Local television news is America’s principal window on the world. Surveys of television viewing (e.g., Roper-Starch, 1994), hours of daily programming (Papper and Gerhard, 1997) and the actual share of the viewing audience captured by local newscasts (Hess, 1991), all demonstrate the dominance of local news. In fact, people can watch live local news almost anytime -- mornings, afternoons, evenings, prime time, and late night. As the amount of news time has increased, so has competition between stations. The drive for audience ratings pushes local news organizations to favor an "action news" format.

Stories about crime provide the necessary ingredients for successful marketing of newscasts – a focus on concrete events, their impact on ordinary people, drama and emotion, and, above all, attention-getting visuals. The special attraction of television to crime is reflected in the content of local television news. In a recent study of fifty-six different cities, no matter how large the media market, crime was the most prominently featured subject in the local news (Klite, Bardwell, and Salzman., 1997). In some cities, crime accounted for more than 75 percent of all news coverage.

Local news coverage of crime follows a standard "script." The presentation has several elements. First, as seen in the news, crime is violent (Elias, 1994; Crispin-Miller, 1998). Second, coverage is episodic in the sense that the news focuses on discrete events rather than collective outcomes or general context (Iyengar, 1991). Third, crime "episodes" often feature a central causal agent, namely, the suspected perpetrator. Given the visual nature of the medium, the importance of the suspect to the script means that crime news is often accompanied by racial imagery (Entman, 1990, 1992; Gilliam, Iyengar, Simon, and Wright, 1996; Gilliam and Iyengar, 1997; Graber, 1976; Peffley, Shields, and Williams, 1996; Romer, Jamieson, and de Coteau, 1998; Worthy, Hagan, and MacMillan, 1997); more often than not the suspect is non-white. In short, the crime news script has readily identifiable elements – crime is violent, and there is an individual perpetrator whose ethnicity is made known to the audience.

Our objective in this paper is to evaluate the relative contribution of two specific elements of the crime news script – violence and race – on public opinion. In general, we find that the latter cue is more influential. Viewers exposed to the "racialized" version of the script become more supportive of capital punishment, mandatory sentencing and other deterrent measures. Not unexpectedly, exposure to this version of the script also serves to substantiate old-fashioned and contemporary racial stereotypes. In closing we consider the implications of these results for intergroup relations, electoral politics, and the practice of journalism.

The theoretical basis for our expectations concerning the effects of crime news on the viewing audience derives from the concept of "scripts." As developed by cognitive psychologists, a script is "a coherent sequence of events expected by the individual, involving him either as a participant or as an observer" (Abelson, 1976, p. 33; also see Abelson 1981; Schank 1990; Schank and Abelson, 1977; Mandler, 1984). In their pioneering work, Schank and Abelson (1977) described "behavioral" scripts, such as, for example, going to a restaurant: people "know" that they eat first, and pay later. Other researchers have expanded the concept to embrace "narrative" or text-based scripts which appear in fiction, humor, advertising and, of particular interest to us, television news reports (see Sulin and Dooling, 1974; Black, Galambos, and Read, 1984; Graesser, Woll, Kowalski, and Smith 1980). Indeed, scripts are characteristic of all forms of story telling. In the case of mystery novels, for example, the "Agatha Christie" script leads readers to expect (in order) a murder, the appearance of assorted suspects and clues, and the final denouement (orchestrated by Monsieur Poirot or Miss Marple) in which all is explained.

All scripts, either behavioral or narrative, facilitate comprehension by distilling experience and knowledge. Because they provide an orderly and quite predictable set of scenarios and roles, scripts allow the "reader," quite effortlessly, to make inferences about events, issues, or behaviors. Because the "target" actions are marked by sequence, there is a clear sense of what is to come. We do not need to see the customer paying the bill or ordering the food to know that this follows the reading of the menu. We do not need to see police officers at the crime scene to be aware of their presence. In many cases script-based expectations are so well developed that when people encounter incomplete versions of the script, they actually "fill in" the missing information and make appropriate (that is, script-based) inferences about what must have happened. In the case of the restaurant script, for instance, the sight of an individual seated at a table reading a text leads observers to understand that this is a customer attempting to decide what to eat. Our evidence indicates that for viewers in Los Angeles and across the country, the expectations prompted by the crime script have achieved the status of common knowledge. Just as we know full well what happens in a restaurant, we also know -- or at least think we know -- what happens when crime occurs. Order and regularity make scripts a powerful tool for human understanding.

As told by television news, the crime news script has three ordered segments. It usually begins with the anchorperson’s terse announcement that a crime has occurred. The viewer is then transported to the scene of the crime for a first hand look supported by accounts from bystanders, relatives of the victim, or other interested parties. Finally, the focus shifts to the identity and apprehension of the perpetrator and the related efforts of law enforcement officials.

The following example, taken from a recent report aired by Los Angeles Channel 9 (KCAL) is typical:

Anchor’s introduction: "A man was shot this afternoon in broad daylight while sitting in his jeep."

Crime Scene Coverage: pictures of jeep and cordoned-off street; concerned neighbor comments ("Imagine something like this happening just in front of your house; I mean, it’s really scary.")

Apprehension of suspect: "Police are looking for this man last seen driving away in a blue Honda Accord (picture of suspect on screen). Police believe the suspect may have argued with the victim before he was shot."

