Communication Research Reports Vol. 27, No. 3, July–September 2010, pp. 207–216 Missing Children in National News Coverage: Racial and Gender Representations of Missing Children Cases Seong-Jae Min & John C. Feaster Downloaded By: [Feaster, John] At: 21:58 27 July 2010 This research explores race- and gender-related selection bias in national television news coverage of missing children cases. When the proportions of race and gender from the news coverage of five national television stations between 2005 and 2007 were compared to official missing children statistics, it was found that African American missing children and female missing children were significantly underrepresented in television news cover- age. It is argued that such things as newsroom diversity, news operation routines, media ownership, and commercial motives of media contribute to the race- and gender-related media bias. Keywords: Children; Media Bias; Minority; Race and Media; Television News; Women in Media In June 2002, a 14-year-old girl named Elizabeth Smart went missing from her home in Salt Lake City, Utah. She was kidnapped by strangers and found alive in March, 2003 in a nearby small city. For the nine months she went missing, her story received frenetic attention from the local media and eventually received national, and even international, coverage. During the time period when Smart was missing, Seong-Jae Min (PhD, The Ohio State University, 2009) is an assistant professor in the Department of Communi- cation Studies at Pace University. John C. Feaster (PhD, The Ohio State University, 2008) is an assistant professor in the Department of Communication Studies at Rowan University. A previous version of this article was presented at the 2008 International Communication Association conference May 22–26, Montreal, Canada. Correspondence: Seong-Jae Min, Department of Communication Studies, Pace University, 41 Park Row, New York, NY 10038; E-mail: smin@pace.edu ISSN 0882-4096 (print)/ISSN 1746-4099 (online) # 2010 Eastern Communication Association DOI: 10.1080/08824091003776289  208 S.-J. Min & J. C. Feaster numerous other children were missing as well. Although some cases involved the deaths of the children, the coverage of these stories was minimal in comparison. Why do we see such a disparity in media’s coverage of missing children cases? Washington Post staff writer, Howard Kurtz (as cited in Cable News Network, 2002), argued that such factors as the missing children’s race and social status matter. He claimed that Smart’s being a young, attractive, Caucasian girl abducted from a rich neighborhood may partially explain her disproportionate level of coverage. Kurtz noted the case of Alexis Patterson, a seven-year-old, African American, from a poor Milwau- kee neighborhood, as support for this claim. She was missing at approximately the same time as Smart, yet received little local media coverage, let alone national attention. The U.S. Department of Justice estimates that there were approximately 800,000 missing children in the nation reported to police or missing children’s agencies in 2002 alone (Sedlak, Finkelhor, Hammer, & Schultz, 2002). Demographic profiles of those missing children show that 51% of them were boys, and 47% were racial minor- ities. Although around one half of the missing children were boys and racial minorities, Downloaded By: [Feaster, John] At: 21:58 27 July 2010 anecdotally, it appears that mainstream television news mostly covers cases of White missing children, especially young Caucasian girls. If this is true, it creates a concern of social justice. Locating missing children is an important task, and news journalists play a prominent role in the endeavor. However, if a particular segment of missing chil- dren gets more news coverage than others—and thereby has a greater chance of being found—there is a potential flaw in the journalistic process. In this project, we employ a content analysis of national television news reports of missing children to determine if any patterns of bias with regard to race or gender representation are present. Children in Television News Although children are generally underrepresented in news media (Kunkel & Smith, 1999), child victimization stories appear to be the notable exception. Pritchard and Hughes (1997) argued that these stories are typically appraised as high in news- worthiness because crimes against children (who often represent purity and inno- cence) are rare, morally abhorred, and constitute high deviance. According to Chermak (1995) news stories with child victims typically have human interest ele- ments and are likely to elicit viewers’ emotions. Not all child victimization cases, however, are looked upon as equal in newsworthiness. Race and gender are potential mediators for the assessment of newsworthiness for child victimization cases. Previous studies dealing with adult cases show that mainstream news