| by M. Mark Miller , Julie Andsager Hate speech tends to be covered as a campus issue with emphasis on speech codes, events and absence of discussion of First Amendment implications. In the pressures of a free society, the media have ample opportunities to elucidate issues of free expression. Hate speech provides one such opportunity; however, some researchers have argued that journalists seldom place concrete events in the larger context of democratic philosophy.(1) This is particularly ironic given their vested interest in free expression and First Amendment ideals. But the news media do not hold sole interest in free expression - the public also has a stake in maintaining its free-speech rights. In addition to protecting their own turf, then, journalists have a unique opportunity to serve the public need in reporting on First Amendment issues. Events such as the April 1995 bombing of the Alfred P. Murrah Federal Building in Oklahoma City and the October 1995 Million Man March in Washington, D.C., thrust the issue of hate speech onto the national agenda. Both events were preceded by virulent speech acts. Following the Oklahoma City bombing, news media documented the statements by radio talk show hosts like Watergate felon G. Gordon Liddy, who instructed listeners on how to kill federal Alcohol, Tobacco and Firearms officials, and Chuck Baker, who called for "an armed revolution" against the federal government. News coverage of the Million Man March was dominated by commentary on the appropriateness of participating in an event organized by Nation of Islam leader Louis Farrakhan, who is notorious for his attacks on Jews and whites. Many researchers hold that the ways that news media frame issues have substantial impact on such things as legitimacy of political groups(2) and the outcome of policy debate.(3) Thus, newspaper coverage of First Amendment issues may provide readers with a context for forming either pro- or anti-regulatory attitudes about certain forms of expression. In the early 1990s, political correctness hit the media spotlight. The idea of being inclusive and sensitive to others, usually the historically oppressed - and all that that entails, from multiculturalism in university curricula to using non-offensive terminology - spread from university campuses to the media to the general public sphere.(4) Books championing the political correctness movement competed in the marketplace with those from the anti-PC side. At issue was the right to freely express ideas versus the right to the opportunity for equal, or at least fair, treatment. The inception of political correctness on university campuses spurred media coverage of academe, according to both journalists and scholars. D. Charles Whitney and Ellen Wartella's search of major newspaper indexes found that the term political correctness appeared nearly 4,000 times in 1991, compared to 638 instances in the previous year; however, they did not search for the occurrence of the term university within the same articles? In 1991, mainstream magazines and syndicated columnists focused so much attention on the PC debate as to make it the top story on higher education.(6) Those who have studied media coverage of political correctness suggest several reasons for the press's fixation on portraying PC as a campus issue.(7) Huntly Collins argued that political correctness became a big story for two reasons: First, university issues became more complicated and budget-centered, and academic research became more esoteric, fueling anti-intellectualism in the popular press.(8) The second reason is economic: As readership declined, newspapers found political correctness on campus as a way to sensationalize often-mundane academic coverage by playing on the "deepest fears of white, middle-class Americans - the very segment of the population that [they] must attract if they are to remain economically viable."(9) Increasing pressures to diversify the journalistic workforce itself may have indirectly caused threatened editors to play up the political correctness debate? More recently, political correctness has been eclipsed by its more virulent cousin, hate speech. Hate speech includes the use of hurtful, biased expression; threats of violence based on sex, race, ethnicity, religion and sexual orientation; and offensive songs, jokes or events, such as fraternity slave auctions.(11) While political correctness tends to be a matter of self-regulation or suppression, hate speech has been the subject of state laws and university speech codes. Arati Korwar's survey of universities across the country found that slightly more than one-third had codes barring verbal harassment of members of certain groups, though the codes seldom hold up when their constitutionality is challenged.(12) The emergence of hate speech as an issue was particularly apparent during 1992. In early 1992, Whitney and Wartella mentioned some universities' "development of sanctions against hate speech."(13) As the year progressed, though, the term hate speech became widespread. The Supreme Court struck down as unconstitutional a hate-crimes statute in its June 1992 decision in R.A.V.v. City of St. Paul, under which a group of white teenagers received extra penalties for burning a cross in an African-American family's yard.