formation, communication, and measurement of citizens’ attitudes toward public affairs” (p.17). This definition is appropriate because it describes public opinion in the context of political behavior. Public opinion is most often captured through the use of survey data. Glynn et al. state that public opinion measured in surveys is a result of the information provided to the public from media elites (18-22). This suggests that survey data,

2

while useful and practical, may reflect elite media opinion as much as they reflect public opinion. Public opinion can be influenced by the information individuals receive as well as the relationship between the receiver and the transmitter of the information (Glynn et al, 1999). If an individual develops a close “bond” with a particular news source and relies on that source to transmit accurate information, this primary news source has great potential to influence the individual’s perceptions of public affairs, and more broadly the public’s perceptions. Perceptions are a major part of the public opinion process. Individuals must collect information to develop opinions, yet individuals are limited (both cognitively and physically) in the amount of information they can process. Perceptions can act as a shortcut in the process of opinion development. Glynn et al describe the term perception as “a summary attitude that is based on all of our past and present sensory information” (178). This definition suggests our perceptions are influenced by a variety of environmental factors, including what we see and hear. LITERATURE REVIEW Public opinion research has developed in several directions. Some research focuses on the importance of public opinion as an independent variable, for example as a factor in decisionmaking. This research emphasizes the importance of public opinion in influencing the decision making process. Powlick finds public opinion is an increasingly considered factor in foreign policy decisions (1995). Other research has treated public opinion as a dependent variable, for example in terms of the public’s policy preferences (Page, Shapiro and Dempsey 1987). These studies suggest public opinion is an important area of research in terms of understanding the policy process. Public opinion research has also examined the factors that affect the development of, and changes in, public opinion. We are specifically interested in the media’s influence on the

3

development of and changes in public opinion, and contribute to the current literature by investigating the news media’s influence on public perceptions. A wide body of research has examined the factors that affect the development of, and changes in, public opinion. As a whole, these studies indicate the media play an important role in affecting public opinion toward events (Bennet 1994, Peffley, Langley and Goidel 1995), elections (Mondak, 1995), political leaders (West 1991, Bartels 1993, Pan and Kosicki 1994, Joslyn and Ceccoli 1996,) and issue areas (Volgy and Schwarz 1980, Erbring, Goldenberg and Miller 1980, Page, Shapiro and Dempsey 1987, Jordan and Page 1992, Jordan 1993, Bartels 1993, Gilens 1996). Much of this research suggests different types of news sources, as well as the amount of news coverage, may have varying effects on public opinion. Volgy and Schwarz examine the influence of different mediums, in particular television and newsprint, on the public’s knowledge about politics, political attitudes and political activity (1980). These scholars find that the use of television and news print mediums are significant predictors of survey respondents’ familiarity with political figures and political outsiders (1980). Volgy and Schwarz find that those respondents who cite television as their primary news source demonstrate less knowledge of politics, are less politically active, and have a lower sense of political efficacy than those whose primary news source is newsprint. These scholars also find that the choice of network is significant. For example, Volgy and Schwarz find ABC news viewers are more concerned about crime than other respondents and NBC news viewers are the least familiar with outside political figures, suggesting that not only does the type of medium matter, but that differences within a particular medium are also important considerations. Other research further supports the finding that different news media sources can be significant predictors of political knowledge (Robinson and Levy 1986, Bennet 1994). Mondak’s analysis of

4

the theory of print superiority finds exposure to a major local newspaper does not enhance public knowledge of national politics, but it does contribute to a self-perception of knowledge about local political campaigns (1995). Scholars have demonstrated the news media can influence the public’s knowledge about politics and political figures, but does the news media influence public opinions and attitudes? Research that examines the impact of various news sources on public opinion has produced mixed results. West’s (1991) analysis of public opinion uses panel data to examine the influence of television content on presidential popularity. West finds evidence that television news coverage harmed public evaluations of the president’s performance in 1980 and 1984. However, Kernell’s analysis of the presidential strategy of “going public”, suggests the news media is one of the president’s most important tools for maintaining popularity with the public (1997). Ovsiovitch’s analysis of public opinion toward policy examines differences in the content of news coverage of human rights issues and finds the New York Times, Time Magazine and the CBS Evening News present similar information to the public (1993). However, Page, Shapiro and Dempsey find that different news sources have varying effects on the policy preferences of the public (1987). For example, these authors find that news commentators and experts have an important impact on public opinion. Jordan’s analysis of the media and policy preferences of the American public similarly concludes that different news sources have different impacts on public opinion (1993). Jordan’s analysis also notes the importance of experts and commentators in wielding influence. These findings suggest the impact of the news media’s influence varies across news sources. Gilens’ analysis of public perceptions of a specific policy area, poverty, finds both network television news and news magazines portray poor Americans as being

5

African American more often than is really the case (1996). This finding suggests information presented by the news media, even if it is similar across news sources, may not be accurate. Scholars have examined the accuracy of news reports across different mediums. Shelley and Ashkins’ (1981) study compares media images of crime trends to police statistics and finds newspaper reports more accurately represent police statistics than television reports. These findings are especially important because Sheley and Ashkins also find that public opinion about the relative distribution of crimes is more closely related to the media reports than the police presentations. In other words, the public’s perceptions about crime in this study were based largely on media reports that failed to accurately report police statistics. Other research suggests the public believes what it hears, reads, and sees. For example, Robinson and Kohut’s (1988) survey of the credibility of thirty-nine news media organizations and personalities finds the majority of those surveyed (2,104 adults) believe most of what they learn from the press. The authors also find that although the public puts the news media into different credibility “groups”, the groupings are not based between television and print news sources. In other words, the public considers some news sources as more credible than others, but does not divide credibility between television and print forms of news media. Scholars note that credibility is also dependent on the audience. For example, Gunther’s (1992) study of mass media credibility finds that respondents’ perceptions of a news source’s credibility were most strongly predicted by group identification. This finding suggests credibility is more a function of who is using the news rather than the source of the news. If the news media do have an influence on public opinion, and may present inaccurate, biased or misconstrued information, can this influence be circumvented? Several studies find the news media’s influence on public opinion is not unlimited, and is mediated by other factors,

6

including prior knowledge and social networks. Robinson and Levy examine variables that affect public comprehension of major news stories (1986) and find that while exposure to the news media enhances understanding of the news, interpersonal discussion is at least as important a factor in public awareness and understanding of the news. Price and Zaller examine the importance of prior knowledge, along with exposure to the media and interpersonal discussion, in their analysis of the factors that affect news recall (1993). These authors find that newspaper exposure and interpersonal discussion are less predictive than prior knowledge in explaining news recall. Finally, Bartels notes the media’s impact on public opinion is mediated by the political views of the public (1993). Bartels finds the media’s ability to influence public opinion is limited because of these prior opinions held by individuals. Additional literature emphasizes the importance of the amount of news coverage when considering news media effects on public opinion. Gerbner, Gross, Morgan and Signorielli (1984) hypothesize that since television news portrays moderate political views, survey respondents who watch more television will have more moderate political opinions. These authors find that those who watch more television are more likely to identify themselves as moderate. In comparing print and television coverage of presidential elections between 1948 and 2000, D’Alessio and Allen (2000) find there were no biases in the amount of coverage given to the two major party candidates in print media. The authors do find a small amount of bias in the amount of coverage among televised news networks. Their findings demonstrate the amount of coverage between mediums may differ, and televised coverage may contain systematic biases. Harrington (1989) compares content and amount of television coverage by examining national economic conditions and network coverage of the economy. This author finds networks devote more coverage (measured in seconds) to bad economic news, which suggests the amount of