in America typically overrepresents White victims, unfavorably depicts racial minorities, and objectifies and trivializes women (e.g., Byerly & Ross, 2006; Dixon & Linz, 2000; Entman, 1994; Keever, Martindale, & Weston, 1997). As with news media’s portrayal of adults, there may be a widespread news bias in the coverage of missing children cases. Research Questions The main task of this study was to determine if there are any patterns of over- or underrepresentation with regard to race or gender in the coverage of missing children  Communication Research Reports 209 cases in national television news. Given the research discussed earlier, there is some base of reason to form expectations as to what will be found. However, given that so little research has been done on coverage of children in the news, we opted to present research questions as opposed to specific hypotheses. Racial Bias News bias favoring White individuals is well-documented. For example, Dixon and Linz (2000) found that African Americans and Hispanics are overrepresented as criminals, whereas White individuals are overrepresented as victims in television news coverage. Any racial bias that may exist in news coverage is not necessarily the result of overt racism. According to Tuchman (1978), news is not the result of journalists’ idiosyncrasies, but the product of specific ways of organizing newswork enmeshed within the culture of an organization. Several factors may contribute to an organizational culture that could potentially foster a ‘‘White bias.’’ First, although Downloaded By: [Feaster, John] At: 21:58 27 July 2010 newsroom diversity has increased over the last years, media owners and journalists are still predominantly White. Van Dijk (1993) argued that journalists who have been socialized and educated in such a context will be more likely to take a White perspec- tive on news events than they would a minority perspective. Second, the commercial drive to present dramatic stories that will draw the audience may influence newsroom culture and behavior. Profit motives lie at the heart of commercial news firms, and their profits are dependent on audience size and advertising revenues. This may lead news organizations to focus on dramatic stories featuring victims of the majority audience group (Fuhrman, 2009). In particular, stories regarding individuals similar to one’s self are more cognitively appealing to a viewer than are stories about dissimi- lar others (Harwood, 1997); thus, news stories about individuals who are similar to the majority are likely to draw a large audience. With a predominantly White audience, this tendency may lead to featuring a greater number of White victims (Dixon & Linz, 2000). Third, journalists typically have ‘‘beats’’ that take them to the same sources, and this may also become a potential source of coverage bias because it is possible that some sources—in this case, more established White groups—are more connected to the journalists and provide news featuring victims of their own interest. Given these concerns, we ask the following research question: RQ1: Are African American missing children cases underrepresented in national television news when compared to their actual rate of incidence?1 Gender Bias Overall, women have been found to be significantly underrepresented in news cover- age (Lens-Rios, Rodgers, Thorson, & Yoon, 2005). When women do appear in news, they are frequently framed as victims who are passive and dependent (Benedict, 1992; Meyers, 1994). Feminist scholars argue that media at large are fascinated with the ‘‘fragile female form and her vulnerability to violation’’ (Byerly & Ross, 2006,  210 S.-J. Min & J. C. Feaster p. 42); and, as a result, female victims are judged as highly newsworthy. When the element of feminine victimization is combined with the fact that the victim is a child, it may generate judgments of greater newsworthiness because children represent even more vulnerability than adults. Furthermore, some argue that the mass media system has a tendency to eroticize and admire girlhood. In what he called media’s ‘‘juvena- tion,’’ Hartley (1998) argued that young girls became the objects of public gaze with their youthfulness and sexuality. This view suggests that the stories featuring female children, particularly when framed as victims, may be judged as more newsworthy than other similar cases. Given these arguments, we pose the following research question: RQ2: Are female missing children cases overrepresented in national television news when compared to their actual rate of incidence? Types of Missing Children Cases Downloaded By: [Feaster, John] At: 21:58 27 July 2010 Beyond coverage of missing children as an omnibus category, the type of case may have some relationship with the race and gender of missing children covered. Threat- ening material has been shown to lead to selective attention (Mathews & Mackintosh, 