(14) Given the media's perhaps disproportionate attention to political correctness, it is not unreasonable to expect similar treatment of hate speech. The purpose of this study is to analyze how newspapers presented the issue of hate speech; to determine the kinds of events that drove news coverage; and to discern the degree to which news stories and opinion pieces (editorials and columns) differed. While news stories may sometimes be limited to the events at hand, editorial writers and columnists have greater latitude to contextualize hate-speech incidents in terms of the First Amendment and its democratic principles. This study addressed the following questions: * Was coverage of hate speech driven by events, or was there sustained discussion of hate speech to elucidate public discourse? * Did newspaper editorials and columns, which are less constrained by norms of objectivity, place hate speech in the broader context of principles of democracy? Method To examine newspaper coverage of hate speech, a computer database was used to compile complete articles containing the words hate speech that appeared in major newspapers from Jan. 1, 1992 to Dec. 31, 1994. The sample included articles from some 20 newspapers of varying sizes across the nation, including the New York Times, the Los Angeles Times, the Houston Chronicle, the Cleveland Plain-Dealer, etc. Most of the cities included have at least one university located in their metropolitan area, but they are large and diverse enough to be beyond the scope of a college town. Starting at the beginning of 1992 showed the impact on coverage of the R.A.V.v. St. Paul Supreme Court decision in June of that year; the next two years incorporated a number of other important hate-speech issues, perhaps most notably the furor over Khalid Muhammad's speaking tour in late 1993 and early 1994. This search strategy resulted in 307 articles (160 news stories, 108 columns and 38 editorials). About three-fourths of the articles (75.6 percent) were of local origin, and the remaining 24.4 percent came from wire services or syndicates. The annual number of articles produced was stable - 96 from 1992, 92 from 1993 and 116 from 1994. The articles were checked for spelling and prepared for analysis in a computerized concept-mapping program called VBPro.15 Each article was coded for time period (January to June, 1992; July to December, 1992; and so on for each year) and article type (news, column or editorial). The prepared text was submitted to the computerized content analysis program. After VBPro calculated the frequency of each word, the 107 most frequently occurring and most interpretable terms were selected. Vague or ambiguous terms such as leaders and states, though they occurred frequently, were eliminated to increase interpretability. Synonyms were put together in a search list, such that VBPro would code for occurrence of any of the combined terms. Thus, university, college and campus were considered synonyms, based on face validity. The selected terms are listed in Chart 1. The VBPro coder was then used to produce a data matrix indicating how many times each term occurred in each article. This data matrix was submitted to the VBPro concept-mapping program, which uses mathematical procedures to create dimensions that can be used to plot the terms on a concept map. Terms appear close together on the map if they tend to occur together in the same stories and far apart if they do not occur together. The concept map indicated, for example, that the terms university and professor tended to co-occur, but they rarely occurred with the terms women and pornography. Because using all 107 terms would result in cluttered maps and unwieldy analysis, the dimensions were submitted to a statistical cluster analysis program? The cluster analysis grouped terms that were closely related based on their occurrences within stories. It was concluded that the most interpretable solution provides eight clusters of at least five terms each. The combinations in each cluster, as well as the newspaper articles in which they occurred, were examined to come up with descriptive labels for them. Article type and time period codes were gathered from the articles themselves. These codes were incorporated into an SPSS file along with the eight term clusters. Analytical statistics were then used to address the research questions. Results and discussion The VBPro program was used to code each case for the total frequency of all terms in each cluster. These coded data were submitted to the concept-mapping program to determine the relative position of the clusters. The resulting concept map is shown in Figure 1. On the map, the labeled clusters are close together to the degree that the terms in them tend to co-occur. Hate speech, speech and codes fell into the first cluster. This cluster also comprised Wisconsin, referring to the University of Wisconsin's speech code, which was being challenged during the time frame. Words related to hate-speech codes also appeared in this cluster - legal, conduct, minority- so the cluster was named Speech Codes. Three other clusters mapped near Speech Codes. Given the terms they comprise, this is not surprising. University Issues was located nearer Speech Codes than any other cluster. It included such terms as university, campus, students, faculty and professors, but also incorporated broader terms that