As we noted earlier, within this brief presentation (the entire story runs for 90 seconds) there are two underlying regularities. First, violent crime is a central script element Spectacular armed bank robberies, homicides, "home invasions," carjackings, police chases and gang-related activities are now a staple of local newscasts. Not surprisingly, little time is devoted to non-violent crimes such as embezzlement, fraud, or tax evasion because they lack the "action" found in most violent crime stories (Gilliam, Iyengar, Simon, and Wright, 1996; Klite et al., 1997). Second, episodic reporting requires a regular "cast" of characters -- the victim or the victim’s family, the suspect, the bystander, and law enforcement officials. The suspect is typically the most important player because he (but sometimes she) is the motivating force behind the news story’s reconstruction of the event. The only information provided about the suspect concerns race or gender.. In sum, as depicted in the local news, crime is violent and criminal behavior is associated with race/ethnicity.

In the next section we present a detailed content analysis that examines the prevalence of the crime news script in the Los Angeles television news market. We then report the results of several experiments in which viewers encountered different versions of the crime script. Finally, we corroborate the experimental results using a survey of Los Angeles County residents.

The centrality of violent crime to local news programs was readily apparent in our study that encompassed all English-language commercial television stations operating in the LA market. These stations aired a total of 3014 news stories on crime during 1996 and 1997 (when we administered our studies) of which 2492 (83%) were about violent crime.

The crime of murder, which accounts for less than 1% of all crime in Los Angeles County, was the focus of 17% of crime stories in the newscasts sampled. In fact, the number of murder stories (510) is equivalent to the total number of non-violent crime stories (522) during the period sampled. As seen in the news, crime is violent, and life is indeed nasty, brutish, and short.

The data in Table 1 also speak to the second element of the crime script -- the focus on individual perpetrators and minority perpetrators in particular. Over one-half (52%) of the violent crime reports made explicit reference to a suspect. Minorities accounted for two-thirds of those cases (768). In other words, when a violent perpetrator was identified in a news report there was a 66% chance that viewers would see a minority suspect. African-Americans (50%) and Latinos (36%) made up the largest percentage of minority suspects. Finally, for the extreme case of murder, minorities accounted for almost 75% of all suspects identified. Once again, African-Americans comprised the largest group of minority suspects (almost sixty percent).

Obviously, racial differences in television crime coverage are partly a reflection of the disproportionate representation of particular racial groups in criminal activity. The relative proportion of blacks in crime news, for example, is not that much out of line with the actual black arrest rate in Los Angeles County (although blacks do not account for the largest absolute number of murders). However, the media’s near exclusive focus on violent crime distorts the real world in the following way: when viewers encounter a suspect in the news he is invariably a violent perpetrator, when in reality the greatest number of felony arrests are for property crimes (Gilliam, 1998). To the extent that people do see nonviolent crime stories, the perpetrator is most typically white (recall the data in Table 1). In the real world, however, minorities actually account for the largest share of nonviolent (property) felonies (Gilliam, 1998). Clearly, the news is not an accurate reflection of the actual world of crime.

In short, the content analysis amply documents the scripted nature of crime news. As depicted in the news, crime is violent and more often than not, news stories about crime make reference to a suspect. Moreover, the race of the suspect varies by the type of crime. The more violent the crime, the more likely the depicted suspect to be a racial minority; and African-Americans constitute the largest share of minority suspects.

In the next section we appraise the impact of the violent crime and race of the perpetrator elements of the crime script on attitudes about crime and race respectively. We begin this section with a description of our experimental design. We move on to discuss the measurement of our dependent variables. We end the section by developing three competing hypotheses about the effect of the crime news script on the viewing public.

Design

We rely primarily on experimental methods to assess the effects of the crime news script. Experiments have the well-known advantage of greater precision in estimating causal effects. We designed the experiments in this study so that the only differences between any two groups of viewers concerned the relevant aspects of the crime news script -- the presence or absence of violent crime, and the race of the alleged perpetrator. Since all other properties of the news presentation were identical we can attribute the observed differences between conditions, if any, to the cues conveyed by the crime script.

Of course, experiments are not without their limitations. Most experiments are administered upon "captive" populations -- college students who must participate in order to gain course credit. Experiments also require a somewhat sterile, laboratory-like environment which bears little resemblance to the cacophony of the real world. Our own research was designed to overcome the artificial nature of the experimental method. As described below, our participants represented a fair cross-section of Los Angeles metropolitan area residents, our experimental manipulation consisted of an actual (and typical) news report on crime, and the experimental setting closely emulated the typical citizen’s encounter with local news.

The principal objective of our manipulation was to manipulate the main elements of the crime news script. Four levels of the manipulation were established. First, some participants watched a story in which the alleged perpetrator of a murder was an African-American male. Second, other subjects were given the same news report, but this time featuring a white male as the murder suspect. A third set of participants watched the news report edited to exclude information concerning the identity of the perpetrator. Finally, a control group saw no crime news story at all.

The most innovative aspect of this design concerns our ability to vary the race of a "target" face (in this case, the alleged perpetrator) while maintaining all other visual characteristics. The original "input" was a local news report which included a close-up "mug shot" of the suspected perpetrator of the crime in question. The picture was digitized, then "painted" to alter the perpetrator’s skin color, and then re-edited into the news report. Beginning with two different perpetrators (a white male and a black male), we were able to produce altered versions of each individual in which their race was reversed, but all other features remained identical. Thus, the perpetrator featured in the "white" and "black" versions of the story was equivalent in all respects but race. Using this method, any differences in the responses of the subjects exposed to the white or black perpetrators can only be attributed to the perpetrator’s race.

Participants watched a fifteen-minute videotaped local newscast (including commercials) described as having been selected at random from news programs broadcast during the past week. The objective of the study was said to be "selective perception" of news reports. Depending upon the condition to which they were assigned (at random), they watched a news story on crime that included a close-up photo of the suspect. Using the method described above, the photo either depicted an African-American or white male. The report on crime was inserted into the middle position of the newscast following the first commercial break. Except for the news story on crime, the newscast was identical in all other respects. None of the remaining stories on the tape concerned crime or matters of race.