7

coverage may influence public opinion, and the amount of coverage may depend on the type of news story. In light of these findings, we also examine the total amount of time our news sources devoted to covering stories about Iraq and possible WMD, and stories addressing a link between Iraq and al qaeda. STATEMENT OF PURPOSE AND HYPOTHESES The purpose of our analysis is to test the influence of an individual’s choice of news source on public perceptions about the existence of 1) weapons of mass destruction in Iraq and 2) a pre-existing relationship between Iraq and al qaeda. Our findings lend understanding to the significance of news choice in the development of public perceptions. Specifically, we ask: Do different news sources present different information to the public? Do public perceptions of the news depend on the source of news or the individual viewer? Does the amount of coverage given to an issue affect public perceptions? The previous literature review suggests the influence of the media on public perceptions may depend on a variety of variables including demographic factors specific to the individual respondent, prior beliefs of the respondent, the choice of news source, the tone of news reported by the news source, and the amount of attention given to an issue by a news source. In the following sections we present our hypotheses and present a research design that describes the influence of the choice of news source, along with other variables, on the public’s perception of the existence of WMD in Iraq and a link between Iraq and al qaeda. Our first hypothesis relates to the demographic variables included in our models: party identification, education level, race, gender, age, income level, and region of residence. Because the Bush administration has continuously defended its position about evidence of Iraqi WMD, those who identify with the Republican party may be more likely than Democratic identifiers to accept the administration’s claims, regardless of the news media’s reports. Thus we predict an

8

individual’s party identification will influence the likelihood of believing 1) evidence of Iraqi WMD have been found, 2) a link exists between Iraq and al qaeda. We expect an individual’s level of education to have a negative impact on his/her likelihood of believing that WMD have been found, with those obtaining higher levels of education being less likely to be in possession of incorrect information. Race is a demographic variable that is often included in models of public opinion (Gilens 1996, Mendelberg 2001). Racial minorities are more likely to support the Democratic party. We suggest race may also influence the likelihood of believing that Iraqi WMD have been found. Previous research also indicates the political behavior of men may be different from the political behavior of women (Kaufmann and Petrocik 1999, Dolan 2004). For example, women are more likely to identify themselves as Democrats than Republicans. Due to these prior findings, gender may also influence perceptions about 1) the discovery of WMD in Iraq, and 2) a link between Iraq and al qaeda. The previous literature review indicates individuals become more “politically socialized” as they age. Younger respondents may not have as much interest in, or knowledge about, public affairs. Thus, younger respondents may be more susceptible to the news media’s influence than older respondents, and may be more likely to believe that 1) Iraqi WMD have been found, and 2) there is a link between Iraq and al qaeda. Finally, prior research suggests demographic variables, including income and region of residence influence public perceptions. Thus, demographic characteristics may have an independent influence on whether or not a respondent perceives WMD have been found in Iraq and a clear link exists between Iraq and al qaeda. In sum, in terms of demographic variables we hypothesize:

9

HYPOTHESIS ONE: Demographic variables may affect the influence of the news media on individual perceptions. Respondents who are male, white, identify with the Republican party, have obtained a higher level of education and income level, or reside in the southern region of the US will be more likely to believe 1) WMD have been found in Iraq, and 2) a link exists between Iraq and al qaeda than respondents who are female, nonwhite, non-Republican, have lower levels of education and income, or live outside the southern region of the US. Individual who responded to the PIPA survey were asked if they planned to vote for President Bush in the 2004 election. We suggest respondents who planned on voting for President Bush may be more likely to believe WMD have been found in Iraq and a link exists between Iraq and al qaeda. Individuals who participated in the PIPA survey were also asked if going to war was the right thing to do. We suggest respondents who feel going to war was the right thing to do may be more likely to believe WMD have been found in Iraq and a link exists between Iraq and al qaeda. In sum, we hypothesize: HYPOTHESIS TWO: An individual’s support for the war in Iraq and feelings toward President Bush will affect the influence of the news media on individual perceptions. Respondents who support the war in Iraq and planned to vote for President Bush will be more likely to believe 1) WMD have been found in Iraq, and 2) a link exists between Iraq and al qaeda than respondents who do not support the war in Iraq or did not plan to vote for President Bush The final three hypotheses relate to differences in perceptions based on the choice of news source and the information provided by the news source. Broadly, we expect the choice of news source and the amount of attention devoted to the issues at hand by the various news sources to influence perceptions about the existence of Iraqi WMD and a link between Iraq and al qaeda. Specifically, we hypothesize: HYPOTHESIS THREE: The choice of news source will influence whether respondents believe WMD have been found in Iraq or a link exists between Iraq and al qaeda. Based on findings in the PIPA survey, respondents whose primary news source is FOX News will be more likely to believe 1) WMD have been found in Iraq, and 2) a link exists between Iraq and al qaeda than respondents who report receiving news from a different primary news source.

10

HYPOTHESIS FOUR: The tone of news will vary by news source. Based on findings in the PIPA survey, we expect the tone of Fox News will be more indicative of 1) WMD being found in Iraq, and 2) a link between Iraq and al qaeda than the tone of other news sources. HYPOTHESIS FIVE: The amount of coverage a news source gives to an issue may influence individual perceptions of whether 1) WMD have been found in Iraq, and 2) a link exists between Iraq and al qaeda. We expect greater amounts of news coverage will lead to fewer misperceptions among individual respondents DATA AND METHODS Pooled survey data from the Program on International Policy Attitudes’ (PIPA) 2003 study on Americans’ misperceptions on the Iraq war is combined with a content analysis of news broadcasts and newspaper articles to assess the impact of variables that may influence the likelihood of an individual believing 1) WMD have been found in Iraq, and 2) a link exists between Iraq and al qaeda.2 The PIPA survey was conducted from June through September of 2003. Randomly selected adult respondents were interviewed in order to investigate the misperceptions the public holds about the war in Iraq, as well as the role of the media in the formulation of these misperceptions. Our analysis focuses on the portion of the survey that asks respondents about the existence of WMD in Iraq. In particular, we are interested in the public’s perceptions of whether WMD have been found in Iraq and the link between Iraq and al qaeda. In order to extend Kull et al’s work and investigate the possibility of bias in news broadcasts, we conduct a content analysis of news transcripts related to Iraq, WMD and al qaeda. We examine news stories reported by the television networks used in the PIPA survey (ABC, NBC, CBS, CNN, FOX) and articles from a national newspaper, USA Today, from December 1, 2002 through August 31, 2003. In addition, we count the number of stories covering 1) WMD and Iraq, and 2) Iraq and al qaeda reported during this period by each of the six news sources.
The authors conducted the content analysis using Lexis-Nexus and the key words Iraq, weapons of mass destruction, WMD, al qaeda, Osama Bin Laden and the title of the relevant news program (for example The NBC Nightly News).
2

11

To conduct the content analysis, we examine news stories on 1) Iraq and WMD, and 2) Iraq and al qaeda that were reported from 6:00 to 7:00 p.m. (Eastern Standard Time) on Monday through Friday from December 1, 2002 through August 31, 2003. 3 This time slot was selected since it is when all the major news networks, including the cable news networks, present national news programs. We selected USA Today as our national newspaper to compare print media with televised media. USA Today is circulated throughout the United States, and its readership may be more representative of the public than other national newspapers, such as the New York Times or the Washington Post. We exclude PBS and National Public Radio (NPR) from our analysis because a small number of respondents in the PIPA Survey indicate that either source is their primary news source (57 respondents). Graph 1 presents the total number of stories on Iraq and WMD that were reported for each news source from December 1, 2002 through August 31, 2003. Insert Graph 1 Here For the most part, with the exceptions of CNN and ABC, the monthly amount of coverage the news media devoted to Iraqi WMD follows similar trends. In the months leading up to Secretary of State Colin Powell’s February presentation to the United Nations, news media sources covered the topic of Iraqi WMD quite heavily, averaging 15 to 25 stories per month. Many of the stories focused on the Bush Administration’s claims of “smoking gun” evidence that Iraq possessed weapons of mass destruction as well as an active nuclear weapons program. Interestingly, ABC’s coverage drops during this period and steadily decreases until May. Also

