communications 2

Tortor11111

ParkHolodyZhang2012.pdf

Home >English homework help >communications 2

Race and Gender

Race in Media Coverage of School Shootings: A Parallel Application of Framing Theory and Attribute Agenda Setting

Journalism & Mass Communication Quarterly 89(3) 475-494 © 2 0 I 2 A E J M C

Reprints and permission: sagepub.com/journalsPermissions.nav

D O l : 10.1177/1077699012448873 http://jmcq.sagepub.com

i)SAGE

Sung-Yeon Park'•^ Kyle J. Holody^ and Xiaoqun Zhang'

Abstract

This study investigated news media coverage of the race of the perpetrator in the 2007 Virginia Tech shootings, using agenda-setting and framing perspectives. More than one-third of newspaper articles contained racial information. The agenda-setting analysis enabled comparison with coverage of the Columbine shootings, in which race was virtually absent; framing analysis revealed that the media framed the VT incident around the perpetrator's ethnicity and generalized criminal culpability to his ethnic group. Racial and ethnic references were also sometimes displayed in prominent positions.

Keywords

agenda setting, Asian, framing, race, school shooting

On April 16, 2007, a Virginia Tech (VT) student named Seung-Hui Cho shot and killed thirty-two faculty members and students on the campus and injured twenty-five more before taking his own life. In the midst of a subsequent flood of media coverage, two articles were published—one on MSNBC.com and Newsweek.com and the other on Salon.com—that specifically discussed Cho's race as it related to the shootings.

'Bowling Green State University, Bowiing Green, O H , USA 'Korea University, Seoul, Republic of Korea 'Coastal Carolina University, Myrtle Beach, SC, USA

Corresponding Author: Sung-Yeon Park, Department of Telecommunications, School of Media & Communication, Bowling Green State University, 302 West Hall, Bowling Green, O H , 43403, USA Email: [email protected]

476 journalism & Mass Communication Quarterly 89(3)

The first article was titled "Eerie Similarities" and featured an extensive interview with Wayne Lo, an Asian perpetrator of a previous school shooting.' Comparisons between the shootings that Lo had committed at Bard's College at Simon's Rock (SR) in 1992 and the 2007 VT shootings were made throughout the article, and he was asked multiple times to provide insight into the mind of the VT shooter, whom he had never met. On the other hand, the second article critiqued undue attention paid to the race of perpetrators in previous media coverage of mass shootings committed by Asians and wondered whether that trend would be repeated in media coverage of the VT incident.^

The current study systematically investigates whether and how major U.S. national newspapers racialized the VT shootings. This incident was chosen because the race of the perpetrator was neither proclaimed by the perpetrator himself nor determined later by investigators to be the primary trigger of the crimes, unlike other shootings perpe- trated by Asians, such as the 2004 Wisconsin deer hunter shootings and the 2009 Binghamton immigrant center shootings. Rather, the VT shootings can be considered to share more characteristics with other school shootings, and race, albeit possibly a factor, did not "need" to be mentioned prominently in news coverage.

As evident from a list of nine recent high-fatality school shootings,^ it is difficult to locate such crimes in which the perpetrator is of Asian descent. Asians compose the least represented racial group in FBI crime statistics,'' and a journalistic recount identi- fied only two other incidents that made national headlines:'̂ the 1991 University of Iowa shootings and the 1992 SR shootings. Thus, one could argue that the very scar- city of such crimes committed by Asians might have made the VT perpetrator's race stand out more to joumalists.

The necessity (or lack thereof) of mentioning the race of perpetrators within news reports may become clearer if one compares media coverage of shootings, committed by whites and Asians, respectively, that are similar in nature. The 1999 Columbine shootings can be considered most comparable to the VT shootings in terms of size and impact. Replication of Chyi and McCombs's measurement scheme developed to examine news coverage of the Columbine incident presents a convenient yet pertinent means of comparison;^ the purpose of the original scheme was to solve the cross- object generalizability problem commonly cited as a main weakness of framing the- ory. Thus, in addition to examining media coverage of the VT shootings, this study also compares media coverage of the VT and Columbine shootings using both the agenda-setting-based analytic scheme developed by Chyi and McCombs and a more traditional race-framing scheme developed for the present analysis.

Media Coverage of School Shootings: The Framing Perspective Because news coverage of school shootings can offer multiple points of understand- ing, it is important to identify which particular aspects of entire incidents were high- lighted at the expense of others. In the case of the Columbine shootings, a systematic

Park et al. All

analysis of news coverage identified eleven different frames, of which gun control and violent popular culture were predominant.^

Of these eleven frames, a few received particular attention from researchers. For example, Scharrer, Weidman, and Bissell pointed out prominent use of the popular cultural product culpability frame.* The notions oï alienated youth and Juvenile super predator were also identified as highly salient frames, though they had rarely been applied to describe previous school shootings.' After the September 11 attacks, an increasing number of stories referencing the Columbine shootings retroactively employed the term terrorism.^'^ In sum, these framing analyses of the Columbine inci- dent collectively illustrate that the crime contained base materials for many different frames and that multiple explanations offered in news coverage of complex crimes like it are, while at least partially inherent, constmcted to a great extent by social, cul- tural, and historical forces influencing the institution and practice of joumalism.

