This study had three purposes. The first purpose of this study was to examine broadcast news visuals of war by specifically employing a content analysis to determine the overall tone of war coverage, and to describe what is covered in television news about the war in Iraq. In addition, an experiment was conducted to examine the impact of broadcast news footage of combat. Finally, to see if inoculation is an effective antidote to the impact of news footage of combat.
Content Analysis Methods
A content analysis was conducted of network television newscasts from March 26, 2004 – March 12, 2006. The analysis focused on the national evening newscasts relating to military operations broadcasted on ABC, CBS, and NBC. The news reports were provided by Vanderbilt University Television Archive. The entire 30-minute news broadcasts were analyzed in the 5:30 to 6:00 PM (CST) time slot. The 2004 date was chosen because it was roughly one year from April 9, 2003, when reporters started using the phrase, “The Fall of Baghdad.” It symbolized the fall of the statue of Saddam Hussein and was considered to be the end of the invasion phase of the war. The dates continued until March 12, 2006, almost three years after the start of the war. Only one broadcast was analyzed for each day. A purposive sample was determined by counting three and five days back from when the public opinion poll was taken. This represented a total of 92 days of television newscast.
To help justify the dates for the study, public opinion data was gathered from a number of polling institutes which included: Gallup Organization, Zogby America Poll, NBC/Wall Street Journal, CBS News/New York Times Poll, ABC News/Washington Post Poll, and National Annenberg Election Survey. Each of these polls asked a question very similar to the following question: “All in all, do you think it was worth going to war in Iraq, or not?” This question was chosen because it provided some insight into the U.S. publics’ perception of the war and, to some extent, the progression of the war. Publicly available polls conducted during the time frame were utilized. A total of 47 public opinion polls from March 2004 until March 2006 were used. The polls were used to stratify the dates used in the content analysis.
The broadcast unit of analysis was each single report by a broadcast journalist about a person, unit, or event with a clear beginning, middle, and end. A unit of analysis (N = 146) commenced with the anchor introducing a story and then either turning to a reporter or showing video of Operation Iraqi Freedom while the anchor reported in support of the video. The ending of each unit of analysis was defined as when a journalist “signed off” (“John Smith, CNN, Fallujah”), or the anchor clearly displayed that the story was over.
Six coders were trained to perform the content analysis. Coding norms were established during a supervised training session. Coders viewed 10% of newscasts. During training, coders established a high degree of standardization resulting in effective inter-coder reliabilities of R = .98 (Rosenthal 1984, 1987). Holsti’s (1969) method was used to examine inter-coder reliabilities on the nominal level data. Coders had a reliability of .95.
Overall tone of coverage was assessed with a global attitude measure adapted from Burgoon, Cohen, Miller and Montgomery (1978). It consisted of six 7-interval scales, including: good/bad, positive/negative, wise/foolish, valuable/worthless, favorable/unfavorable, and acceptable/unacceptable (a = 1.0, N = 146). The extent to which a broadcast embodied opinion was also assessed using a single-item indicator. The 7-interval scale ranges from opinion/interpretation (a fair and balanced news story).
The extent to which each unit employed framing was measured with a single 7-interval scale, episodic/thematic. The scale was used previously by Pfau (2004).
The news stories were coded for content focus according to the percent of the story devoted to the topic (0 –100%). Topics coded included: U.S. money spent in Iraq, megawatt hours, oil production, effectiveness of Iraqi forces, effectiveness of U.S. forces, Iraqi forces combat casualties, U.S. forces combat casualties, political instability, the prospect of democracy, public opinion ratings about the war in Iraq, Fallen Hero stories, or U.S. troops wounded in combat.
To examine the coverage of casualties, some questions were asked using nominal level scales. One question coded asked what specific casualties were discussed in each story (e.g., U.S. military casualties, Iraqi civilian casualties, insurgent casualties, Iraqi forces casualties, or other). The cause of casualties was also coded (e.g., road side bomb, gun fire, accident, or other). The broadcast news stories were also coded for discussion of wounded (e. g., U.S. military casualties, Iraqi civilian casualties, insurgent casualties, Iraqi forces casualties, or other). The possibility of the news story discussing the total number of deaths or wounded was also coded.
To examine the impact of broadcast news footage of combat and to assess the efficacy of inoculation in deflecting this impact, an experiment was conducted.
The investigation employed broadcast news stories discussing casualties of war. Specifically, the news stories featured reports of casualties without footage and reports of combat with footage involving U.S. forces, Iraqi forces, and Iraqi civilians. Broadcast news stories were selected from CBS evening news.
Participants were recruited from introductory communication classes at a midwestern university. A total of (N = 146) research participants completed both phases of the study (a retention rate from Phase 1 of 97%). Participant involvement with the issue (e.g. low, medium, and high) was used to randomly place participants in one of three casualty categories and into either an inoculation treatment or control group.