The evening news broadcasts examined were: NBC Nightly News, ABC World News Tonight, CBS Evening News, Wolf Blitzer Reports (CNN), and the Big Story With John Gibson (FOX). Saturdays and Sundays were excluded from the analysis for two reasons: 1) fewer people watch news programs during the weekend, and 2) the national newspaper we examined, USA Today, does not publish on the weekends

3

12

noteworthy is CNN’s sharp increase in coverage from December’s 20 stories to over 35 in the beginning of February. Once the war began, coverage of Iraqi WMD dropped among all networks, mainly because coverage shifted to conflict updates such as battles, captures and casualties. The drop reaches its lowest point in May, when the Bush administration announced the end of major combat operations. During this period the average monthly coverage was five stories. However, as US and other officials began to again question evidence of the existence of Iraqi WMD, news coverage began to rise again, particularly with FOX news. In July the Bush administration reported that evidence of Iraqi WMD would most likely not be found. Interestingly, the amount of news coverage during this period increased only slightly. News coverage of Iraqi WMD reached its lowest levels in August. Graph 2 presents the total number of stories on Iraq and al qaeda that were reported for each news source during the same period. Insert Graph 2 Here At first glance it appears that the amount of news coverage focusing on links between Iraq and al qaeda mirror that of WMD stories, with a high peak in the earlier months of analysis, a sharp decrease in once the war in Iraq begun in March through the end of combat in May, and ending with a slight increase in the summer months. However, when examining the quantity of these stories, we see that the media’s attention to a potential link was about one-third of that to their attention on Iraq WMD. Furthermore, while there was never a month where any given media source failed to report an Iraq WMD story, there are six months where at least one source never had a story on potential links between Iraq and al qaeda.

13

In examining the media sources individually, we see that one source, CNN, is consistently amongst the sources that present the largest quantity of Iraq/al qaeda link stories. CBS, ABC, and USA Today are the three sources that consistently present the fewest stories, while NBC and FOX tend to fluctuate amongst the two groups. However, it is important to note that while both graphs 1 and 2 give us some indication of media coverage on both issues, these graphs alone tell us little about what type of stories are being presented, to which we now turn. In conducting the content analysis of the tone of coverage we employ guidelines to identify whether the news story suggests 1) WMD have been found, or 2) a link exists between Iraq and al qaeda.4 Each news story is coded as “1” if the report presents information that clearly suggests no Iraqi WMD have been found or there is no link between Iraq and al qaeda; “2” if there is conflicting (“mixed”) reports of whether or not Iraqi WMD have been found, or whether there is a link between Iraq and al qaeda; and “3” if the report clearly presents the view that WMD have been found or there is a link between Iraq and al qaeda.5 After coding each report, the daily average tone of the news coverage was calculated using the formula Yi = (Ai*3) + (Bi*2) + (Ci*1) / Di Where Y = Average Tone for day i A = Number of stories that day clearly presenting the view that WMD have been found or there is a link between Iraq and al qaeda B = Number of stories that day presenting conflicting reports C = Number of stories that day presenting information that clearly suggests the view that WMD have not been found or that there is no link between Iraq and al qaeda D = Total sum of all stories in day i Graph 3 presents the results of the tone of coverage for each news source toward Iraqi WMD during the time period previously identified.
See Appendix A for coding guidelines used. To ensure intercoder reliability, articles were selected at random to be reviewed and coded by others and the results were compared with the previous codes. Conflicting scores were decided upon in meetings between the authors. The overall intercoder reliability rate ranged from 83 to 96 percent.
5 4

14

Insert Graph 3 Here In reviewing trends in the tone of coverage toward Iraqi WMD between December 2002 and August 2003, if the news media presented information that was neutral and unbiased, we would expect to see the tone scores remain around ‘2’. However, all six sources fluctuate over time and four of the six sources (FOX and ABC are the outliers) follow the same trend. In the months leading up to Secretary of State Colin Powell’s presentation before the United Nations in February, the tone tended to remain around 2.5 (fairly positive), with the exception of Fox, which had the highest tone, and ABC, which swayed more toward a negative score. The tone during this period reflects UN inspectors’ reports of unhelpful Iraqi officials, countless accusations by the Bush administration, and reports that Powell would present “smoking gun” evidence that a nuclear program existed in Iraq. Once Powell presented the US case before the United Nations, USA Today and FOX jumped toward tone ratings of 2.7. The tone of the other news sources also rose during this period, but not as high. ABC’s tone, while rising, remained the lowest of the sources. The trend among all the news sources remained steady until April, when the tone of all of the news sources dropped to a level equal or slightly less than neutral. This trend reflects that as major combat operations were ending, and no WMD were found, the news media began to question whether evidence ever existed. Interestingly, FOX news stories continued to possess a positive tone, which actually increased during this time period. In May, the tone levels of all of the news sources (with the exception of ABC) reached the levels of FOX news. This trend accompanied the capture of Rahib Tah al-Azawi al-Tikriti (“Dr. Germ”), who many believed would admit to an Iraqi nuclear weapons program. ABC’s tone did rise during this period, but it never surpassed the “neutral” or “mixed” mark, suggesting that while other sources were quick to assert that Iraqi WMD would be found, ABC remained

15

more cautious in its reporting. Once many realized that “Dr. Germ” would remain loyal to her claims that no such evidence existed, questions rose again and we see a sharp decline from May until July. In July tone scores reached their lowest levels, particularly with USA Today and CBS, when the administration announced that evidence of Iraqi WMD would not be found. Note Fox still remained at a “mixed/neutral” score, indicating that while news sources reported the administration’s admission, FOX still presented mixed signals through the remainder of the period under analysis. In August all sources, especially ABC, returned to a “mixed/neutral” score when the Bush administration began to refute criticisms of justifying the war based on perceptions of Iraqi WMD, by defending its stance and pointing out that Iraq once had WMD and may have removed evidence before inspections and the war. During this same time period Ali Hassan al-Majid (“Chemical Ali”), who was believed to have information about an Iraqi nuclear weapons program, was also captured. In terms of the tone of coverage related to Iraq and WMD, FOX and ABC are the outliers. During the period of analysis, FOX coverage never dropped below a neutral score and had tone scores over 2.5 (positive) a majority of the time. This indicates FOX most consistently presented “positive” Iraqi WMD stories, even when other sources reported the opposite. ABC, in contrast to FOX (with the exception of January and February) remained relatively neutral compared to other news sources and consistently remained in the tone range of 1.5-1.7 for the majority of the period of analysis. This may suggest ABC continuously challenged beliefs that Iraqi WMD evidence existed. Graph 4 presents the results of the tone of coverage for each news source toward a link between Iraq and al qaeda.