Whereas deductive approaches taken in the research reviewed above may be effective in uncovering framing by inclusion, a comparative approach may reveal framing by omission. An example is the comparative analysis of media coverage of Columbine and of another shooting at Red Lake Senior High School in Minnesota. Although it was the largest school shooting since the Columbine incident, it was largely ignored by national media." According to the authors, the markedly differ- ent media responses could be largely attributed to the race and class of perpetrators and victims: Because the shootings took place on a low-income American Indian reservation, the news media determined the story devoid of the mass appeal of the Columbine incident involving mostly white suburban students. By neutralizing its racial elements, the media constructed the Columbine shootings as a crime that can happen to anyone.

The absence of a "race" angle in the Columbine coverage has been noted by other scholars. Weissman questioned why the media hardly made a connection between the perpetrators' obsession with Hitler and the targeted killing of a black victim.'^ Similarly, Frymer noted the reluctance of the media to mention the racist beliefs of the Columbine perpetrators.'^ More directly, Jaekson stated "there was never any mention of the race of the killers . . . none of them [news media] even made mention of either the fact that there were 12 other school shootings initiated by young White males or that the self-labeled 'trenchcoat mafia' was an organized gang."'''

In the case of the VT incident, the matter of race and ethnicity becomes even more complicated due to the immigrant status of the perpetrator. To protect collective self- image, Greer and Jewkes noted, the mainstream media tend to construct an "outsider" status for perpetrators of exceptional offenses. The othemess of perpetrators of egre- gious crimes can be even more accentuated if they happen to belong to a group that is already marginalized or viewed as inherently criminal, such as immigrants.'^ Despite strong empirical evidence demonstrating the opposite, the American public often misperceives immigrants as more prone to criminal behavior.'* Emphasis in media coverage on the race and immigrant status of the VT perpetrator runs the risk of rein- forcing this misperception that, in tum, becomes a cause of racial injustice.

478 Journalism & Mass Communication Quarterly 89(3)

Media Coverage of School Shootings: The Agenda-Setting Perspective

Because agenda-setting theory is applicable at both the issue and attribute levels,'^ media coverage of school shootings can be analyzed at these two levels as well. Although no particular theory was mentioned in it, a study examining the salience of fourteen school shooting incidents in network television news can be considered an example of issue agenda-setting research.'* By measuring the amount of news cover- age of the incidents, the researchers showed that shooting incidents that resulted in more deaths or greater injuries were covered more. They also found that the amount of coverage was highly correlated across major television networks, demonstrating their "herd mentality."

Most existing agenda-setting research set in the context of school shootings, how- ever, has adopted attribute agenda setting as its theoretical framework. The analytic scheme developed for the first study in this line of research comprised two axes: time and space. Chyi and McCombs tested this measurement scheme with New York Times (NYT) coverage of the Columbine incident.''

Two years later, the two-dimensional scheme was reapplied to media coverage of nine school shootings, including Columbine.^" Muschert and Carr traced shifting salience of specific time and space dimensions for the shootings and found the mea- surement scheme useful in tracking emergence, peak, and decline of school shootings as a media agenda item. Contrary to the original intent of Chyi and McCombs, how- ever, the authors found it difficult to use the measurement scheme for cross-case com- parisons because the scheme requires that objects have a similar level of basic salience to even allow comparisons. Although the usage patterns for the two most dominant levels did not, the patterns for two secondary levels corresponded closely with the total number of articles, illustrating high sensitivity of the analytic scheme to total amount of coverage and thus limited applicability of the scheme for cross-case comparisons.

Two other studies have also investigated media coverage of school shootings in reference to this analytic scheme. To address limitations faced in the earlier analysis of nine school shootings, Muschert identified three major themes that were evident in news coverage of the Columbine incident: the shooting event itself, reactions to it, and commentaries about it.^' Subsequent application of these themes revealed that the media were initially focused on the incident itself and then moved on to reactions outside of the Littleton community.

The second study examined media coverage of the VT shootings.^^ The authors expanded their investigation by incorporating collectivism into the time and space measurement scheme and by comparing U.S. and Korean newspapers. Overall, news- papers from the two countries were similar in their deployment of the time and space dimensions. At the same time, there was a much heavier usage of the collectivistic frame in Korean newspapers.

In sum, these studies reveal that high-profile shootings received more media cover- age than did lower-profile shootings and thus presented a reliable account of how an

Park et al. 479

issue stayed on the media agenda. When they were on the media agenda, the incidents went through phases in which actions of individuals became salient first, followed by discussion of broader social perspectives and long-term ramifications.