Design and Independent Variables
One part of the study featured a 2 x 3 Multi Variable Analysis of Covariance (MANCOVA) to examine the hypotheses and research questions. Independent variables were news condition, which was operationalized as a broadcast news story of combat operations in Iraq featuring a news report with footage of combat operations and without accompanying footage. Topic was operationalized as the focus of the story involving U.S. operations, Iraqi Operations, or Iraqi civilians. The second part was conducted on participants exposed to one of the news footage conditions. It featured a one-way MANCOVA to examine the efficacy of inoculation, operationalized as print inoculation or print inoculation with photograph. The effectiveness of inoculation was assessed by comparing attitudes, elicited involvement, and elicited affect of inoculated participants. Reliability of all scales was gauged using Cronbach’s coefficient alpha.
Receiver prior attitude, initial issue involvement, and gender were employed as covariants. Attitude toward the continued U.S. military presence in Iraq was assessed using six bipolar adjective pairs employed in recent inoculation research (Burgoon, Cohen, Miller & Montgomery, 1978). Adjective pairs included negative/positive, bad/good, unacceptable/acceptable, foolish/wise, wrong/right, and unfavorable/favorable. The reliability coefficient for prior attitude was a = .97, N = 143. Issue involvement was operationalized as the importance or salience of continued U.S. military presence in Iraq and was assessed using a version of the Personal Involvement Inventory (PII) (Zaichkowski, 1985). Six items of the PII were employed in this study to include: unimportant/important, of no concern/of much concern; means nothing/means a lot; doesn’t matter/matters to me; insignificant/significant; and irrelevant/relevant. Reliability for the issue involvement scale was a = .95, N = 145. Participant gender was operationalized as female and male.
The first part of the experiment attempted to assess the impact of broadcast news footage of combat. The broadcast news segments were 17 minutes in length, including commercials. Each segment contained one story specific to the casualty condition (U.S. military, Iraqi forces, Iraqi civilians). For example, the Iraqi civilian condition included a CBS evening newscast. The newscast was edited down to 17 minutes from the original 30 minutes. The newscast included a story (or package) about Iraqi civilian casualties. The package was 2:00 minutes in length. These same procedures were employed for each of the other two combat conditions. The packages discussing casualties with photos all featured the same reporter. The control package did not depict casualties. It only employed audio from the anchor discussing casualties, but no video showing specifics. There was a total of four newscasts. One newscast was employed for each of the three casualty conditions, and then the control video.
The second part of the experiment featured inoculation. The inoculation message was a generic preemption against the influence of visual images in forming personal opinion. The inoculation messages had a word count of 360 words. One inoculation message asked participants to examine a color photo of civilian casualties. The text of the inoculation messages was the same. The only difference between the inoculation messages was one referred participants to examine an attached still, color photo; the other did not contain a photo. Because inoculation theory posits that threat is a motivating catalyst in resistance, the first paragraph of the inoculation message was designed to elicit threat. Threat was operationalized as a warning of an impending news story featuring potentially influential pictures. The remainder of the inoculation message raised arguments that warned of the impact of visual imagery on their position.
The study was conducted in two phases. Phase 1 and 2 experimental booklets were prepared for participants. During Phase 1, demographic information was collected on research participants to include, name, gender, age, and year in school. Additionally, an exposure and attention measure of TV news use was collected.
Phase 1 was conducted over a period of 5 days. Participants were randomly assigned to one of the cells in the design: news condition (footage or no footage) and topic (U.S. forces, Iraqi forces, and Iraqi civilians). Those assigned to the news condition footage cell were assigned to one of three inoculation conditions (print, print plus photograph, and control). Subjects were assigned randomly with the exception that care was taken to ensure conditions were relatively balanced in regards to initial involvement. Additionally, inoculation cells were assigned more subjects than control cells (2:1 ratio). Also, care was taken to balance across the two types of inoculation treatments (one with message, the other with a message and a still photo). Phase 1 booklets contained an inoculation message warning against the impact of visual imagery on their opinions or a control message, a questionnaire that assessed the number of days participants spent watching TV news, and measure of threat, counter-arguing, involvement in Iraq, and attitude about U.S. military operations in Iraq.
One week later, over a period of five days, research participants were given the Phase 2 questionnaire. First participants went to the video lab to watch a video. The video was a 17- minute newscast. The newscasts were similar; however each newscast contained a 2:00 package about casualties. Participants either saw the news story without images of combat or the story with images of combat. After viewing the broadcast newscast, research participants completed a questionnaire. The questionnaire assessed attitude toward the continued U.S. military presence in Iraq, emotional response to the story, experiential and rational processing of the story, involvement, and counterarguing.
Research participant attitude about combined U.S. military presence in Iraq war was assessed using six bipolar adjective pairs developed for use in resistance research by Burgoon and colleagues (1978). Adjective opposite pairs were: unacceptable/acceptable, foolish/wise, unfavorable/favorable, negative/positive, bad/good, and wrong/right. Alpha reliability of the
attitude scale was a = 1.0.