On their arrival, participants were given their instructions and then completed a short pre-test questionnaire concerning their social background, party identification and political ideology, level of interest in political affairs, and media habits. They then watched the videotaped newscasts. The viewing room was furnished casually and participants were free to browse through newspapers and magazines, snack on cookies, or chat with fellow participants. At the end of the videotape, participants completed a lengthy questionnaire that included questions about their evaluations of various news programs and prominent journalists, their opinions concerning various issues in the news, their reactions to particular news stories and, depending on the study, questions tapping their beliefs about the attributes of particular racial/ethnic groups. After completing the questionnaire, subjects were debriefed in full and were paid the sum of fifteen dollars.

Using this basic design, we have administered five separate studies between April 1995 and November 1997. Study 1 was administered at the UCLA Media Research Laboratory (which consists of a two-room suite on campus). Studies 2, 3,and 6 were conducted at a major shopping mall in the city of Los Angeles. Studies 4 and 5 were conducted at a smaller mall in an outlying section of the metropolitan area known as Simi Valley which is located in Ventura County. Each study was designed (in part) to evaluate different attitudes about crime and race. Studies 1, 2, 4 and 6 addressed attitudes towards the criminal justice process in general. Study 3 focused on questions of juvenile crime. While Studies 2 and 3 focused on traditional racial stereotypes, Studies 4 and 5 were designed to investigate the effects of the crime news script on more subtle racial attitudes. Finally, participants in Studies 1, 3, and 4 completed measures of "free recall" of the crime news story that enables us to validate the experimental manipulation and assess viewer’s reconstruction of the news story. To maximize the reliability of the analysis, we pooled all five experiments. However, because some indicators were not common to all five studies, the number of cases varies across analyses.

The experimental "sample" consisted of 2331 residents of the Los Angeles metropolitan area who were recruited through flyers and announcements in newsletters offering $15 for participation in "media research". The age of the participants ranged from 18 to 74. Fifty-three percent were white, 22 percent were black, 10 percent were Asian, and 8 percent were Latinos (footnote about low Latino and Asian samples). Fifty-two percent were women. The participants were relatively well educated (49 percent had graduated from college) and, in keeping with the local area, more Democratic than Republican (45 percent versus 25 percent) in their partisan loyalty.

In order to assess the validity of our manipulation, we began by examining participants’ ability to recall the details of the news story. As noted above, a subset of subjects were asked to recall the content of the crime report. At the end of the questionnaire subjects completed a section which began with the following instructions: "Now we want to know what you remember and how you felt about some of the stories you just saw. On the next page, some of the stories are briefly described." Subjects were asked to recall what the story was about, their thoughts and reactions to the story, and to identify the race, age, and gender of the "suspect in the story". The question about the race of the suspect was used to construct a test of accuracy in recall by comparing across the three experimental conditions.

Table 2 presents the results of these comparisons. While subjects were generally accurate in their recall of the presence of a perpetrator (an average of about 67%), they responded more accurately in the black perpetrator condition (70%) than in the white perpetrator condition (64%). This difference is statistically significant at the .05 level

(t = 1.87). Similarly, subjects in the white perpetrator condition were about 50% more likely to be unable to recall seeing a suspect than subjects in the black perpetrator condition

(t = 1.71; p < .10). These results are in keeping with the extensive literature in social and cognitive psychology indicating that people are more likely to discard discrepant information and retain information that is consistent with their prior beliefs (see, for example, Graesser, Singer, and Trabasso, 1994; Roediger and McDermott, 1995).

Turning to the especially interesting case of the condition that did not feature a perpetrator, Table 2 shows that over 60% of the respondents who watched the story with no reference to a perpetrator falsely recalled having seen a perpetrator. Even more striking, in seventy percent of these cases, the perpetrator was identified as African-American. Taken together, these data reveal that the crime script generates strong expectations about crime, thus allowing viewers to fill in gaps in the script. Lacking concrete evidence about the perpetrator, viewers fall back on the crime script to infer what must have happened. Overall, the recall data validate the notion that the crime script is no mere journalistic device; instead, it is a powerful filter for observing daily events.

Measurement

Crime Attitudes: General trends in American public opinion suggest that the growing reach of local news has contributed to increased concern for crime, a belief that crime is a product of individual failings, and greater support for punitive remedies. For example, recent polling by CBS News and the New York Times shows that majorities continue to believe that

crime and lawlessness are a serious concern (CBS News/NY Times, 1994-1996). Similarly, a number of polls suggest that Americans believe that a "moral decline" and the "influence of drugs" are the primary causes of crime (see, Belden and Russonello, 1997). Growing national concern about the "crime problem" has meant that support for the death penalty has risen from about 50% in 1976 to almost 80% in 1996 (Lee and Ladd, 1997).

Preliminary analysis revealed that vulnerability to crime (as measured by gender, criminal victimization, socio-economic status, and place of residence) dominated exposure to the script as a determinant of concern for and fear of crime. In other words, fear of crime is more responsive to personal experience than to news media coverage.

Following the public opinion data crime-related attitudes were measured with two indices. Dispositional explanations of crime were measured with the following question:. "Now, here is a list of potential reasons that, according to some people, help explain why there is so much crime in this country". For each, tell us if you strongly agree, agree, neither agree nor disagree, disagree, or strongly disagree with the proposed remedy." Our index consisted of "failure of some groups in society to instill proper morals and values in their children", "breakdown of the family structure", and "people are just born criminals". Subjects who "strongly agreed" and "agreed" on each of the items were coded as one, all other responses were given the value of zero. We created an index by summing the responses and dividing by three (X=. 57, S.D.=. 22, Cronbach’s alpha = .57).