Beyond these very broad observations, however, the time- and space-based analytic scheme did not quite live up to its promises. Chyi and McCombs had proclaimed, "this measurement scheme solves the cross-object generalizability problem inherent in object-specific approaches while increasing sensitivity to a variety of news perspec- tives."^^ As demonstrated, the measurement scheme did solve the cross-object gener- alizability problem, at least for objects that had similar levels of salience. However, the scheme's ability to reveal a variety of news perspectives is less obvious. Although Chyi and McCombs stated time "corresponds to the 'when' in the five W's of joumal- ism" and space "refers to the 'where' and may also include the 'who,' the 'what,' and even the 'why,"^" the analytic scheme reduced rich details of the shooting incidents to the point that it was difficult to determine what the findings meant in terms of audience understanding of an incident.

Even the attempt by Muschert to overcome this limitation appears to be afflicted by a similar problem.^' "Unconcerned with the thematic elements of the Columbine cov- erage," he argued, Chyi and McCombs "limited their analysis to demonstrate the frame-changing dynamic of salience maintenance."^* However, the analytic scheme Muschert devised to "provide deeper insight into the thematic discourses about Columbine" was likewise devoid of content.^^

The limitations of the Chyi and McCombs measurement scheme and its alternatives highlight the debate among researchers concerning the relationship between framing and agenda setting.^* At the risk of oversimplification, the landscape can be summa- rized into three camps. One camp, led by McCombs and colleagues, maintains that framing is the same as increasing the salience of a news object's attributes, and thus represents a "second level" of agenda setting beyond issue salience. The second camp, represented by Entman, defines the four fimctions of frames as problem definition, causal analysis, moral judgment, and remedy promotion and argues that issue agenda setting and attribute agenda setting correspond respectively to the first and the latter three functions of framing.^' The third and rather loose constellation of scholars consid- ers a convergence of the two theories unproductive and even unobtainable, at least as their definitions now exist. Whether focusing on framing research outside of the social psychological tradition,'" on the functions of framing beyond initial problem defini- tion,^' or on more advanced stages of framing research in news production and differ- ences in the cognitive processing involved in the effect processes,''^ scholars in this category regard convergence of agenda-setting and framing theories as neither feasible nor productive, although they do encourage continued exploration of the relationship.

To date, this discussion has been mostly conceptual and not based on tangible data. By applying analytic schemes, developed from agenda setting and framing, to media coverage of one incident—the VT shootings—and presenting the findings side by side, we intend to provide empirical grounds for researchers to frirther develop their arguments.

480 journalism & Mass Communication Quarterly 89(3)

Research Questions

The first set of research questions is bom of agenda setting. First, the overall salience of the "race" attribute is gauged.

RQla: How salient was the race attribute of the perpetrator in media coverage oftheVT incident?

Next, media coverage of the VT and Columbine incidents is compared to test cross-object generalizability of Chyi and McCombs's scheme, using data about the VT coverage collected here and data about the Columbine coverage provided by Chyi and McCombs." Whether the salience of the race attribute was related to specific levels of time and space is also examined.

RQlb: Were there differences in the usage of the time and space dimensions between media coverage of the VT and Columbine incidents?

RQlc: Was the salience of the race attribute in media coverage of the VT inci- dent related to specific levels of the time and space dimensions?

The following research questions can be considered as more typical of framing research. Under the metaframe of "racialization of the VT shootings," media coverage is analyzed for racialization of the perpetrator and the crimes, racial generalization of criminal culpability, prominence of racialization, and racialization via implicit stereo- typing.

RQ2a: How much did the media racialize the VT perpetrator by revealing his racial/ethnic minority status in detail?

RQ2b: How much did the media racialize the VT shootings by cormecting them to the race/ethnicity of the perpetrator?

RQ2c: How much did the media generalize culpability of the VT shootings to a wider group of ethnic/racial minorities?

RQ2d: How prominent—in headline, in first paragraph, or in body text—were racial/ethnic references in media coverage of the VT incident?

RQ2e: How much did the media implicitly racialize the VT perpetrator by using common Asian stereotypes to describe him?

Method

To compare media coverage of the Columbine and VT incidents, the unit of analysis for examining VT coverage was kept constant with that of the Chyi and McCombs study: a newspaper article. The time frame and method of article collection were also consistent: newspaper stories within a one-month period following the date of the VT incident were obtained from the LexisNexis database. In addition to NYT, USA Today

Park et al. 481

(UT) articles were included to increase generalizability. From the initial pool, news- paper articles were screened to determine whether they actually centered on or were written as a consequence of the shootings. The decisions were made as part of the official coding process and were thus included in the intercoder reliability test. This process yielded 136 articles to be analyzed for their actual content {N = 70; TV = 66). Three coders went through multiple intercoder reliability training sessions and tests; after analyzing a random sample of 10% of all the articles, they achieved Krippendorff s a reliability estimates of .81 or higher.

Coding of Attribute Agenda-Setting Variables

Space and time. The existing coding scheme for time and space attributes was used to make this study comparable to the two previous analyses of school shootings.^* Those previous studies diverged on only one point in the coding scheme: a story focus- ing on any single community other than Columbine High School or Littleton was coded at the community level in the earlier study, but at the societal level in the later study.̂ * The current study adopted the later study's position. Other than this, the space scheme, comprising individual, community, regional, societal, and international, and the time scheme, encompassing past, present, and ftjture, were replicated here (see the appendix for descriptions and examples of all coding categories and levels).