seemed to have been frequently used in discussing hate speech. Thus, racist, offensive, diversity and debate grouped into this cluster. Equidistant from Speech Codes but in the opposite direction from University Issues lay a cluster that incorporated more abstract terms. This cluster included free speech, rights, American and policy; this cluster was named American Ideals. Its nearest cluster, Political Correctness, fell slightly farther from Speech Codes and was fairly distant from University Issues. This cluster included the terms PC, harassment, society and police (which probably came from the term thought police). These four clusters appear to encompass hate speech itself. That the university-related terms grouped in most of these clusters suggests that the concept of hate speech was seen primarily as a campus or campus-created issue. An intriguing aspect of this is the positioning of Speech Codes between American Ideals and University Issues. The newspapers, then, appeared to frame hate speech and hate-speech codes as a conflict between the university and the traditional democratic values of free expression. The remaining four clusters were quite distinct from those discussed above. At the center of the map was a cluster containing the terms First Amendment, Constitution, words, protection and American Civil Liberties Union. This cluster was called First Amendment because it encompassed the terms most related to abstract democratic principles guaranteed in the Bill of Rights. The nearest cluster to this was a group of terms dealing with the June 1992 R.A.V.v. St. Paul decision - Supreme Court, Scalia, majority, ordinance. This cluster, named R.A.V. Ruling, stood alone on the map, and it was the most distant from the four previously discussed clusters, especially University Issues. The clear distinction between the university frame and democratic principle suggests several interpretations. By framing the Supreme Court decision and the First Amendment as separate from the university frame, the newspapers may have reinforced the notion that Whitney and Wartella described of academia as a liberal threat to democracy.(17) The distance between speech codes and the R.A.V. case also suggests that newspaper coverage of hate speech issues is in line with Dorothy Bowles' finding that historically the media have done a poor job of explaining court decisions to the public in a contextual way.(18) The last two clusters were consistent with University Issues on one dimension, perhaps because both had academic links. The closest was Divisive Speeches, which related to the university because it grouped the terms in the spate of news coverage on Khalid Muhammad's college speaking tour. Muhammad, a member of the Nation of Islam and former spokesman for Minister Louis Farrakhan, delivered anti-Semitic speeches at a number of universities around the nation, but his most notable appearances were at Kean College in November 1993 and at Howard University in April 1994. This largest cluster contained black, white, Jewish and individual names related to the controversy. The last cluster, Gender and Sex, contained the terms women, men, sex, pornography and censorship. Its proximity to University Issues was weaker, in that only a bit more than half of the cases containing these terms referred to campus issues, usually with regard to sexual harassment or required reading lists. This cluster was the most detached from the others. Isolating the racial/ethnic conflict and gender clusters from the other frames suggests that these groups were defined as existing outside of the discourse on hate speech versus free speech. That is, because the racial/ethnic and gender clusters were not located near either the university frame nor the democratic principles frame, the newspapers studied seemed to marginalize these groups. This is ironic, given that hate speech is directed at women or individuals from different racial or ethnic groups. Coverage over time Using the coded cases, a series of one-way analyses of variance tracked differences in newspaper coverage of the various clusters during the three-year period studied. This procedure allowed determination of whether differences in coverage over time were real or if they occurred by chance. Time period, defined in six six-month groups, was the independent variable, with the concept clusters serving as dependent variables. Scheffe' post hoc tests were used to determine differences between groups. The ANOVAs were significant for all clusters but one, Political Correctness. Time period had a significant effect on three clusters at the p [less than] .0001 level. (Means for each cluster by time period are displayed in Table 1.) Post hoc tests on R.A.V. Ruling indicated that coverage of this cluster was significantly higher in January-June 1992 than any subsequent period, with an average of 22.9 occurrences of its terms per case? Occurrences of these terms during the second half of 1992 were also significantly higher than they were during 1993 and 1994. First Amendment was also affected by time? Like the R.A.V. cluster, its incidence was significantly higher in the January-June 1992 period than in nearly all of the other periods. Given the proximity of these two clusters on the concept map, many of the First Amendment terms must have co-occurred with the R.A.V. Ruling terms. Divisive Speeches showed a similar pattern of peaking dramatically in one six-month period.