16

Insert Graph 4 Here A stark contrast from the tone of news for WMD stories is apparent by viewing the sharp spikes throughout graph 4. Once again, we see the fact that there were months where at least one source failed to report on possible links between Iraq and al qaeda reflected in the overall tone equating to 0. All six sources fluctuate over time, but many of the sources fail to follow the same trend. In December, only NBC, CBS and FOX presented stories of a potential link between Iraq and al qaeda with the tone hovering around 1.5, which is fairly negative. However, in the months from Powell’s speech to the U.N. through the beginning of the Iraq war, we see that not only do all the media sources report on possible links, but the tone of all media sources (except ABC) hover around 2.5 (fairly positive) to 3.0 (positive). ABC appears to be the lone source exercising caution in reporting a possible link, with the tone ranging from 2.0 (neutral) to 1.5 (fairly negative) until the beginning of the war when it becomes only source with a 3.0 tone. The tone during the period of January through March reflects assertions from the Bush administration that Iraq has harbored and funded al qaeda operatives, and reports that Powell’s presentation to the U.N. would include evidence of a link between Iraq and al qaeda. This trend remained steady until April, when we see extreme fluctuations between all news sources where the tone ranged from 0 (due to no stories being presented) to a 3 (positive). Whereas before we saw that as major combat operations ceased, the news media began to questions whether WMD would ever be found in Iraq, here we see that the news media either reported that a link exists or neglected the subject altogether. During the period of April through June, we see a large inconsistency between news sources. The three major television news networks, ABC, CBS and NBC all tend to follow the same pattern, with tone ranging between 2.5 and 3.0 in April, decreasing in May, then

17

completely decreasing to zero since all three networks failed to discuss the topic in June. However, USA Today and FOX follow the complete opposite pattern with tone averaging 0 in April, and in May we see USA Today’s tone rise to a 3.0 while FOX remains at zero. In June, USA Today joins the other networks at zero while FOX increases from zero to a 1.0. Also note that only CNN continued to report on possible links during this period, and that the tone of these stories hovered around the neutral point (2.0). We suspect what has made this period so

turbulent is that during this time, reports focused on the question of whether or not evidence of Iraqi WMD ever existed since this was the major justification for war, and because no evidence had yet been found, the media begun placing more scrutiny on this issue while ignoring others. In July we see the media begins reporting on the subject once again, with reports from ABC, NBC, and CNN having a negative tone while reports from the remaining sources have either a neutral or slightly positive tone. Finally, in August we see the tones of ABC “jump” from 1.0 to 3.0 and CBS maintain its 2.5 tone score, while the tone of NBC and CNN stories move from a negative tone to a neutral tone. The outliers, USA Today and FOX fail to report on the subject altogether. During this time reports focused on captured Iraqi insurgents with

“suspected” ties to al qaeda, thus “suggesting” the missing link between Iraq and al qaeda, which was classified as “neutral” reports by our coding standards. Other reports noted that although these insurgents had “suspected” ties to al qaeda, it was later established that those suspicions were falsified. It was also during this time that families of 9/11 victims were battling countries suspected of harboring or funding al qaeda operatives in the courts. Iraq was one of the countries named in the compliant and after courts found Iraq could be named as a defendant, conflicting reports suggesting that either this could be interpreted as “evidence” of a link or should still be viewed with caution until the cases further proceeded, had dominated news stories.

18

Given the content of the news stories from December 2002 through August 2003, we suggest public perceptions about the existence of WMD in Iraq and possible links to al qaeda are influenced by factors specific to individual respondents as well as the respondent’s choice of a primary news source. The following section provides our method for operationalizing the variables we use to develop a model of the influences of public perceptions. DEPENDENT VARIABLES Evidence WMD have been found: Our first dependent variable is the perception that WMD have been found in Iraq. The variable is dichotomous, and is coded 1 if the respondent perceives evidence of WMD have been found and 0 if the respondent perceives there is no evidence that evidence of WMD have been found. The data were obtained from the PIPA survey. Link between Iraq and al qaeda: Our second dependent variable is the perception that there is of a link between Iraq and al qaeda. Specifically, we examine individual’s responses to the PIPA survey question, “Has the United States found clear evidence that Iraq was working closely with al qaeda.” The variable is dichotomous, and is coded 1 if the respondent perceives evidence of a link between Iraq and al qaeda have been found and 0 if the respondent perceives there is no evidence of a link between Iraq and al qaeda. EXPLANATORY VARIABLES Choice of News Source: The purpose of this analysis is to examine the influence of the choice of news source on public perceptions about 1) the discovery of WMD in Iraq, and 2) a link between Iraq and al qaeda. The previous literature review indicates different news sources present different types of information to the public. We code the news source identified by respondents as his or her primary news source. The variable is coded as “1” if the news source is

19

USA Today, “2” for NBC, “3” for CBS, “4” for ABC, “5” for FOX, and “6” for CNN. “Multiple news sources” is the baseline category. The data were obtained from the PIPA survey. Tone of News Coverage: The previous literature review suggests different news sources may present different information. We describe the type of information presented as the tone of the coverage. Tone is measured on a weekly basis as a categorical variable. News stories are coded as “3” if the coverage indicates WMD exist or will be found in Iraq, “1” if the coverage indicates WMD have not or will not be found in Iraq, and “2” if the news report contains mixed or contradictory information. News stories are coded as ”3” if the coverage indicates there is a link between Iraq and al qaeda, “2” if the coverage indicates there is no link between Iraq and al qaeda, and “1” if the news report contains mixed or contradictory information. The data were obtained from the content analysis. Amount of News: We code this variable as the number of news stories reported by each news source in a given week. The data were obtained from the content analysis. Party Identification: This is a categorical variable with respondents who identify themselves as Republican coded as “3”, Independents/Others as “2”, and Democrats as “1”. The data were obtained from the PIPA survey. Education Level: Education level is measured on an ordinal scale with each respondent placed into one of four categories based on the highest level of education that the respondent completed: less than high school is coded as “1”, high school is coded as “2”, some college is coded as “3”, and bachelor’s degree or higher is coded as “4”. The data were obtained from the PIPA survey. Race: Respondents who classify themselves as “white” are coded as “1” and respondents who classify themselves as “nonwhite” are coded “0”. The data were obtained from the PIPA survey.

20

Gender: Male respondents are coded as “2” and female respondents are coded as “1”. The data were obtained from the PIPA survey. Age: Respondents’ age is measured in years. The data were obtained from the PIPA survey. Income Level: This variable is measured using an ordinal scale of 19 categories, with the lowest income category coded as “1” and the highest income category coded as “19”. The data were obtained from the PIPA survey. Region of residence: The last demographic variable in our model measures whether a respondent's region of residence is significantly related to his or her perceptions about WMD in Iraq and links between Iraq and al qaeda. Each respondent is placed into one of two categories based on the respondent’s place of residence. Respondents who report living in the South are coded as “1” and other respondents (“non-south”) are coded as “0”. Data were obtained from the PIPA survey. Vote for Bush: Individuals who responded to the PIPA survey were asked if they planned to vote for President Bush in the 2004 election. This variable is measured as a dichotomous variable. Respondents who reported they planned on voting for President Bush in the 2004 election are coded as “1” and respondents who did not plan on voting for President Bush are coded as “0”. War in Iraq was Right: Finally, individuals who participated in the PIPA survey were asked if going to war was the right thing to do. Respondents were coded in one of four categories: “1” if respondent felt “war was the right decision and the best thing to do”, “2” if respondent felt “war was the right decision and support the president”, “3” if respondent felt “war was the right decision but don’t know if it was the best thing to do” and “4” if respondent felt war was the “wrong decision.” This variable is not measured as a dichotomous variable since other variables