Threat elicited by the inoculation treatment was measured using six bipolar adjective pairs employed in all recent inoculation studies. It was assessed in Phase 1, following administration of the inoculation treatment. A six-point scale consisting of bipolar adjective pairs was used to evaluate perceived threat against the person’s thoughts regarding the possibility of persuasive counterarguments influencing their position on the continued presence of the U.S. military in Iraq (perceived threat). Adjective pairings consisted of; not dangerous/dangerous, non-threatening/threatening, calm/anxious, not scare/scary, not harmful/harmful, and not risky/risky.
A thought-listing technique (Brock, 1967; Greenwald, 1968) was used to establish potential arguments against their position regarding the U.S. military presence in Iraq and subsequent responses to these potential arguments. After completing their list, subjects were asked to rate their arguments on a 1 (weak) to 7 (strong)-point scale, and then rate their thoughts and feelings on the responses to these arguments from 1 (weak) to 7 (strong). Lastly, the importance of continued U.S. military presence and involvement in Iraq was studied again using a six-point bipolar adjective scale (Zaichkowski, 1985). The scale included; unimportant/important, no concern/of much concern, means nothing/means a lot, doesn’t matter/matters to me, insignificant/significant, and irrelevant/relevant.
During Phase 2, participants were asked to complete an open-ended measure on which they identified possible arguments contrary to their own position and then listed potential responses to those arguments in the spaces provided. The procedure is based on the thought-listing technique pioneered by Brock (1967) and Greenwald (1968). However, past use of this technique alone has proven to be inadequate in inoculation research (Pfau et al., 1997). Eagly and Chaiken (1993) have argued that thought-listing does not reflect the amount of cognitive effort expended. In addition, thought listing, by itself, fails to acknowledge the prospect that respondents may view their own thoughts as varying in power and intensity, both in cognitive and affective terms. Therefore, after generating their list of arguments contrary to their position and responses to those arguments, respondents rated perceived strength of arguments contrary to their position and strength of responses using a 1 to 7-point scale.
Multiple item indicators were used to evaluate emotion. The emotion scale was based on the previous work of Dillard and colleagues (Dillard, Plotnick, Godbold, Freimuth & Edgar, 1996; Smith & Dillard, 1997). Featured emotions included anger (angry, irritated, and annoyed) a = .79, surprise (surprise, astonished, and amazed) a = .81, puzzled (puzzled, bewildered and confused) a = .85, sad (sad, dreary, and dismal) a = .79, fear (fearful, afraid, and scared) a = .90, and pride (dignity, honor, and gratification) a = .80. The category of pride is added to this scale for the purpose of this particular study.
Task specific experiential and rational cognition were also measured.
There were 13 items that asked participants how the news story made them feel, the scale had a = .91. Those items were: the story made them sad, they felt comfortable with the story, they felt fearful, they felt good, they were disgusted, they angry, they felt joyful, unpleasant, happiness, bad, uncomfortable, and/or felt agitated.
Items on the next scale were included on two different dimensions, experiential and rational. Experiential items reached an a = .81 and included: use of instincts, using the heart as a guide for reactions, feelings, intuition, use of hunches to make decisions, gut feelings, use of free association, flashes of insight, first impressions, and ideas “popped” on the subjects head. The rational processing items reached an a = .77 and include: reasoning, analytical assessment, systematical judgments, focus on steps to process the story, use of rules, awareness of mental process, focus on task before arriving to judgment, and careful use of information in order to arrive to conclusions.
The task specific scale items are as follows. There were 18 items used to make the student reflect on the process of watching the news cast. “I reasoned things out carefully.” “I used my instincts.” “I approached and assessed the news story analytically.” “I used my heart as a guide for my reactions.” “I assessed and judged the news story systematically.” “I went by what felt good to me.” “I was very focused on the steps involved in judging the video.” “I relied on my sense of intuition.” “I trusted my hunches.” “I used clear rules.” “I was very aware of my thinking process.” “I used my gut feelings.” “I was very focused on what I was doing to arrive at my judgment.” “I used free-association, where one idea leads to the next.” “I had flashes of insight.” “I relied on my first impressions.” “Ideas just popped into my head.” “I arrived at my assessments by carefully assessing the information in front of me.” (Novak & Hoffman, 2005).
Research participant attitude about combined U.S. military presence in Iraq was assessed using six bipolar adjective pairs developed for use in resistance research by Burgoon and colleagues (1978). Adjective opposite pairs were: unacceptable/acceptable, foolish/wise, unfavorable/favorable, negative/positive, bad/good, and wrong and right. Alpha reliability of the attitude scale was a= .97.
Lastly, the importance of continued U.S. military presence and involvement in Iraq was studied again using a six-point bipolar adjective scale (Zaichkowski, 1985). The final issue involvement scale’s a = .94. The scale included the following items: unimportant/important, of no concern/of much concern, means nothing/means a lot, does not matter, matters to me, insignificant/significant, and irrelevant/relevant.