We measured support for punitive criminal justice policy using a composite three-item index. The items were worded as follows. "Now, here is a list of potential remedies that, according to some people, help explain why there is so much crime in this country". For each, tell us if you strongly agree, agree, neither agree nor disagree, disagree, or strongly disagree with the proposed remedy." Our index consisted "enforcement of the death penalty for people convicted of murder", "three strikes and you’re out legislation", and "putting more police on the streets." Subjects who "strongly agreed" and "agreed" on each of the items were coded as one, all other responses were given the value of zero. We created an index by summing the responses and dividing by three (X=. 53, S.D.=. 33, Cronbach’s alpha = .69).

Racial Attitudes: The public image of the "anti-social black" is nothing new. Classic models of prejudice are built around the assumption that one’s views on racial matters ultimately turn on how one feels about blacks. According to Webster’s, racial prejudice refers to a "natural aversion", "repugnance", or "instinctive opposition in feeling" toward a particular group and its members. Thus the acceptance of negative racial stereotypes is taken as evidence of racial prejudice (Allport, 1954). While we have chosen to call this old-fashioned" racism, it has also been referred to as "classical racism" (Sidanius et al, 1996), "red-neck racism" (McConahay, 1986), and "blatant racism" (Pettigrew and Meertens, 1995). Nomenclature aside, it generally includes traits associated with both genetic inferiority (e.g., lesser intelligence) and "environmental liberalism" (e.g., failings of character).

In our study, "old-fashioned" racism was measured by asking participants to rate African-Americans in terms of the applicability of the following traits -- "law abiding," "unintelligent," disciplined," and "lazy." The stereotype battery was worded as follows. "In this section we want you to rate various groups in terms of particular attributes that may or may not characterize them. Please consider the group named at the top of the list of attributes and then rate how well each attribute applies to that group in general. A score of 1 would mean that you think that the trait applies very well, while a score of four would mean that the attribute does not apply to the group at all. If you have no opinion about how well a particular attributes applies to the group, you may choose don’t know." Each rating was dichotomized (depending on the item we collapsed the "very well" and "quite well" or "not so well" and "not well at all" ratings) and then summed. Coefficient Alpha for this scale was .57. The index score, therefore, measures the probability that any given participant rated African-Americans negatively.

Challenges to white hegemony during Radical Reconstruction and then again in the Modern Civil Rights Movement (1942-1973), however, called into question the basic assumptions of traditional prejudice models (McAdam, 1982; Morris, 1985). By the 1960s opinion polls showed that white majorities rejected key tenants of racial prejudice in favor of racial equality. For example, support increased for the principles of school integration, fair housing, and to a lesser degree, intermarriage. And support decreased for the view that blacks were from a "less able race" (Schuman, Steeh, and Bobo, 1985; but see, Jackman, 1978). And in this spirit discrimination was deemed unlawful (e.g., Brown v. Board of Education, 1954; Civil Rights Act, 1964; Voting Rights Act, 1965, Fair Housing Act, 1968).

It is now common knowledge that people should not openly espouse racial animus. And for the most part, the society abides by this norm. Nonetheless, whites continue to resist many forms of racial change and pejorative stereotypes about African-Americans remain alive and well (Kinder and Sanders, 1996; Sniderman and Piazza, 1993). As a result scholars began to rethink classic models of prejudice.

The "new racism" is thought to be "symbolic", "subtle", "covert", "hidden", or "underground". Although the meaning and measurement of the new racism has varied widely from study to study and has been the basis of much controversy, there is general agreement that racial attitudes have become increasingly tied to support for traditional American values Kinder and Sanders offer what we believe is a sensible measure. They identify four central elements of the new racism: a) a denial that discrimination against African-Americans continues; b) a sense that blacks have violated traditional American values of hard work and self-reliance; c) a perception that blacks make illegitimate demands; and d) the belief that blacks receive undeserved benefits from government. These attitudes are captured in a battery of questions, originally designed for the 1986 National Election Study, and reproduced in our experiments.

We are now in a position to distill the preceding discussion into a few simple propositions. As operationalized in our experiments, the effects of the crime script can be represented by the following equation:

Hypothesis #1: b₀ < b₁ = b₂ = b₃

This model predicts that exposure to violent crime, regardless of the presence or absence of a perpetrator of a particular race, will heighten the tendency to attribute crime to individual failings, increase support for punitive crime policies, and increase negative attitudes toward African-Americans. The underlying logic of this prediction is that violent crime, in and of itself, stimulates support for punitive and racist attitudes.

Hypothesis #2: b₀ < b ₂ < b ₁ < b ₃

On this line of reasoning, race takes precedence over violence in determining viewer attitudes. Exposure to the black perpetrator in the news is expected to elicit higher levels of support for dispositional attributions, punitive remedies, and racist attitudes. Given the findings from our recall analysis, we also anticipate that the effect of the no perpetrator condition will be similar to the effect of the black perpetrator condition. In any event, the core prediction is that the black perpetrator version of the script will be significantly different from the null and white perpetrator conditions.
 
 

Hypothesis #3: b₀< b₁ < b ₂ = b ₃

Standing between hypotheses 1 and 2 is the view that the simple presence of any perpetrator is enough to influence crime and race attitudes. The race of the perpetrator, then, is of little consequence. Seeing a face personalizes crime and in this way leads viewers to harsher attitudes about crime. The one modification here is that exposure to either a white perpetrator or black perpetrator should lead people to cite dispositional attributions for crime and support a punitive crime policy agenda. In the case of racial attitudes, however, the expectation is that exposure to the black perpetrator should heighten anti-black sentiment; exposure to the white perpetrator, on the other hand, should weaken racial prejudice (i.e., the presence of a white perpetrator is evidence against the dominant stereotype).