21

may influence the different categories. For example, those who intended to vote for President Bush may have been more likely to pick category “2” than the other categories. ANALYSIS We develop a series of models to better understand the factors that influence public perceptions. Before we analyze the impact of each news source’s content and total coverage on the likelihood an individual will believe 1) Iraqi WMD have been found, and 2) there is a link between Iraq and al qaeda, we first examine the extent to which traditional demographic variables including party identification, age, gender, race, education level, income and region of residence influence these perceptions. Our first model, labeled “Demographic Model WMD” is specified as follows: Evidence Iraqi WMD have been found = + 1 Party Identification + 2 Age + 3 Gender + 4 Race + 5 Education Level + 6 Income Level + 7 Region of Residence + Insert Table 1 Here Table 1 reports the influence of demographic variables on individuals’ perceptions of the existence of Iraqi WMD. The results indicate party identification, age and education level each has a significant influence on the likelihood an individual believes Iraqi WMD have been found. Specifically, the results indicate party identification, age, and education all have a negative relationship with the dependent variable. This shows that Republicans, younger people, and less educated people are significantly more likely than others to believe weapons of mass destruction had been found in Iraq. The results of our model suggest race, gender, income level and region of residence do not have a statistically significant impact on perceptions of WMD in Iraq. In sum, the results of the first model support our hypothesis that demographic variables do have an

22

independent influence on public perceptions. Our second model, labeled “Demographic Model al qaeda”, is specified as follows: Evidence of a link between Iraq and al qaeda = + 1 Party Identification + 2 Age + 3 Gender + 4 Race + 5 Education Level + 6 Income Level + 7 Region of Residence + Insert Table 2 Here Table 2 reports the influence of demographic variables on individuals’ perceptions of a link between Iraq and al qaeda. The results indicate party identification, age, race and education level each have a significant influence on the likelihood an individual believes there is a link between Iraq and al qaeda. For example, respondents who identified themselves as Republican are more likely to believe in the link than other respondents. Despite the significant findings, the R-squared indicates the model has poor explanatory power. We suspect other factors, beyond demographics, contribute to perceptions about Iraqi WMD and the link between Iraq and al qaeda. The third logit model explores the extent to which an individual's support for the war in Iraq, or support for President Bush, influences his or her beliefs about finding WMD in Iraq. This model includes, in addition to the variables in the previous model, two dichotomous variables: the intent to vote for President Bush if the election were the day of the survey and the belief that the war in Iraq was the “right thing to do.” Our third model, labeled “Support for Bush and War in Iraq: WMD” is specified as follows: Evidence Iraqi WMD have been found = + 1 Party Identification + 2 Age + 3 Gender + 4 Race + 5 Education Level + 6 Income Level + 7 Region of Residence + 8 Vote for Bush + 9 War in Iraq was right + Insert Table 3 Here The results reported in Table 3 indicate that when we add an individual’s intention to vote for Bush and support of the Iraq war, two of the previously significant variables are no

23

longer statistically significant, although education remains close to statistical significance at the .05 level. The effects of party identification may be captured in an individual’s intention to vote for President Bush. The results in Table 3 also indicate that when we add the intention to vote for Bush and support for the Iraq war variables, the influence of gender, race and region of residence becomes statistically significant. It appears that minorities, men, and nonsoutherners are all more likely to believe weapons of mass destruction had been found in Iraq than whites, women, and southerners. The impact of both of the two new variables (intent to vote for Bush and support of the war in Iraq) is statistically significant and in the expected direction. Individuals who reported the intent to vote for Bush or support for the war in Iraq were more likely to believe Iraqi WMD have been found. The fourth logit model explores the extent to which an individual's support for the war in Iraq, or support for President Bush, influences his or her beliefs about the link between Iraq and al qaeda. This model also includes the intent to vote for President Bush and the belief that the war in Iraq was the “right thing to do.” Our fourth model, labeled “Support for Bush and War in Iraq: al qaeda” is specified as follows: Evidence of Link between Iraq and al qaeda = + 1 Party Identification + 2 Age+ 3 Gender+ 4 Race + 5 Education Level + 6 Income Level + 7 Region of Residence + 8 Vote for Bush + 9 War in Iraq was right + Insert Table 4 Here The results reported in Table 4 indicate that when we add an individual’s intention to vote for Bush and support of the Iraq war, the impact of the demographic variables changes. Party identification, age, and race lose statistical significance, while income, and gender reach statistical significance in the expected direction. Education is statistically significant, but not in the expected direction: higher levels of education in this model seem to increase the likelihood of

24

believing there is a link between Iraq and al qaeda. Both new variables are statistically significant and in the expected direction. Moreover, the R-squared of .20 indicates, as expected, these variables should be included in the model. Our final set of models examine the influence of the choice of primary news source on the likelihood individuals’ perceptions of whether 1) WMD have been found in Iraq and 2) a link exists between Iraq and al qaeda. The fifth logit model includes, in addition to the variables in the previous models, a variable to measure respondents’ choice of primary news source. Our fifth model, labeled “Primary News Source Model: WMD,” measures the influence of the variables on the belief WMD have been found in Iraq. The model is specified as follows: Evidence Iraqi WMD have been found = + 1 Party Identification + 2 Age + 3 Gender + 4 Race + 5 Education Level + 6 Income Level + 7 Region of Residence + 8 Vote for Bush + 9 War in Iraq was right + 10 USA Today + 11 NBC+ 12 CBS + 13 ABC + 14 FOX + 15 CNN+ Insert Table 5 Here The results reported in Table 5 indicate all of the statistically significant variables from Table 3 remain significant with the inclusion of the primary news source variables. In addition, the results indicate a statistically significant relationship between the choice of FOX news as the primary news source and perceptions about Iraqi WMD: respondents who reported FOX as their primary news network were more likely to believe Iraqi WMD have been found. The sixth logit model measures the influence of choice of primary news source on perceptions about the link between Iraq and al qaeda. Our sixth model, labeled “Primary News Source Model: al qaeda,” is specified as follows: Evidence of Link between Iraq and al qaeda = + 1 Party Identification + 2 Age + 3 Gender + 4 Race + 5 Education Level + 6 Income Level + 7 Region of Residence + 8 Vote for Bush + 9 War in Iraq was right + 10 USA Today + 11 NBC+ 12 CBS + 13 ABC + 14 FOX + 15 CNN+

25

Insert Table 6 Here The results reported in Table 6 indicate gender, education and belief the war in Iraq was the right choice each remain statistically significant in influencing respondents’ perceptions of the link between Iraq and al qaeda. In addition, the choice of FOX news as the primary news source has an independent influence on respondents perceptions about the link: respondents reporting FOX news as their primary source of news were more likely to believe in the existence of a clear link between Iraq and al qaeda. More broadly, the results from Tables 5 and 6 suggest there is systematic variation in perceptions based on the choice of news source. While our results indicate the choice of news source influences perceptions of the existence of Iraqi WMD and perceptions about the relationship between Iraq and al qaeda, we must examine whether it is the tone of the coverage that is driving these perceptions. Model seven employs Ordinary Least Squares regression to measure whether or not there is a significant relationship between the news sources and the tone of the news coverage. In this model, the tone of the news coverage related to Iraqi WMD is the dependent variable, and the different news sources represent the explanatory variables. Our seventh model, labeled “Tone of News Source: WMD” is specified as follows: Tone of Coverage WMD= + 1 USA Today + 2 NBC + 3 CBS + 4 ABC + 5 FOX + 6 CNN + Insert Table 7 Here The results reported in Table 7 indicate that the tone of the coverage of Iraqi WMD does systematically vary among news sources. For example, the coefficients for the tone of coverage range from –0.202 for NBC to 0.093 for ABC, all of which are statistically significant. Model eight also employs OLS regression to measure the relationship between news sources and the