(21) Mentions of terms in this cluster were relatively flat until January-June 1994, which was significantly higher than the remaining five periods; at a mean of 27.7 mentions per case, the terms occurred more than three times as often than the next highest period. A glance at the cases during the period indicates this finding can be attributed to Muhammad's speech at Howard University. [TABULAR DATA FOR TABLE 1 OMITTED] Time period affected American Ideals, such that the 1993 groups had significantly higher means than January-June 1994.(22) The means in Table 1 show that Political Correctness also reached its peak during 1993, which supports their spatial relationship on the concept map. It is likely that the tremendous jump in coverage of Divisive Speeches displaced American Ideals from the discourse in early 1994 because the abstract terms included in the American Ideals cluster may not have clear connections to the very concrete terms encompassed by the speeches cluster. Certainly the concept map suggests a gap between the two. While journalists' vested interest in the freedom of expression guaranteed by the First Amendment would suggest that occurrences of terms related to democratic principles and free speech would remain consistently high in news coverage of hate speech, the sporadic means for the First Amendment and American Ideals clusters indicate that those principles did not figure prominently in the news stories studied. The other two clusters in the university frame - University Issues(23) and Speech Codes(24) - were also significantly affected by time at the .05 level. Both clusters had their highest means in January-June 1993, although that period was not significantly higher than the rest for Speech Codes. Concern over hate-speech codes on college campuses seemed to have pushed the university frame into the newspapers, sparked by the University of Wisconsin debate over its speech code (and especially by former UW chancellor Donna Shalala's confirmation hearings for Secretary of Health and Human Services) and the water buffalo incident at the University of Pennsylvania. According to the means in Table 1, newspaper coverage of the speech codes at universities was increasing while coverage of the Supreme Court ruling that struck down a similar city ordinance was waning. If this was the case, it appears that journalists did not provide a broader context of the import of hate-speech codes for their readers. The mean occurrences of the terms in the clusters, moreover, indicate that University Issues received more mentions than any other cluster in two-thirds of the time periods. Because the cases included in this data set were selected for the words hate speech, this suggests that the media closely associated hate speech with the university; the concept map reinforces the finding. The cluster situated at a distance from the rest, Gender and Sex, varied significantly by time, with peak occurrences coming in January-June 1993.(25) The mean number of terms occurring in this cluster for each time period indicate that Gender and Sex was the least frequently covered issue in half of the six-month periods. Along with the concept map, this finding suggests that issues of sex and gender - women's issues in particular, since women and woman were included in the cluster - were not as newsworthy as other topics. Overall, the finding that seven of the eight clusters were significantly affected by time period indicates that newspaper coverage of hate speech and its associated issues was heavily event-driven. The peaks in coverage of each cluster can be traced to specific events occurring during the period. The most notable examples are the R.A.V. Ruling and Divisive Speeches clusters. Only Political Correctness coverage was stable over the three-year period, but it remained near the bottom of topics mentioned; no clearly defined events occurred relating to P.C. during the years studied. Type of article Straight news stories are expected to be event-driven, but columns and editorials are more free to place hate-speech issues within a context of values. To examine occurrence of the term clusters by type of article, a series of one-way ANOVAs was performed using article type - news, column or editorial - as the independent variable. Again, the term clusters served as dependent variables. Article type had a significant effect on three of the clusters. For University Issues, the effect was the strongest? News stories were most likely to contain terms comprised by University Issues, while editorials were least likely. (See means for each of the three clusters by article type displayed in Table 2.) This disparity suggests that while universities may have spurred the coverage on hate speech through news events, journalists did not seem to find those events important enough to merit editorial comment, either in columns or editorials. The university frame was primarily developed, then, in news itself. | |
Table 2: Mean occurrence of hate-speech clusters by type of article.
Cluster News Columns Editorials
Stories
University Issues 20.39(*) 12.11 8.32
R.A.V. Ruling 9.93 6.49 4.92
American Ideals 8.21 9.06 4.76(*)
* Mean is significantly different than both others in the row at
p [less than] 05.