Finally, we consider the question of differential effects of exposure to the crime script among whites and blacks. The case for race-specific effects is based on the vast racial differences in social and political beliefs and experiences (see, Kinder and Sanders, 1996). For example, African- Americans hold more nuanced views of blacks as a group (see, for example, Smith, 1996; Dyson, 1996); have greater interaction with one another (Oliver, 1988; Bienenstock, Bonacich, and Oliver, 1990), and have access to alternative media outlets which are less prone to rely on the crime script (Hunt, 1997). Thus blacks should reject the racial implications of the crime script because it is an attack on their in-group (Tajfel, 1978). On the other hand, it is possible that blacks, because of their vulnerability to crime, find the violent crime script just as compelling as whites. If so, the effects of the violence cue on crime-related attitudes should be uniform for both groups.

In summary, if the violent crime element of the script is dominant, we expect exposure to any crime story to influence crime and race attitudes equally for black and white subjects. If the racial element is more powerful, we expect that exposure to the black perpetrator version of the script will influence attitudes above and beyond exposure to the other conditions for white, but not black participants. Finally, if the simple presence of a perpetrator is the most notable element of the crime script, we expect exposure to any perpetrator in the news (regardless of race) to have a measurable impact on crime and race attitudes for both black and white subjects. The following analysis assesses each of these possibilities.

The Pooled Local News Experiment

We model the impact of the experimental manipulation by specifying separate dummy variable terms for each script element. So violent crime is expressed as a value of one for subjects who watched a violent crime story that did not depict a perpetrator and zero for all remaining conditions. The race of the suspect is measured by two dichotomous variables set equal to one when the suspect is either African-American or white and zero for all remaining conditions. For all equations, then, the null condition is expressed as the constant.

Although random assignment makes the need for statistical controls less necessary, we also controlled for standard political and demographic variables thought to be related to public opinion on crime and race (social class, education, gender, age, ideology, partisanship, criminal victimization, place of residence, and self-reported exposure to local television news). Lastly, given our interest in race-specific effects, we report separate results for white and black subjects.

The basic logic of our approach is to estimate the independent contributions of each script element to attitudes concerning crime and race after taking into account the effects of other relevant predispositions. We should remind readers that our manipulation is extremely subtle. The racial cue, for example, is operationalized as a five-second exposure to a mug shot in a ten-minute local news presentation. Consequently we have rather modest expectations about the impact of any given coefficient. Instead, we are looking for a pattern of results that is consistent with a particular hypothesis.

Table 3 documents the impact of the crime news script on crime-related attitudes. The coefficients for the treatment conditions represent the percentage of people endorsing the respective position (i.e., dispositional attributions and punitive solutions for crime) above and beyond the null condition (i.e., the constant term). The results in the first two columns of the table provide mild support for the "race" hypothesis. For example, among white subjects, exposure to the black perpetrator significantly raised support for the view that crime is caused by dispositional factors when compared to the null condition (+4%). Similarly, exposure to the no perpetrator condition significantly heightened support for dispositional attributions by 6% compared to the null condition. As predicted by Hypothesis 2 (race > violence), exposure to the no face and black face conditions both exerted significant effects on whites’ causal attributions, while exposure to the white perpetrator had no measurable impact. Also in keeping with the race hypothesis, the attributions of African-American participants were unaffected by exposure to any element of the crime news script.

The third and fourth columns in Table 3 provide even stronger support for the race

hypothesis. In the first instance, the pattern of coefficients among white subjects is directly in line with a priori expectations. Exposure to the black perpetrator had the greatest impact, on support for punitive policies (+6 percent) followed by the no perpetrator and white perpetrator conditions respectively. Further, the effects of exposure to the black perpetrator condition proved significantly more powerful than exposure to the white perpetrator condition (t = 2.2; p < .01). Once again, the results from our black sub-sample provided support for the race hypothesis: as shown in the last column, exposure to any element of the crime script actually served to reduce their support for punitive crime policies.

Perhaps the most interesting finding from the control variables was that whites’ who said they frequently watched local television news were significantly more likely to cite dispositional attributions and support punitive crime remedies. On the other hand, black frequent viewers were less likely to hold these views. The other finding of note is the inconsistent effects of education and ideology among our white participants. Somewhat unexpectedly, liberals and the more educated were more likely to cite dispositional attributions. But when it came to punitive solutions, and more in keeping with the literature, these groups held more liberal views. There is no obvious explanation for these results.

We next turn to the implications of the crime script for racial attitudes. Table 4 displays the results. The first column indicates that, generally speaking, the crime script had no effect on old-fashioned racism among white subjects. However, the pattern of differences was most consistent with the race hypothesis. Thus the coefficient for the black perpetrator condition had the correct sign and was larger than the coefficients for both the white perpetrator and no perpetrator conditions (although these differences were statistically insignificant).

The race hypothesis fared better when we examined the views of African Americans. Exposure to any element of the crime script reduced negative stereotyping and the effect was significant for both the white and black perpetrator conditions. In other words, the violence cue, in and of itself, had virtually no impact on blacks’ stereotypes of themselves. When the race of the perpetrator was known, however, black subjects whole-heartedly rejected the racial implications of the crime script. Seeing either a white or a black perpetrator in the news reduced the percentage of blacks with negative stereotypes by about 10%.