26

tone of news coverage related to a link between Iraq and al qaeda. . Our eighth model, labeled “Tone of News Source: al qaeda” is specified as follows: Tone of Coverage al qaeda= + 1 USA Today + 2 NBC + 3 CBS + 4 ABC + 5 FOX + 6 CNN + Insert Table 8 Here The results reported in Table 8 indicate that the tone of the coverage of the relationship between Iraq and al qaeda does systematically vary among news sources. For example, the coefficients for the tone of coverage range from –0.447 for USA Today to 0.494 for CNN. Once again, all the variables are statistically significant. In sum, the results from Tables 7 and 8 demonstrate that the tone of coverage varies among news sources. Based on these results, we expect that the varying tones of news sources will impact an individual’s perceptions on issues. However, in order to better understand the relationship between news sources and individual perceptions of Iraqi WMD and a link between Iraq and al qaeda, we must include a measure for the total amount of news coverage. We suspect the amount of news coverage respondents consume may influence perceptions independent of the tone of news coverage. For example, if viewers are continuously bombarded with information that suggests Iraqi WMD have been found, we might expect viewers to be more likely to hold this misperception than individuals who have only viewed a few stories suggesting Iraqi WMD have been found. Our final two models predict the extent to which the tone and the total amount of news coverage influence respondents’ perceptions. These models include the variables presented in the previous models, and variables to measure tone and amount of news coverage. The ninth model, labeled “Media Effects Model: WMD” is specified as follows:

27

Evidence Iraqi WMD have been found = + 1 Party Identification + 2 Age + 3 Gender + 4 Race + 5 Education Level + 6 Income Level + 7 Region of Residence + 8 Vote for Bush + 9 War in Iraq was right + 10 USA Today + 11 NBC+ 12 CBS + 13 ABC + 14 FOX + 15 CNN + 16 Tone of news coverage + 17 Total amount of stories reported + Insert Table 9 Here The results reported in Table 9 indicate the amount of news coverage has a significant influence (at .07) on perceptions of Iraqi WMD, independent of the influence of the other included variables. In addition, a number of the demographic and support variables remain statistically significant. Interestingly, when amount of coverage is included in the model, the tone of the coverage is no longer statistically significant. This finding suggests the more news coverage an individual is exposed to, the less likely he or she is to believe WMD have been found in Iraq, regardless of the tone of coverage. Perhaps as individuals are exposed to more news coverage, they increase their chances of viewing the issue from all perspectives and are more aware of both sides of the story. The final model in our analysis predicts the extent to which the total amount of news coverage influences respondents’ perceptions of the link between Iraq and al qaeda. The final model, labeled “Media Effects Model: al qaeda,” is specified as follows: Evidence of Link between Iraq and al qaeda = + 1 Party Identification + 2 Age + 3 Gender + 4 Race + 5 Education Level + 6 Income Level + 7 Region of Residence + 8 Vote for Bush + 9 War in Iraq was right + 10 USA Today + 11 NBC+ 12 CBS + 13 ABC + 14 FOX + 15 CNN + 16 Tone of news coverage + 17 Total amount of stories reported + Insert Table 10 Here The results reported in Table 10 indicate neither the amount of news coverage nor the tone of news coverage have a statistically significant influence on perceptions of the link between Iraq and al qaeda, independent of the influence of the other included variables. In this

28

model, the choice of USA Today, NBC or ABC news, as well as support for President Bush and the war, have a statistically significant influence on perceptions about the link between Iraq and al qaeda. In addition, gender and education continue to have an independent influence on this perception. It is possible that the poor explanatory power of tone and amount of news in this model is related to the range in tone of the coverage of the link between Iraq and al qaeda: perhaps a larger range of views were reported about this topic than Iraqi WMD, which diluted the impact of the overall tone and amount of coverage on respondents. DISCUSSION The purpose of this analysis has been to empirically test whether the news media’s coverage of 1) the search for WMD in Iraq, and 2) links between Iraq and al qaeda has influenced public perceptions. Kull et al. demonstrates that an individual’s primary media source has an influence on perceptions. We extend the findings of Kull et al. by examining the actual content of the news stories. Our findings indicate the choice of news source can have a significant influence on public perceptions about Iraqi WMD and links between Iraq and al qaeda. Moreover, we find the amount of exposure an individual has to news coverage can have an even stronger influence on these perceptions. Overall, it appears variance in news coverage does play a strong role in influencing public perceptions about WMD in Iraq and a link between Iraq and al qaeda. Previous research suggests the news media’s influence is dependent on the individual viewer. Our findings support this perspective. Our analysis indicates factors specific to the individual viewer influence perceptions. For example, the influence of gender and level of education remain statistically significant throughout the analysis. Overall, our results suggest the choice of news source varies in importance by the type of individual making the choice.

29

However, given the differences in tone and amount of coverage among the news sources, especially with FOX news, we conclude an individual’s choice of news source, and the amount of attention paid to an issue by a news source should be included in models that seek to explain the influence of the news media on public perceptions. Finally, future research might empirically examine the role of the news media in the formation of other individual perceptions about the war in Iraq. For example, perceptions about the war in Iraq are not limited to questions of WMD and links with al qaeda. The choice of news source likely affects other perceptions including, for example, perceptions of how successful U.S. efforts have been in Iraq. Future research might also apply our model to other issue areas. For example, are tone and amount of coverage as important in explaining perceptions about domestic issues as they are in explaining perceptions in Iraq? We suspect the choice of news source may be quite important with especially salient domestic issues, such as a presidential election. Finally, future research might further examine differences in coverage among various types of news sources. Does broadcast media have an identifiably different influence on perceptions than print media? Clearly scholars may take many different directions in better understanding the role of choice of new coverage in influencing public perceptions.

30

BIBILOGRAPHY Bartels, Larry M. 1993. “Messages Received: The Political Impact of Media Exposure.” The American Political Science Review. 87:267-285. Bennett, Stephen Earl. 1994. “The Persian Gulf War’s Impact on Americans’ Political Information.” Political Behavior. 16:179-201. D’Alessio, D. and M. Allen. 2000. “Media Bias in Presidential Elections: A Meta-Analysis.” Journal of Communication. 50:133-156. Erbing, Lutz, Edie N. Goldenberg and Arthur H. Miller. 1980. “Front-Page News and RealWorld Cues: A New Look at Agenda Setting by the Media.” American Journal of Political Science. 24:16-49. Gerber, George, Larry Gross, Michael Morgan and Nancy Signorielli. 1984. “Political Correlates of Television Viewing.” The Public Opinion Quarterly. 48:283-300. Gilens, Martin. 1996. “Race and Poverty in America: Public Misperceptions and the American Correlates of Television Viewing.” The Public Opinion Quarterly. 48:283-300. Glynn, Carroll, Susan Herbst, Garrett J. O’Keefe and Robert Shapiro. 1999. Public Opinion. Boulder, Colorado: Westview Press. Gunther, Albert C. 1992. “Biased Press or Biased Public? Attitudes Toward Media Coverage of Social Groups.” The Public Opinion Quarterly. 56:147-167. Harrington, David E. 1989. “Economic News on Television: The Determinants of Coverage.” The Public Opinion Quarterly. 53:17-24. Jordan, Donald L. 1993. “Newspaper Effects on Policy Preferences.” The Public Opinion Quarterly. 57:191-204. Jordan, Donald L. and Benjamin I. Page. 1992. “Shaping Foreign Policy Opinions: The Role of TV News.” The Journal of Conflict Resolution. 36:227-241. Joslyn, Mark R. and Steve Ceccoli. 1996. “Attentiveness to Television News and Opinion Change in the Fall 1992 Presidential Campaign.” Political Behavior. 18:141-170. Kernell, Samuel. 1993. Going Public. 2nd Edition. Washington D.C.: Congressional Quarterly Press. Kull, Steven, Clay Ramsey and Evan Lewis. 2003. “Misperceptions, the Media, and the Iraq War.” Political Science Quarterly. 118:2003-2004.