| | Although it is beyond the scope of this study, future research might examine the flaming effects of the placement of news stories on hate speech; because casual readers may be more likely to see headlines for news stories on the first few pages of the newspaper, the university frame may have been particularly pronounced. R.A.V. Ruling was significantly affected by type of article, but post hoc testing showed no significant differences between pairs of means.(27) The trend was in the same direction as that of University Issues, however: The decision was most likely to be covered in news stories and least likely to be mentioned in editorials. This is not surprising, because many papers carried the decision with a brief explanation of the court's rationale. In a historical analysis of editorial coverage of Supreme Court decisions on free-speech cases, Bowles found that editorials did not explain well the implications of those decisions.(28) Similarly, Doris Graber has charged that news stories do not necessarily provide an adequate context for readers' understanding of the broader scope of the Supreme Court's rulings, mostly due to the complexity of the rulings themselves.(29) Perhaps an increased attention from columns or editorials might have framed the R.A.V. ruling nearer the other hate-speech frames and contributed to the public's understanding of their underlying connections. Type of article had a significant effect also on the American Ideals cluster? While editorials were significantly less likely to address these terms than the other types of articles, discussion of American Ideals was slightly more likely to occur in columns than in news stories. Journalists' inherent interest in free speech and rights would suggest that, contrary to these findings, editorials might contain more instances of the terms included in American Ideals. It might be expected to see the terms appear in editorials at least as often as in news stories because editorial writers have greater leeway to incorporate more abstract ideas than journalists writing straight news. Future studies might investigate if this finding is an artifact of hate-speech coverage or whether it holds true across First Amendment issues. Conclusion Although hate speech offers journalists an excellent opportunity to discuss freedom of expression and its place in a diverse, democratic society, these findings suggest that little discussion has actually appeared in major newspapers. Indeed, terms related to the First Amendment seldom occurred in the newspapers studied after the initial flurry of reports surrounding the R.A.V. v. St. Paul ruling in 1992. That coverage of the various clusters fluctuated by time period suggests that newspapers tended to report events without necessarily linking them to related issues. By separating the clusters over time, journalists may have inhibited public comprehension of the hate-speech issue. Similarly, the finding that news stories devoted more attention than opinion pieces - which offer more opportunity for interpretation of issues - to several of the clusters supports this. If, as researchers have argued, the ways that news media frame issues can shape policy and confer legitimacy to various groups, hate-speech coverage would appear to serve as an exemplary case study. Other researchers have noted that the media's tenacious coupling of political correctness and academe reflected an attack against the liberal university by the political Right. This study suggests that, in much the same way, newspapers have tended to position hate speech as a campus issue, with speech codes implicating the university as an opponent of the democratic principle of freedom of expression. Political correctness - at least during the early 1990s - remained fairly close to the university frame. If political correctness has been used derisively against higher education, hate speech would appear to be moving in the same direction. This study suggested that hate speech has been discussed in terms of various groups. While members of those groups may be the targets of hate speech (given their minority status), the concept map indicates that African-Americans, Jews and women have seldom been granted legitimacy in the hate-speech debate. Their clusters were far removed from those connoting authority, such as the Supreme Court's ruling on a hate-speech ordinance, the First Amendment and democratic principles. Distancing minority groups from the more general discussion of hate-speech issues reinforces their position as outsiders. This analysis does not account for the valence of the terms, however, and this contention is based upon a reading of the news coverage. This study suggests that, in the first half of the 1990s, newspapers tended to discuss hate speech as primarily a university-related issue, often focusing on speech codes and controversial events that occurred on campuses. Since 1995, however, hate