Race. Newspaper articles were coded as either having or not having a race attribute.

Coding of Framing Variables

Radalization of the perpetrator. This was measured using three categories. The first was whether the perpetrator's race was identified. The second was whether the perpe- trator's ethnicity was identified. The third, most detailed category was whether the perpetrator's immigration status or history was mentioned. Because a more detailed category (e.g., Korean) "assumed" a less detailed category (e.g., Asian) either explic- itly or implicitly, an article registering ̂ ê* on a detailed category of racial information was checked yes on all less detailed categories of this subframe.

Radalization of the crimes. The extent to which news articles linked the crime to the perpetrator's race, rather than specific qualities, was measured using three categories. The first category was marked yes when a news article described the perpetrator using Asian stereotypes with a clear acknowledgment of their cultural currency because such an article ascribed the crime at least partially to stereotypical characteristic(s) of the Asian race. The second category was marked yes if an article mentioned similar crimes committed by an Asian(s) because such an article implicitly generalized such violence to the race. The third category was coded >'ei if an article explicitly attributed, at least partially, the crime to the race of the perpefrator. Although the third category can be considered the most direct race-based attribution of the crimes, the presence here of one category did not logically necessitate the presence of the other categories. Thus, each newspaper article was examined for these three categories independently.

482 Journalism & Mass Communication Quarterly 89(3)

Racial generalization of criminal culpability. Five increasingly broad categories cap- tured the degree of racial generalization of an article. If a story referred to the race or ethnicity of the individual perpetrator, it was coded at the individual level. If it reported on shame and distress experienced by the family of the perpetrator, it was coded at the family level. If it highlighted guilty feelings expressed by Korean Americans or South Koreans, it was coded at the ethnicity level. If it discussed reactions from various Asian ethnic communities—not only the ethnicity of the perpetrator—or backlash against Asian Americans throughout the country, it was coded at the race level. Finally, if it employed a perspective that encompasses other, non-Asian minorities as well as Asians, it was coded at the all minorities level. This variable measures extended burden of the crimes to broader minority groups, whereas racialization of the perpe- trator focused attention on the perpetrator's minority status by revealing increasingly more details about it.

Prominence of racialization. This was determined by the location of reference(s) to the perpetrator's race/ethnicity: body text other than the first paragraph, first para- graph, headline. Although headline was considered the most prominent position, fol- lowed by first paragraph and body text, each article was examined separately for the three categories because headline presence of racial information did not automatically warrant its presence in the first paragraph or body text. Racial references in quotes were treated the same as other references because joumalists make decisions about use of quotes just as they would other information in news articles.

Racialization via implicit stereotyping. According to the aversive racism paradigm in social psychology, people are affected by egalitarian norms as well as deeply and often subconsciously ingrained biases against minorities.'* In the current study, elite jour- nalists operating under close public scrutiny may not commonly make explicit race- based attributions. However, jotimalists may subconsciously engage in racialization by using common Asian stereotypes to describe the perpetrator without explicitly link- ing the traits to his race. Therefore, each article was scmtinized for usage of these common Asian stereotypes: asocial, good student, hard working, asexual, untrust- worthy, family honor, small physique.

Results

Preliminary Analysis: Volume and Pattern

The seventy NYT and sixty-six UT articles illustrated a similar pattem of distribution during the one-month period examined here. More than half of articles by both NYT (n = 37, 53%) and UT (n = 41, 62%) were published within the first four days. In addition, both newspapers virtually stopped covering the incident after the sixteenth day. This condensed cycle was a departure from the Columbine coverage, in which ten days passed before approximately half of all articles were published and in which more than 25% of articles were published after the sixteenth day (see Figure 1).

Park et al. 483

1 2 3 4 5 6 7 8 9 10111213141516171819 20 2122 23 24 25 26 27 28 29 30 Day

Figure I. Thirty-day Virginia Tech (VT) shootings coverage in the New York Times and USA Today: Comparison with Columbine (CO) shootings coverage in the New York Times

Agenda-Setting Results

RQI a: Salience of the race attribute. Of all newspaper articles about the VT shoot- ings, 38% contained some information indicating the perpetrator's race or ethnicity, though the salience of the race attribute varied over time. On the first day, 20% of articles contained a race attribute, compared to 45% the following day. It peaked between the third (68%) and fourth days (69%) and then stayed visible until the tenth day (see Figure 2).

R Q I b : Time and space comparison: VT versus Columbine. Shifts along the three lev- els of time during the first sixteen days of VT coverage revealed a pattem similar to that in Columbine coverage for thirty days found by Chyi and McCombs.'* "Past" peaked on the third day, occupying 27% of total VT coverage, only to wane gradually and completely disappear after the eighth day. "Present" was most dominant near the beginning of issue cycle. One major departure from the Columbine coverage was a relatively strong presence of the "future" category throughout the issue's life span. Whereas the "future" category maintained a 10% or lower presence until the fifteenth day of the Columbine coverage and never surpassed the "present" category, "ftiture" hit a 30% mark on the second day of the VT coverage and tied or surpassed the "pres- ent" category for four days toward the end of the sixteen-day news cycle.