The last column of Table 4 shows the influence of the crime script on the "new

racism." While exposure to any element of the crime script served to strengthen new racism, exposure to the black perpetrator condition had the greatest impact. Seeing a black perpetrator in the news increased whites’ new racism scores by 17% compared to the null condition. Using the constant as the baseline, a rough calculation suggests that seeing a black perpetrator in the news raised the percentage of whites who support the new racism from one-third to over one-half. Further, the black perpetrator coefficient proved significantly larger than the white perpetrator condition (t = 2.8; p < .001). These effects suggest that while more subtle racial attitudes -- such as new racism -- can be strengthened by mere exposure to violent crime, it is the presence of a black perpetrator that has the most dramatic effect on public attitudes.

Finally, the effects of the control variables were intuitive. Among whites, women and liberals were least likely to support racist views. Although the difference was not statistically significant, frequent viewers of television news were more likely to hold negative views about African-Americans. For blacks, self-reported exposure to the news had the opposite effect: frequent viewers were less likely to endorse negative views about blacks.

In sum, the local news experiment demonstrates that exposure to the crime script significantly influences attitudes about both crime and race. Our results show that it is the racial element of the crime script, however, that is the dominant cue. This conclusion is supported by two pieces of evidence. First, the coefficients for the black perpetrator conditions are always larger than the coefficients for both the null and the white perpetrator conditions. The black perpetrator coefficients are significantly different from the null condition in three of the four comparisons, and from the white perpetrator condition in two of the four comparisons (the latter comparison is more stringent, of course, because it controls for the presence of the violence cue). Second, the African-American sub-sample generally rejects the crime script. Six of the nine relevant coefficients carry negative signs. In particular, exposure to the crime script leads blacks to lowers their support for punitive criminal justice policies and reduces their willingness to accept negative characterizations of their group. This pattern is in stark opposition to the findings for our white study participants.

While the experimental results are revealing, we are sensitive to the argument that controlled experiments have certain limitations. The most common concern has to do with external validity. Experimental samples, for instance, are typically not drawn with the rigor of probability samples common to most public opinion surveys. To take account of this possibility the following section introduces a survey replication that generally corroborates the findings reported above.

Each year, the Institute for Social Science Research at the University of California, Los Angeles surveys the social and political attitudes of Los Angeles County. From our perspective, this survey is advantageous because the sample frame includes the areas from which we recruited a significant portion of our experimental subjects. Moreover, the survey was administered at approximately the same time as our news experiments. Thus we have the ability to match the experimental and survey findings.

The survey questions tapped a wide range of political topics as well as standard background indicators. We are in a position to reconstruct the experimental results because the survey included questions concerning respondents’ exposure to broadcast media. Specifically, people were asked about their local television news viewing habits ("How often do you watch local news like ‘Eyewitness News’ or ‘Action News’? Every day, three or four times a week, once or twice a week, or hardly ever?" We assign a score of 1 to those viewers who report watching "hardly ever", a score of 2 to viewers who watch "once or twice a week", a score of 3 for people watching "three or four times a week", and a score of 4 for daily viewers of local television news. Our assumption is that frequent television news viewers are more likely to be exposed to the crime news script. In turn, we assume that this group of viewers is also more likely to be exposed to the racial element of the script. Thus, for the findings reported above to be sustained, we believe, the local news exposure measure should have a significant effect on attitudes concerning both crime and race. In other words, while we might expect higher levels of exposure to impact crime attitudes as a simple function of the violence element of the script, it would take some exposure to the racial element of the script to move racial attitudes.

Respondents were also asked their views about crime and race. As in the experiment, we included survey indicators of causal attributions, support for punitive crime policies, racial stereotypes, and new racism. The results are presented in Table 5.

There were several parallels between the experimental and survey data. In all four cases, the sign for the local news exposure measure was in the anticipated direction. Further, in three of the four comparisons, exposure to local television news had a statistically significant impact on viewer attitudes. For instance, participants who report watching every day were 16% more likely than those who hardly ever watch local news to support punitive remedies and endorse the view that blacks are less intellectually able. More dramatically, daily viewers were 28% more likely to subscribe to the new racism than viewers who rarely watch the news. In sum, frequent news watchers were more likely to support punitive crime policies and to endorse negative characterizations of African-Americans. Thus, the survey and experimental results match quite closely.

In conclusion, our experimental and survey evidence both suggest that exposure to local news coverage of crime conditions attitudes toward crime and race. In particular, the racial element of the crime script (as opposed to the violence element) has the most demonstrable impact. Our experiments show that for white viewers, a brief five-second exposure to a black perpetrator in the news is sufficient to increase the percentage of people who believe crime is caused by individual failings and who support punitive crime policies. In addition exposure to the black perpetrator strengthens racial stereotypes and fosters the view that African-Americans are out of step with the cultural mainstream. On the other hand, the crime script has generally the opposite effect on African-American viewers. This pattern is supported by our survey replication. Using a rough exposure measure as a proxy for the crime script, we find that white frequent viewers are more likely to endorse punitive crime policies and express negative beliefs about African-Americans. In short, what appears in the news on a regular basis -- violent crime and racial imagery -- does not go unnoticed.

The scope of our experimental effects is impressive given the scale of the manipulation (a one-minute exposure in the case of the violent crime script element and a five-second exposure in the case of the racial element). Our confidence in the generalizability of the observed effects is boosted not only by the range of attitudes surveyed, the multiplicity of experimental locales, and our reliance on adult samples instead of college sophomores, but also by the convergence of the experimental and survey results. Across both methods and a variety of measures, the crime script influenced viewers’ attitudes.