31

Mondak, Jeffery J. 1995. “Newspapers and Political Awareness.” American Journal of Political Science. 39:513-527. Ovsiovitch, Jay S. 1993. “News Coverage of Human Rights.” Political Research Quarterly. 46:671-689. Page, Benjamin I. Robert Y. Shapiro and Glenn R. Dempsey. 1987. “What Moves Public Opinion?” The American Political Science Review. 81:23-44. Pan, Zhongdang and Gerald M. Kosicki. 1994. “Voters’ Reasoning Processes and Media Influence During the Persian Gulf War.” Political Behavior. 16:117-156. Peffley, Mark, Ronald E. Langley and Robert Kirby Goidel. 1995. “Public Response to the Presidential Use of Military Force: A Panel Analysis.” Political Behavior. 17:307-337. Powlick, Philip J. 1995. “The Sources of Public Opinion for American Foreign Policy Officials.” International Studies Quarterly. 39:427-451. Price, Vincent and John Zaller. 1993. “Alternative Measures of News Reception and their Implications for Research”. The Public Opinion Quarterly. 57:133-164. Robinson, Michael J. and Andrew Kohut. 1988. “Believability and the Press.” The Public Opinion Quarterly. 52:174-189. Robinson, John P. and Mark R. Levy. 1986. “Interpersonal Communication and News Comprehension.” The Public Opinion Quarterly. 50:160-175. Shelley, Joseph F. and Cindy D. Ashkins. 1981. “Crime, Crime News, and Crime Views.” The Public Opinion Quarterly. 45: 492-506. Volgy, Thomas J. and John E. Schwarz. 1980. “On Television Viewing and Citizens’ Political Attitudes, Activity and Knowledge: Another Look at the Impact of Media on Politics.” The Western Political Quarterly. 33:153-166. West, Darrell M. 1991. “Television and Presidential Popularity in American.” British Journal of Political Science. 21:199-214.

32

Appendix A Coding Rules Employed in Content Analysis Code stories as having weapons of mass destruction or a link with al qaeda if the following information is included: A member of the Bush administration saying that weapons of mass destruction would be found in Iraq / a link between Iraq and al qaeda exists. A member of Congress saying that weapons of mass destruction would be found in Iraq / a link between Iraq and al Qaeda exists. An “expert” such as an academic, book author, etc. saying that weapons of mass destruction would be found in Iraq / a link between Iraq and al-Qaeda exists. A member of the media saying that weapons of mass destruction would be found in Iraq / a link between Iraq and al-Qaeda exists. A statement by any of the above people stating that al-Qaeda operatives have been trained at terror camps in Iraq. Tony Blair saying that weapons of mass destruction would be found in Iraq. Mention of specific pieces of “evidence” that Iraq has WMD, such as acquiring uranium from Africa, US has documents, etc. Weapons inspectors saying that Iraq is not cooperating with the inspection process. Mention of Iraq using weapons of mass destruction in the past. Mention of the United States and countries surrounding Iraq preparing for the possible use of weapons of mass destruction during the war. Mention of the finding of materials used to make weapons of mass destruction in Iraq by weapons inspectors. (Cont. on Next Page)

33

(Cont.)

Mention of the capturing of Iraqi scientists who have been known as participants in the weapons of mass destruction program. Members of Congress saying that weapons of mass destruction will be found in Iraq. A journalist saying that weapons of mass destruction will likely be found in Iraq. Saddam Hussein or a member of the Iraqi government saying that there are no weapons of mass destruction in Iraq / no relationship with al qaeda. Osama bin Laden or member of al qaeda saying that there is no relationship with Iraq. A member of the Bush administration saying that they doubt that weapons of mass destruction will be found in Iraq / a link can be established. Tony Blair saying that weapons of mass destruction will not be found in Iraq. A member of the United Nations saying that weapons of mass destruction will not be found in Iraq / no relationship with al qaeda exists. Weapons inspectors saying that weapons of mass destruction have not been found. A journalist saying that weapons of mass destruction have not been found in Iraq. Members of Congress saying that weapons of mass destruction will not be found in Iraq. A statement is made by any of the above people that Saddam Hussein is not supportive of al qaeda. A statement by any of the above people that indicates that Saddam Hussein was openly unsupportive of terrorist groups such as Iraq.

34

A statement by any of the above people that indicates that there is no evidence that there is a connection in any way between Iraq and al qaeda. Direct quotes from Hussein or members of his administration indicating that Iraq has no ties to the al qaeda organization. Direct quotes from Bin Laden or members of al qaeda indicating that Iraq has no ties to the al qaeda organization. A combination of information from the two categories shown above.

Code the stories as mixed if they contain the following information:

35

Appendix B Survey Questions, Response, Frequencies and Percentages
Frequency If the upcoming election for president were held today, For whom would you vote? Vote for President Bush 1,292 Will not vote for President Bush 1,213 Gender Male 1,620 Female 1,714 Race White 2,617 Nonwhite 717 Income Less than $5,000 87 $5,000-7,499 71 $7,500-9,999 79 $10,000-12,499 103 $12,500-14,999 112 $15,000-19,999 191 $20,000-24,999 255 $25,000-29,999 220 $30,000-34,999 252 $35,000-39,999 300 $40,000-49,999 434 $50,000-59,999 349 $60,000-74,999 365 $75,000-84,999 163 $85,000-99,999 145 $100,000-124,999 108 $125,000-149,999 53 $150,000-174,999 17 $175,000 + 30 Party Identification Republican 1,065 Democrat 1,081 Other 1,188 Education Less than High School 390 High School 1,258 Some College 936 Bachelor’s degree or higher 750 Was the war in Iraq the right decision? Right decision, best thing to do 1,437 Right decision, support the President 481 Right decision, don’t know if best thing to do 21 Wrong decision 930 Has the US found evidence of Iraqi WMD? US Has 568 US Has not 2,115 Has the US found clear evidence that Iraq was working closely with al qaeda? US Has 1,342 US Has not 1,261 (Cont. on Next Page) Percentage

52 48 49 51 78 22 3 2 2 3 3 6 8 7 8 9 13 10 11 5 4 3 2 1 1 32 32 36 12 38 28 23 47 16 .69 31 21 79 54 46

36

(Cont.) How closely are you following the news about the situation in Iraq now? Very closely 516 Somewhat closely 1,495 Not very closely 954 Not closely at all 351 Where do you tend to get most of your news? Newspaper/Magazines 669 NBC 358 CBS 221 ABC 273 FOX 471 CNN 407 NPR/PBS 57 Multiple Source 819

16 45 29 11 20 11 7 8 14 12 2 25

37

Graph 1: Monthly Media Coverage for WMD Stories
40

35 Total Amount of Iraqi WMD Stories

30 CNN CBS 20 NBC ABC FOX 15 USA

25

10

5

0 Dec Jan Feb Mar Apr Month May Jun Jul Aug

Graph 2: Monthly Media Coverage for al qaeda Stories

Total Amount of Al Qaeda Link Stories

12 10 CNN 8 6 4 2 0 Dec Jan Feb Mar Apr
Month

CBS NBC ABC FOX USA

May

Jun

Jul

Aug

38

Graph 3: Monthly Tone of News for WMD Stories
3

2.5

2 CNN CBS Tone 1.5 NBC ABC FOX USA 1

0.5

0 Dec Jan Feb Mar Apr Month May Jun Jul Aug

Graph 4: Monthly Tone of News for al qaeda Stories

3.5 3 2.5
Tone

CNN CBS NBC ABC FOX USA

2 1.5 1 0.5 0 Dec Jan Feb Mar Apr
Month

May

Jun

Jul

Aug

39

Table 1: Demographic Model WMD—Model Specification 1
Dependent Variable: Likelihood of belief that WMD had been found (1=WMD, 0=No WMD) Independent Logit Analysis Variables Results Gender Coefficient SE Z Statistic Race 0.109 0.120 0.91 Income Level -0.015 0.013 -1.15 Party ID -0.235 0.059 -3.98* Age -0.017 0.003 -5.62* Education Level -0.160 0.053 -3.00* Region 0.159 0.101 1.58 Psuedo R-Squared *p 0.05, one-tailed test 0.0248 0.116 0.096 1.21