Overall, Columbine and VT coverage were similar on the space dimension. The "societal" category was dominant throughout both news cycles, while the "commu- nity" and "individual" categories exhibited continual decline. One noticeable differ- ence was the relatively high percentage of the "individual" category and its intermittent peaks in use after the midpoint of the VT issue cycle. Furthermore, 10% of VT articles

484 journalism & Mass Communication Quarterly 89(3)

1 2 3 4 5 6 7 9 10 11 12 13 14 15 16

Figure 2. Salience and distribution of race attribute in Virginia Tech sbootings coverage: First sixteen days

registered at either the "regional" or "international" space level, whereas the two lev- els were virtually absent in Columbine coverage.

When the time and space dimensions were combined, the list of the most com- mon categories was similar between the two incidents: societal-present was most common, followed by societal-future, community-present, individual-present, and individual-past. At the same time, heavy concentration in the societal- present category found in Columbine coverage was dispersed more evenly in VT coverage, resulting in a noticeable increase in the societal-future category and some coverage at the international and regional levels of the space dimension (see Table 1).

RQI c: Relationship among time, space, and race. To determine whether the salience of the race attribute in media coverage of the VT incident was related to specific levels of time and space dimensions, two chi-square tests were conducted relating presence of a race attribute (the dependent variable) and the time and space attri- butes (as respective independent variables). The time attribute was related to race salience. Whereas articles with past and present time levels were equivalent in their inclusion of a race attribute, in 43% and 44% of articles each, the rate was much lower at the future level (20%), x\2, N= 136) = 6.16, p = .046. The salience of the race attribute varied across space levels as well. Whereas the individual (38%), com- munity (36%), and societal (34%) levels were similar, their race salience was lower than for the regional (67%) and international (57%) levels, but not significantly, x\4, N = 136) = 3.1S, p = .436.

In addition, distribution of the race attribute was examined. Although articles with "present" time (28%) and "societal" space (18%) levels exhibited higher percentages

Park et al. 485

Table I. Space by Time by Race: Comparing Coverage of the Columbine (CO) and Virginia

Tech (VT) Shootings

CO/VT

Space

Race

International

Societal

Regional

Community

Individual

Yes

Past(

— 2

• —

10 Í

0 2 0 1 8 4

Time

Present (%)

CO VT

— 4

39 28

— 4

24 19

7 10

? 28

Future

13 —

1 0

(%)

Race

Yes(%

) ) 7

The information about newspaper coverage of the Columbine incident was from Chyi and McCombs, "Media Salience and the Process of Framing: Coverage of the Columbine School Shootings," Journo/ism & Moss Communication Quarterly 8\ (spring 2004): 22-35. Because the study did not report figures for the regional and international levels broken down by the time levels, percentages of articles in the two space levels for Columbine coverage are marked with a dash. In addition, race attribute salience figures in Columbine coverage are noted with a question mark because of a lack of data.

of articles with a race attribute(s), the percentage in the "societal" level reflected the large number of articles within the category and not by disproportionately higher salience of race attribute at that particular space level. What may deserve more atten- tion in the distribution presented in Table 1 is what is missing: information on the race attribute in Columbine coverage.

Framing Results

RQ2a: Raciallzatlon of the perpetrator. Fifty-one articles (38%) identified the VT perpetrator's race, Asian, and all but one of these (37%) identified his ethnie- ity, Korean. Furthermore, 8% of articles revealed the perpetrator's immigration status.

RQ2b: Racialization of the crimes. Only two articles about the VT incident adopted this subframe. One article (1%) mentioned similar crime(s) committed by an Asian perpetrator, and another article (1%) attributed the crimes to the perpetrator's ethnicity by referring to a lack of mental health awareness in the Korean American community. This second article also explicitly discussed several stereotypes of Korean Americans in relation to the crimes.

R Q 2 c : Racial generalization of criminal culpability. Of the fifty-one newspaper articles that identified the VT perpetrator's minority status, a majority (69%) discussed racial/ ethnic identity at the individual level. A substantial percentage of articles (22%)

486 journalism & Aloss Communication Quarterly 89(3)

portrayed Koreans as the ethnic in-group that shares the shame and burden of the per- petrator's crimes. The family (4%) and race levels (6%) were rare, and the "all minori- ties" level was completely absent (0%).

RQ2d: Prominence of radalization. When the three categories were examined, 36% of articles contained racialization information in the body text, 13% in the first para- graph, and 4% in the headline.

When classified based on their highest level of racial prominence, the lowest prominence—racial reference(s) in body text only—accounted for 63%, the next higher prominent level accounted for 27%, and the highest prominence accounted for 10% of all articles with racial references.