Although our results indicate that the crime script is a meaningful cue, the scale of effects is quite modest. After all, our manipulation is minute when compared with a lifetime’s worth of socialization. Moreover, some of the weaker results can be explained, in part, by the weakness of our measures. For example, the scales for dispositional attributions and old-fashioned stereotypes are not highly reliable (at least as measured by Cronbach’s alpha). On the other hand, the punitive crime policy and new racism scales are considerably more reliable and this is reflected in both the strength of the coefficients and the increase in explained variance. Finally, despite our use of multiple methods, questions remain about the generalizability of our results. Los Angeles is fairly typical in its crime coverage compared to other media markets (see Klite, et al, 1997), but one wonders if similar effects would be obtained elsewhere.

The commercial realities of our time dictate that local news will continue to cultivate misperceptions and prejudice. Local television stations reach huge audiences, but face intense economic pressures. Crime dominates other news because its emphasis on vivid pictures and emotional personal accounts is believed to attract viewers. As we noted earlier, the news rarely presents non-racial attributes of criminal suspects – educational attainment, age, employment status, family background, etc. Information about race is conveyed automatically, due to the visual nature of the medium; other individuating characteristics are seemingly not newsworthy. While reporters cannot be expected to compile detailed bibliographies of suspects (who are frequently not apprehended when the story airs), they can. consider other ways of reporting on crime (see, Guensburg, 1999).

Despite the limitations of our evidence, we believe that journalists need to rethink their reliance on the crime script. The civic and commercial objectives of news organizations are not necessarily zero-sum in nature. Stations could de-emphasize reporting on particular episodes of violent crime while providing more substantive, thematic coverage of local communities. Stations like KUVE (in Austin, Texas) and KTVU (in San Francisco) have developed important initiatives in which crime coverage is allocated in accordance with a set of guidelines based on "community impact." Initial results indicate that the omission of graphic accounts of violence has not diminished ratings. (see Holley, 1996). Further, the goal of achieving more balanced reporting about crime might be facilitated by hiring reporters with better knowledge of the communities they cover, and by increasing the ethnic diversity of the newsrooms -- especially in the important "gatekeeping" positions.

In closing, we note that the effects of the crime script extend well beyond the views of ordinary citizens. Our evidence shows that local news programming "racializes" political discourse by making policy opinions increasingly intertwined with questions of race. The audience’s heightened sensitivity to matters of race, as history amply attests, is grist for vote-seeking politicians. Racial appeals – explicit or coded – are now common in political campaigns. In short, the emergence of local news has made race an even more central component of American life.
 
 

Abelson, R.P. 1976. "Script Processing in Attitude Formation and Decision Making," p.33-46 in J.S. Carroll and J.W. Payne eds. Cognition and Social Behavior. Mahwah, NJ: Lawrence Erlbaum Associates.

Abelson, R.P. 1981. "The psychological status of the script concept." American Psychologist, 36, 715-29.

Allport, G.W. 1954. The Nature of Prejudice. Garden City, NJ: Doubleday.

Ball-Rokeach, S., M. Rokeach, and J. Grube. 1984. The Great American Values Test: Influencing Behavior and Belief Through Television. New York: The Free Press.

Belden, M. and J. Russonello, 1997. Reporting of existing public opinion data on juvenile justice issues. Belden and Russonello: Washington, D.C.

Black, J.B., J.A. Galambos, and S.J. Read. 1984. "Comprehending Stories and Social Situations," pp.45-86 in R. S. Wyer Jr., and T. K. Srull eds., Handbook of Social Cognition. Vol. 3. Mahwah, NJ: Lawrence Erlbaum Associates.

Bienenstock, E.J., P. Bonacich, and M. Oliver. 1990. "The Effect of Network Density and Homogeneity on Attitude Polarization." Social Networks 12: 153-172.

California Department of Justice. 1997. Crime 1997 in selected California jurisdictions. Sacramento: Division of criminal justice information services.

Campbell, C. 1995. Race, Myth and the News. Thousand Oaks, C.A.: Russell Sage.

Campbell, A., P.E. Converse, W.E. Miller, and D.E. Stokes. 1960. The American Voter. New York: John Wiley and Sons.

Crispin-Miller M., 1998. Crime News in Baltimore: The economic cost of local TV’s bodybag journalism. New York: Project on Media Ownership.

Dovidio, J.F. and S.L. 1986. Gaertner. Prejudice, Discrimination, Racism: Theory and Research. New York: Academic Press.

Dyson, M.E. 1996. Race Rules: Navigating the color line. Reading, Mass.: Addison-Wesley.

Elias, R. 1994. "Official Stories: Media coverage of American crime policy," Humanist. 54: 3-8.

Entman R.M., B.H. Langford, D. Burns-Melican, I. Munoz, S. Boayue, C. Groce, A. Raman, B. Kenner, and C. Merrit, 1998. Mass Media and Reconciliation. Cambridge: John F. Kennedy School of Government.

Entman, R.M. 1992. "Blacks in the News: Television, modern racism, and cultural change," Journalism Quarterly. 69: 341-62.

Entman, R.M. 1990. "Modern Racism and the Images of Blacks in Local Television News," Critical Studies in Mass Communication. 7: 332-46.

Gilliam, F D., Jr. 1998. "Race and Crime in California" in M. B. Preston, B. A. Cain, and S. Bass (eds.), Racial and Ethnic Politics in California. (Berkeley: Institute for Governmental Studies Press, University of California).

Gilliam, F.D., Jr., and S. Iyengar. 1997. "Prime Suspects: The effects of local news on the viewing public." Paper presented at the Annual Meeting of the Western Political Science Association, Portland.

Gilliam, F.D. Jr., S. Iyengar, A. Simon and O. Wright. 1996. "Crime in Black and White: The violent, scary world of local news," Harvard International Journal of Press/Politics. 1: 6-23.