speech seems to have spread from academe to the general public sphere; witness the Oklahoma City bombing, President Clinton's strong stand against talk radio and the uproar over Louis Farrakhan's surge in visibility. Future research needs to examine whether these news events will dislodge the university's hold on hate speech and lead journalists to examine the issue in a much needed broader context. Chart 1: Terms forming the clusters SPEECH CODES speech: hate speech; code, codes; legal, Wisconsin, conduct, minority, minorities, ethnic UNIVERSITY ISSUES university, universities, university's, college, colleges, college's, school, schools, campus, campuses; student, students; faculty, professors; debate; education; racist; offensive; action; academic; newspaper; issues; administration; administrators; rules; diversity DIVISIVE SPEECHES black, blacks, African-American, African-Americans; white, whites; Jews, Jewish; Muhammad, Muhammad's; Howard (University); hate; Nation of Islam; (Louis) Farrakhan; remarks; bigotry; history; media; speak; anti-Semitism, anti-Semitic; Kean (College); man; Jenifer (former Howard president Franklyn); Anti-Defamation League RAV RULING law, laws; Supreme Court; decision; ordinance; St. Paul; justices; ruling; majority; (Justice Antonin) Scalia; crime, crimes; gender; bias; hateful FIRST AMENDMENT First Amendment; word, words; race, racial; civil; fighting; constitution, constitutional; protection; American Civil Liberties Union, ACLU; color; discrimination; punish AMERICAN IDEALS right, rights; free speech; public; expression; policy; federal; American; lawyer, lawyers; ideas; conservative; California POLITICAL CORRECTNESS harassment; political; free; society; political correctness, PC; support; police GENDER & SEX women; sexual; violence; pornography; men; sex; censorship; woman Notes 1. Dorothy Bowles, Missed Opportunity: Educating Newspaper Readers About First Amendment Values. Newspaper Research Journal, Winter 1989, pp. 39-53; Doris A. Graber, Mass Media and American Politics, 3rd. Washington, D.C.: CQ Press, 1989. 2. See, for example, Todd Gitlin, The Whole World is Watching: Mass Media and the Making and Unmaking of the New Left. Berkeley, California: University of California Press, 1980. 3. Herbert J. Gans, News Media, News Policy, and Democracy: Research for the Future. Journal of Communication, Summer 1983, pp. 174-184; Zhongdang Pan and Gerald M. Kosicki, Framing Analysis: An Approach to News Discourse. Political Communication, 1993, pp. 55-75; Gaye Tuchman, Making the News: A Study in the Construction of Reality. New York: The Free Press, 1978. 4. Paul Berman, ed., Debating P.C.: The Controversy over Political Correctness on College Campuses. New York: Dell, 1992. 5. D. Charles Whitney and Ellen Wartella, Media Coverage of the "Political Correctness" Debate. Journal of Communication, Spring 1992, pp. 83-93. 6. Huntly Collins, PC and the Press. Change, January/February 1992, pp. 12-16. 7. Whitney and Wartella, Media Coverage, op. cit., pp. 85-86. 8. Collins, PC and the Press, op. cit., pp. 15-16. 9. Ibid.,p. 16. 10. Whitney and Wartella, Media Coverage, op. cit., p. 88. 11. Arati R. Korwar, War of Words: Speech Codes at Public Colleges and Universities. Nashville, Tennessee: The Freedom Forum First Amendment Center, 1994. 12. About 72 percent of universities surveyed responded to Korwar's survey. She analyzed their student handbooks, most of which were from 1992-93 or 1993-94. 13. Whitney and Wartella, Media Coverage, op. cit., p. 85. 14. R.A.V. vs. City of St. Paul, 505 U.S. __, 112 S. Ct. 2538 (1992). 15. M. Mark Miller, User's Guide to VBPro. Knoxville, Tennessee: M. M. Miller, 1993. 16 The SPSS for Windows program was used, specifying the cosine distance metric for the clustering algorithm. While only three dimensions can be plotted using normal methods, the VBPro map program, which uses some of the mathematics of relativity theory in physics, can calculate as many dimensions as there are terms. We examined the clustering using up to 10 dimensions and concluded that the clustering solution using four dimensions yielded the most interpretable solution. 17. Whitney and Wartella, Media Coverage, op. cit., p. 88. 18. Bowles, Missed Opportunity, op. cit., p. 47. 19. F5, 303 = 26.08; p [less than] .0001. 20. F5, 303 = 7,40; p [less than] .0001. 21. F5, 303 = 19.20; p [less than] .0001. 22. F5, 303 = 5.16; p [less than] .001. 23. F5, 303 = 2.90; p [less than] .05. 24. F5, 303 = 2.64; p [less than] .05. 25. F5, 303 = 4.19; p [less than] .01. 26. F2, 305 = 9.28; p [less than] .001. 27. F2, 305 = 3.65; p [less than] .05. 28. Ibid. 29. Graber, Mass Media and American Politics, op. cit., pp. 266-267. 30. F2, 303 = 4.58; p [less than] .05. Miller is professor of journalism at the University of Tennessee in Knoxville and Andsager is assistant professor of journalism at Middle Tennessee State University in Murfreesboro. |
|