R Q 2 e : Usage of stereotypes. Of the seven common Asian stereotypes examined, "asocial" was used most often, in 6% of newspaper articles, followed by family honor (2%), asexual (2%), and small physique (2%). The other descriptors were not observed.

Discussion

The core question, whether and how the media racialized the VT incident, was pur- sued via two separate routes, one based on attribute agenda setting and the other based on framing theory. Both investigations revealed that 38% of the newspaper articles contained racial or ethnic references to the perpetrator. Beyond this point, however, their foci diverged.

Agenda-Setting Analysis

The agenda-setting-based analysis clarified the primary results, allowing comparison with a study of the Columbine shootings.^' Although the two incidents were more similar than different along time and space dimensions, media coverage of the VT incident as a whole consistently exhibited one attribute—race—that Columbine cov- erage had been criticized for overlooking.

Although we cannot say that the race attribute was a driving force behind the media coverage, it was undoubtedly an important element of the crimes, as recon- structed in the media reports, in two ways. First, the race attribute was particularly salient during the first few days when public attention was highest. Second, the race attribute served a role(s) other than simply being the perpetrator's demo- graphic trait because it was also present in newspaper articles published much later in the issue cycle, when his race or ethnicity itself could no longer have been newsworthy.

In addition, the scarcity of a race attribute in future-oriented articles, as opposed to present- or past-oriented articles, may support the argument that more pmdence is needed before making racial references. "Future" articles discussed remedies and

Park et al. 487

corrective actions, so the dearth of racial references in those articles suggests that race/ ethnicity of the perpetrator was not really at the heart of the problem and that, possibly even worse, an early focus on race may have distracted readers from tmderstanding the real issue.

Last but not least, the contrast between Columbine and VT coverage in the distribu- tion of the race attribute illustrates that interobject comparison can be a powerful method to uncover frames that researchers cannot effectively demonsfrate by focusing on a single object or event.

Framing Analysis

The framing analysis focused on the content of the VT incident articles containing a race attribute. Although the perpetrator's ethnic Korean heritage—as opposed to other Asian ethnicities—was deemed not particularly relevant to either his mental illness or crimes, his ethnicity was still frequently identified. References to his immi- gration status, the most racializing information, were not as common, but occurred nonetheless.

In terms of generalization, the results indicated that the perpetrator's "other- ness" was limited primarily to the individual. At the same time, his Korean heri- tage was projected to a substantial degree to the larger group of Korean Americans or South Koreans. Asians and Cho's family, albeit to a far lesser extent, were also cast in several articles as "others" distinguished from "us," the general public.

Racializing information was found not only in body text but also in prominent locations, such as the first paragraph or headline. Furthermore, examining the loca- tion of racial references should not obsctire the fact that the articles frequently con- tained multiple references. In addition to capturing readers' attention, (racial) information in prominent locations—headlines in particular—in prestigious national newspapers may tend to be recycled by other media outlets, blogs, and individuals. On a positive note, there was little raeialization of the crimes. The MSNBC/Newsweek interview with Lo tumed out to be a rare exception, and direct racial atfribution was equally scarce.

Last, the few incidents where "common" Asian stereotypes were utilized to describe the perpetrator raise a complicated question: when does an adjective stop being simply descriptive and become a stereotype? Being "asocial" is a char- acteristic shared widely by criminals, certainly including the VT perpetrator, and thus its usage could be based more on the individual's personality than on his race. On the other hand, usage of the other words such as "family honor," "asex- ual," and "small physique" could be considered to have been prompted by the race of the perpetrator. In a larger context, presence or absence of these words may not be as meaningful when the perpetrator's racial/ethnic identity was clearly

488 Journalism &. Mass Communication Quarterly 89(3)

stated—and sometimes discussed in detail—in more than one-third of all news- paper articles. Still, it is reassuring to know that the stereotypes were used sparingly.

This investigation also speaks to the need to continue efforts to make the media inclusive in employment as well as representation. Asian American joumalists were among the first who alerted the public and the joumalism community of potentially excessive racialization of the VT incident. Also, with more common representations of Asian Americans, some positive and some negative, Cho's race might have escaped the attention of reporters, just as it did for the Columbine shooters Harris and Klebold.

Summary and Suggestions for Future Research

The current study lends support to the idea that agenda-setting and framing theo- ries are distinct. The two sets of findings concerning our one research object sug- gest attributes and frames are not interchangeable. Rather, they can make unique contributions, and their deployment should be determined by the purpose of inquiry.

Furthermore, the coding process of this study revealed a limitation of both the agenda-setting and framing approaches to understanding the news coverage examined here: a quantitative analysis can be only as good as its categories.''" A few articles examined here were written specifically to report on unwarranted attacks against Asian or Korean Americans after the VT incident. Such enhanced racial sensitivity by joumalists deserves to be recognized. However, using the current coding schemes, such articles were likely coded the same as ones that identified the perpetrator's race or ethnicity with no clear purpose. The only type of analysis that could have captured this unexpected, albeit scarce, phenomenon would have been a qualitative framing analysis guided by the critical paradigm.'"