40

Table 2: Demographic Model al qaeda—Model Specification 2
Dependent Variable: Likelihood of belief there is a link between Iraq and al qaeda Independent Logit Analysis Variables Results Gender Coefficient SE Z Statistic Race 0.402 0.104 3.87* Income Level 0.001 0.011 -0.06 Party ID 0.427 0.051 8.36* Age -0.008 0.003 -3.08* Education Level -0.282 0.046 -6.18* Region 0.120 0.087 1.38 Psuedo R-Squared *p 0.05, one-tailed test 0.0368 0.072 0.082 -0.87

41

Table 3: Support for Bush and War in Iraq: WMD—Model Specification 3
Dependent Variable: Likelihood of belief that WMD had been found (1=WMD 0=No WMD) Independent Logit Analysis Variables Results Party ID Coefficient SE Z Statistic Age -0.015 0.004 -3.90* Gender 0.242 0.119 2.03* Race -0.457 0.155 -2.95* Income Level -0.042 0.017 -2.51* Region 0.125 0.124 1.01 Education Level -0.065 0.065 -1.00 Support for President Bush -0.985 0.159 6.20* Belief War in Iraq Was Right Choice -0.478 0.064 -7.48* Psuedo R-Squared *p 0.05, one-tailed test 0.1323 -0.096 0.087 -1.11

42

Table 4: Support for Bush and War in Iraq: al qaeda—Model Specification 4
Dependent Variable: Likelihood of belief there is a link between Iraq and al qaeda Independent Logit Analysis Variables Results Party ID Coefficient SE Z Statistic Age -0.001 0.003 -0.21 Gender -0.263 0.103 -2.55* Race -0.109 0.132 -0.83 Income Level -0.031 0.014 -2.13* Region 0.129 0.109 1.19 Education Level -0.209 0.057 -3.68* Support for President Bush 0.842 0.127 6.64* Belief War in Iraq Was Right Choice -0.616 0.046 -13.29* Psuedo R-Squared *p 0.05, one-tailed test 0.1978 0.085 0.069 1.23

43

Table 5: Primary News Source Model: WMD—Model Specification 5
Dependent Variable: Likelihood of belief that WMD had been found (1=WMD, 0=No WMD) Independent Variables Coefficient SE Z Statistic Party ID Age Gender Race Education Level Income Level Region Support for Pres. Bush Belief War in Iraq Was Right Choice USA Today NBC CBS FOX CNN ABC Pseudo R-Squared *p 0.05, one-tailed test -0.106 -0.014 0.273 -0.399 -0.072 -0.046 0.104 -0.944 0.090 0.004 0.123 0.159 0.067 0.017 0.127 0.163 -1.18 -3.56* -2.22* -2.50* -1.08 -2.71* 0.82 -5.80*

-0.487 0.032 -0.053 0.176 0.387 0.158 -0.075 0.1377

0.066 0.186 0.222 0.262 0.189 0.220 0.249

-7.40* 0.17 -0.24 0.67 2.05* 0.72 -0.30

44

Table 6: Primary News Source Model: al qaeda—Model Specification 6
Dependent Variable: Likelihood of Belief There is a Link Between Iraq and al qaeda Independent Variables Coefficient SE Z Statistic Party ID Age Gender Race Education Level Income Level Region Support for Pres. Bush Belief War in Iraq Was Right Choice USA Today NBC CBS FOX CNN ABC Pseudo R-Squared *p 0.05, one-tailed test 0.103 0.002 -0.324 -0.027 -0.200 -0.023 0.159 0.706 0.082 0.004 0.122 0.153 0.067 0.017 0.128 0.150 1.25 0.53 -2.67* -0.18 -2.97* -1.40 1.25 4.70*

-0.609 -0.224 0.031 0.405 0.595 0.397 -0.163 0.2044

0.055 0.205 0.230 0.268 0.226 0.229 0.186 -0.88

-10.98* -1.09 0.14 1.51 2.63* 1.73

45

Table 7: Tone of News Source: WMD—Model Specification 7
Dependent Variable: Tone of Networks Independent Variables USA Today Coefficient SE T Statistic NBC

Regression Analysis Results

0.031 0.012 -2.61* -0.202 0.015 -13.87*

ABC -0.093 0.016 5.80* FOX 0.069 0.013 5.19* CNN -0.033 0.014 -2.40* R-Squared Adj. R-Squared *p 0.05, one-tailed test 0.1073 0.1056

46

Table 8: Tone of News Source: al qaeda—Model Specification 8
Dependent Variable: Tone of Networks Independent Variables USA Today Coefficient SE T Statistic NBC

Regression Analysis Results

-0.447 0.024 -18.28* -0.265 0.027 -9.83*

ABC -0.394 0.029 -13.83* FOX -0.409 0.026 -15.90* CNN 0.495 0.021 18.74* R-Squared Adj. R-Squared *p 0.05, one-tailed test 0.5401 0.5391

47

Table 9: Media Effects Model: WMD—Model Specification 9
Dependent Variable: Likelihood of belief that WMD had been found (1=WMD, 0=No WMD) Independent Variables Coefficient SE Z Statistic Party ID Age Gender Race Education Level Income Level Region Support for Pres. Bush Belief War in Iraq Was Right Choice USA Today NBC ABC FOX CNN CBS Tone Total Close Psuedo R-Squared *p 0.05, one-tailed test -0.106 -0.015 0.284 -0.408 -0.090 -0.043 0.112 -0.910 0.090 0.004 0.125 0.161 0.068 0.017 0.128 0.164 -1.17 -3.73* 2.28* -2.53* -1.31 -2.53* 0.88 -5.54*

-0.495 0.022 0.046 -0.102 0.492 0.128 0.133 0.362 -0.154 -0.057

0.067 0.192 0.231 0.253 0.206 0.223 0.267 0.277 0.074 0.077 0.1409

-7.42* -0.11 -0.20 -0.40 2.39* 0.58 0.50 1.31 -2.09* -0.73

48

Table 10: Media Effects Model: al qaeda—Model Specification 10
Dependent Variable: Likelihood of belief there is a link between Iraq and al qaeda Independent Variables Coefficient SE Z Statistic Party ID Age Gender Race Education Level Income Level Region Support for Pres. Bush Belief War in Iraq Was Right Choice USA Today NBC ABC FOX CNN CBS Tone Total Psuedo R-Squared *p 0.05, one-tailed test 0.102 0.002 -0.318 -0.037 -0.202 -0.023 0.155 0.710 0.082 0.004 0.122 0.153 0.068 0.017 0.128 0.151 1.24 0.48 -2.61* -0.24 -2.99* -1.36 1.22 4.71*

-0.610 -0.780 -0.517 -0.549 0.039 -0.061 0.491 -0.347 0.160 0.2057

0.056 0.254 0.270 0.284 0.230 0.312 0.279 0.218 0.138

-10.98* -3.07* -1.91 -1.94 0.14 -0.20 1.76 -1.59 1.16

49


				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:22
posted:5/30/2009
language:English
pages:49