Graber, D. 1976. Verbal Behavior and Politics. Urbana: University of Illinois Press.

Graesser, A.C., S.B. Woll, D.J. Kowalski, and D.A. Smith. 1980. "Memory for Typical and Atypical Actions in Scripted Activities," Journal of Verbal Learning and Verbal Behavior, 18, 319-32.

Graesser, A.C., M. Singer, and T. Trabasso. 1994. "Constructing Inferences During Narrative Text Comprehension," Psychological Review 101: (3) 371-395.

Guensburg, C. 1999. "Taming the Beast." American Journalism Review. July.

Hess, S. 1991. Live from Capital Hill: Studies of congress and the media. Washington, D.C.: Brookings Institute.

Holley, J. 1996. "Should the Coverage Fit the Crime?," Columbia Journalism Review, May/June, pp. 27-32.

Hunt, D. 1997. Screening the Los Angeles "Riots": Race, seeing, and resistance. New York: Cambridge University Press.

Hurwitz, J. and M. Peffley. 1997. "Public Perceptions of Crime and Race: The role of racial stereotypes," American Journal of Political Science. 41: 375-401.

Iyengar, S. 1991. Is Anyone Responsible?: How television frames political issues. Chicago: University of Chicago Press.

Jackman, M. and Muha, M. 1984. "Education and Intergroup Attitudes: Moral enlightenment, superficial democratic commitment, or ideological refinement?" American Sociological Review. 49: 751-769.

Kinder, D. R. and L. M. Sanders. 1996. Divided by Color: Racial politics and democratic ideals. Chicago: University of Chicago Press.

Klite, P., R.A. Bardwell, and J. Salzman. 1997. "Local TV News: Getting away with murder," Harvard International Journal of Press/Politics. 2:102-112.

Lee, H. C. and C. Ladd, 1997. "Criminal Justice: An unraveling of trust?" The Public Perspective, 8 (4): 6-7.

Mandler, J.M. 1984. Stories, Scripts, and Scenes: Aspects of schema theory. Mahwah, NJ: Lawrence Erlbaum Associates.

Massey, D.S. and N.A. Denton. 1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge: Harvard University Press.

McAdam, D. 1982. Political Process and the Development of Black Insurgency, 1930-1970. Chicago: University of Chicago Press.

McConahay, J. B. 1986. "Modern Racism, Ambivalence, and the Modern Racism Scale." In J. F. Dovidio and S.L. Gaertner (eds.). Prejudice, Discrimination, Racism: Theory and Research. New York: Academic Press.

Morris, A. M. 1984. The Origins of the Civil Rights Movement: Black Communities Organizing for Change. New York: Free Press.

Neuman, W., M. Just, and A. Crigler. 1992. Common Knowledge: News and the Construction of Political Meaning. Chicago: University of Chicago Press.

Oliver, M. L. 1988. "The Urban Black Community as a Network: Toward a social network perspective." Sociological Quarterly. 29(4): 623-645.

Papper, R. and M. Gerhard. 1997. "Newsrooms Still Earn Profits," Communicator. 7-8.

Peffley, M., T. Shields, and B. Williams. 1996. "The Intersection of Race and Crime in Television News Stories: An experimental study," Political Communication. 13: 309-328.

Pettigrew, T.F. and R.W. Meertens. 1995. "Subtle and Blatant Prejudice in Western Europe." European Journal of Social Psychology. 25: 57-75.

Roediger, H. L. III, and K. B. McDermott, 1995. "Creating False Memory: Remembering words not presented in lists," Journal of Experimental Psychology, 21: 803-14.

Romer, D., K. H. Jamieson, and N. J. de Coteau. 1998. "The Treatment of Persons of Color in Local Television News - Ethnic blame discourse or realistic group conflict?" Communication Research 25 (3): 286-305.

Roper Starch Worldwide. 1994. Roper Reports, 93: 22-23.

Schank, R.C. 1990. Tell Me a Story: A new look at real and artificial memory. New York: Scribners.

Schank, R.C., and R. P. Abelson. 1977. Scripts Plans, Goals, and Understanding. Mahwah, NJ: Lawrence Erlbaum Associates.

Scheingold, S.A. 1995. "Politics, Public Policy, and Street Crime." Annals of the American Academy of Political and Social Science. 539: 155-68.

Schuman, H., C. Steeh, and L. Bobo. 1985. Racial Trends in America: Trends and Interpretations. Cambridge, MA: Harvard University Press.

Sears, D.O. 1988. "Symbolic Racism," in P.A. Katz and D. A. Taylor Eds. Eliminating Racism: Profiles in controversy. New York: Plenum. pp. 52-113

Sidanius, J., F. Pratto, and L. D. Bobo. 1996. "Racism, Conservatism, Affirmative Action, and Intellectual Sophistication: A Matter of Principled Conservatism or Group Dominance?" Journal of Personality and Social Psychology. 70: 1-15.

Smith, R.C. 1996. We Have No Leaders: African Americans in the post-civil rights era. Albany: State University of New York Press.

Sniderman, P. and T. Piazza. 1993. The Scar of Race. Cambridge, MA: Harvard University Press.

Sulin, R.A., and Dooling, D.J. 1974. Intrusion of a Thematic Idea in Retention of Prose. Journal of Experimental Psychology, 103, 255-62

Tajfel, H. 1978. Differentiation Between Social Groups: Studies in Intergroup Relations. London: Academic Press.

Wortley, S., J. Hagan, and R. Macmillan. 1997. "Just Des(s)erts? The racial polarization of perceptions of criminal injustice," Law and Society Review 31: (4) 637-676.