Finally, there remains one critical question that demands more discussion: should the media discuss race at all when reporting crimes? Without preexisting sociopoliti- cal conditions and the undue infiuence of stereotyping, racial identification of a crim- inal could simply be part of comprehensive news reporting. Furthermore, for crimes as complicated as the VT shootings, understanding the perpetrator, including his race and ethnic background, may be crucial in understanding the incident and which aspects of his background may have contributed. Still, one could argue that not all of the mentions of race and ethnicity were necessary, and some of the connections reporters made between the crime and the perpetrator's race/ethnicity were rather speculative and published ahead of formal investigations. Hopefully, this discussion will lead to future research that identifies more considerate and appropriate ways of referring to race in news.

tlfl

I o

1 -

-ac 1)

O •{j

§0 01

OJ _ l

o c

M O

c n

c W) t:!

0) ç c e «

1.11 Ko re

co n

ta l

m e n

tí

S ti

Suo:

_5 tí = c tj ' ^ .2 Q Si !ïï JJ

•

M 3 E 2 E -̂

E : ^ i E c i i <u o E -Q

3 u <U ë - ? E

g !_ O a)

ffI 2

ta 3 E

o U

II .1

0) JJ

>5! o 2 (U «-J VI

TO ( - C

^ J -5

489

ni X

O D n.

§0

a (4

"O V 3

C O u

a. a.

;s a

n d /

c 1- <u <u '" s

ns C or

ea r

0 _

iC ti

on r

ca ns

L .

;u d e n t;

c ra

le ge

c o

is es

nS U

A ra

b s

u m

S E

¿- S-g o ^ 2 g b o .£ *-• "î" J .E ¿ E Í3 .E

b g.

0 . E « ^

c 0

pa s (/>

'o JZu

1 "5 a.

le s;

M—

oo c

lo si

oo c

'ñ -au

n j

(/> E

ic ti

C o (J UJ

. (U c

ai c 5 •S f . 8 o S c c r o>O b I Si i e

sn o i

tin gs

o n

0 C / l

13 U O

o o l

_c u

ic ti

> C!

im ai

o o l

uos,

i.i E c — S) —

1- a) c 3 -C .E 1/1 w <->

o o

490

1 bM i § 1

" Sí! r̂

111

U >̂ o -o 3 O ^

LE I

• D C (U

(U Q

o u

a a <

« O -Ö El

•5 £r S .. . 1

o a) c u rt ^ E J5 ^ L.

O ¡3 ~

= 11 u ^ O O (U w

s.s c « o E

l _

491

492 Journalism & Mass Communication Quarterly 89(3)

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interests with respect to the authorship and/or publication of this article.

Financial Disclosure/Funding The author(s) received no financial support for the research and/or authorship of this article.

Notes

1. Samantha Henig, "Eerie Similarities," MSNBC.com, May 2, 2007, http://www.newsweek. com/2007/05/Ol/eerie-similarities.html.

2. Jeff Yang, "Killer Reflection," Saion, April 19, 2007, http://www.salon.com/news/opin- ion/feature/2007/04/19/cho_shooting.

3. Glenn W. Muschert and Dawn Carr, "Media Salience and Frame Changing across Events: Coverage of Nine School Shootings, 1997-2001," Journalism & Mass Communication Quarterly 83 (winter 2006): 747-66.

4. FBI Uniform Crime Reports, "Crime in the United States 2010," http://www.fbi.gov/about- us/cjis/ucr/crime-in-the-u.s/2010/crime-in-the-u.s.-2010/tables/table-43/10tbt43a.xls.

5. Yang, "Killer Reflection." 6. Hsiang Iris Chyi and Maxwell E. McCombs, "Media Salience and the Process of Framing:

Coverage of the Columbine School Shootings," JoMrnafem & Mass Communication Quar- terly 81 (spring 2004): 22-35.

7. Regina G. Lawrence and Thomas A. Birkland, "Guns, Hollywood, and School Safety: Defining the School-Shooting Problem across Public Arenas," Social Science Quarterly 85 (5, 2004): 1193-1207. The frames school programs/security, teen life, parental/adult/ community oversight, and mental health were employed far less frequently than the flrst two frames, yet still more often than were the frames sociai breakdown, secularism, lax criminal justice system, individual character of the perpetrators, and illicit drugs.

8. Erica Scharrer, Lisa M. Weidman, and Kimberly L. Bissell, "Pointing the Finger of Blame: News Media Coverage of Popular-Culture Culpability," Journalism & Communication Monographs 5 (2, 2003): 49-98.

9. Benjamin Frymer, "The Media Spectacle of Columbine: Alienated Youth as an Object of Fear," American Behaviorai Scientist 52 (10, 2009): 1387-1404.

10. David L. Altheide, "The Columbine Shootings and the Discourse of Fear," American Behavioral Scientist 52 (10, 2009): 1354-70.

11. Patricia Leavy and Kathrjoi P. Maloney, "American Reporting of School Violence and People Like Us: A Comparison of Newspaper Coverage of the Columbine and Red Lake School Shootings," Critical Sociology 35 (2, 2009): 273-92.

12. Dick Weissman, "Some Thoughts on the Columbine Shootings," Popular Music & Society 23 (autumn 1999): 29-30.

13. Frymer, "Media Spectacle of Columbine."

Park et al. 493

14. Ronald L. Jackson, "Black Masculine Scripts," in Scripting the Black Masculine Body: Identity, Discourse, and Racial Politics in Popular Media 2006, ed. Ronald L. Jackson (Albany: State University of New York Press, 2006), 81.

15. Chris Greer and Yvonne Jewkes, "Extremes of Otherness: Media Images of Social Exclu- sion," Social Justice 32 (1, 2005): 20-31.

16. John Hagan, Ron Levi, and Ronit Dinovitzer, "The Symbolic Violence of the Crime- Immigration Nexus: Migrant Mythologies in the Americas," Criminology & Public Policy 7 (February 2008): 95-112.

17. Maxwell E. McCombs and Salma I. Ghanem, "The Convergence of Agenda Setting and Framing," in Framing Public Life: Perspectives on Media and Our Understanding of the Social World, ed. Stephen D. Reese, Osear H. Gandy, Jr., and August E. Grant (Mahwah, NJ: Lawrence Erlbaum, 2001), 67-81.

18. Brendan Maguire, Géorgie Ann Weatherby, and Richard A. Mathers, "Network News Coverage of Sehool Shootings," Social Science Journal 39 (3, 2002): 465-70.

19. Chyi and McCombs, "Media Salience and the Process of Framing." 20. Muschert and Carr, "Media Salience and Frame Changing aeross Events." 21. Glenn W. Muschert, "Frame-Changing in the Media Coverage of a School Shooting: The

Rise of Columbine as a National Coneem," Social Science Journal 46 (1, 2009): 164-70. 22. Kyounghee Hazel Kwon and Shin-Il Moon, "The Bad Guy Is One of Us: Framing Com-

parison between the US and Korean Newspapers and Blogs about the Virginia Tech Shooting," Asian Journal of Communication 19 (September 2009): 270-88.

23. Chyi and McCombs, "Media Salienee and the Process of Framing," 25-26. 24. Chyi and McCombs, "Media Salience and the Process of Framing," 25 25. Muschert, "Frame-Changing in the Media Coverage of a School Shooting." 26. Muschert, "Frame-Changing in the Media Coverage of a School Shooting," 166. 27. Muschert, "Frame-Changing in the Media Coverage of a School Shooting," 166. 28. Chyi and MeCombs, "Media Salienee and the Process of Framing." 29. Robert M. Entman, "Framing Bias: Media in the Distribution of Power," Journal of Com-

munication 57 (Mareh 2007): 163-73. 30. Stephen D. Reese, "The Framing Project: A Bridging Model for Media Research Revis-

ited," Journal of Communication 57 (Mareh 2007): 148-54. 31. David H. Weaver, "Thoughts on Agenda Setting, Framing, and Priming," Journal of

Communication 57 (Mareh 2007): 142-47. 32. Dietram A. Seheufele and David Tewksbury, "Framing, Agenda Setting, and Priming: The

Evolution of Three Media Effects Models," Journal of Communication 57 (Mareh 2007): 9-20. 33. Chyi and McCombs, "Media Salience and the Process of Framing." 34. Chyi and MeCombs, "Media Salienee and the Proeess of Framing"; Musehert and Carr,

"Media Salienee and Frame Changing across Events." 35. For this difference, Musehert and Carr reasoned, "[C]ommunities (other than the actual

sites of the shootings) were being discussed as part of the diseourse surrounding the rel- evanee of the partieular shooting for eommunities aeross the nation." Musehert and Carr, "Media Salienee and Frame Changing aeross Events," 764.

494 journalism & Mass Communicatíon Quarterly 89(3)

36. Christopher L. Aberson and Tara E. Ettlin, "The Aversive Racism Paradigm and Responses Favoring African Americans: Meta-analytic Evidence of Two Types of Favoritism," Social Justice Research 17 (March 2004): 25-46.

37. Ki-Young Lee and Sung-Hee Joo, "The Portrayal of Asian Americans in Mainstream Magazine Ads: An Update," Joumalism & Mass Communication Quarterly 82 (autumn 2005): 654-71.

38. Chyi and McCombs, "Media Salience and the Process of Framing." 39. Chyi and McCombs, "Media Salience and the Process of Framing." 40. Daniel Riffe, Stephen Lacy, and Frederick G. Fico, Analyzing Media Messages: Using Quan-

titative Content Analysis in Research, 2nd ed. (Mahwah, NJ: Lawrence Erlbaum, 2005). 41. Paul D'Angelo, "News Framing as a Multiparadigmatic Research Program: A Response to

Entman," Journal of Communication 52 (December 2002): 870-88.

Journalism & Mass Communication and its content may not be copied or emailed to multiple sites or posted to a

listserv without the copyright holder's express written permission. However, users may print, download, or

email articles for individual use.