1998; Yiend & Mathews, 2001). As such, more violent missing children cases, such as those resulting in death, may be deemed as more newsworthy than less violent ones. Therefore, we ask the following research question: RQ3: Is there any relationship between missing child situation types and the race and gender of the children covered in national television news? Method To address the research questions, we content analyzed missing children stories in national television news, and compared race and gender proportions to official miss- ing children statistics provided by the U.S. government. Television was considered to be an optimal medium for investigating this subject because of the medium’s capacity to provide up-to-the-moment updates compared to other channels, such as newspa- pers. With little direction from previous research, we elected to focus on missing children cases at the national level because of the large number of cases journalists at this level have to chose from at any given time. By comparison, any specific local region may have few or no cases to cover at a particular point in time. Although bias likely exists at both the national and local levels, we deduced that national news would reveal more about potential bias (and is of greater interest) than local news. The content analysis of television news has been employed in cultivation research and many studies concerning media coverage of social minorities (Dixon & Linz, 2000; Romer, Jamieson, & de Coteau, 1998). For this study, news clips of missing children cases were obtained through the Vanderbilt Television News Archive. A dataset documenting real-world statistics of missing children was obtained from the National Crime Information Center of the Federal Bureau of Investigations  Communication Research Reports 211 (FBI). The Vanderbilt archive contains all evening newscasts from major national broadcasters for the past several decades. From this archive, we searched for evening newscasts from ABC, CBS, NBC, CNN, and FOX News that contained any missing children cases between 2005 and 2007. The time frame of 2005 and 2007 was selected because the FBI dataset that was used as a base of comparison had only been compiled since 2005. The search terms used to build the sample of clips within the archive were, ‘‘missing child(ren),’’ ‘‘missing boy(s),’’ ‘‘missing girl(s),’’ and ‘‘missing kid(s).’’ The use of search terms regarding specific incidents, such as ‘‘kidnapping,’’ were considered but not used due to concerns about potential, as yet unknown, biases in the coverage. After eliminating duplicates, our sample contained 161 video clips. The unit of analysis for this study was reference to a specific missing child. Stories ranged from having one case to dozens of cases. Only cases involving missing individuals described as being under 18 were included. In addition, only instances where individuals were identified by name and photo were included. For example, several stories featured dozens of pictures of missing children, but within those Downloaded By: [Feaster, John] At: 21:58 27 July 2010 stories, only the pictures that were assigned names were included in the analysis. Also, we did not include missing children cases due to Hurricane Katrina in 2005.2 After applying these criteria, our final sample yielded a total of 185 cases. We each coded every case in the sample by reading the descriptions of the stories provided in the Vanderbilt database and by visual inspection of the video clips containing the stories. Our codes were compared to check for consistency. When disagreements arose, we revisited the stories and discussed them until consensus was reached. Excluding 11 cases that were uncodable on one of the variables due to a lack of pictorial referencing in the story, we achieved complete agreement on each variable (Cohen’s j ¼ 1.0). Within the stories analyzed, each reference to a child was coded based on the network on which the story was aired, gender, race, and situation of the case. In nearly every case, gender and race was coded using visual inspection of a picture featured in the story, and the situation of the case was coded using the spoken content of the story. Within the sample, there were 115 male cases, 70 female cases, and three cases where gender was indeterminable. Race was coded as African Amer- ican and non-African American. Although it would have been preferable to use Cau- casians as a reference group given the claims made in this study, Caucasians (the U.S. racial majority) and Hispanic Americans (a racial minority) were included in a single ‘‘White’’ category in the FBI dataset. Thus, it was necessary to use African American as a reference group. Within the sample, there were 36 African American cases and 149 non-African American cases. For eight cases, race could not be identified due to lack of necessary description within the stories. The situation involved in a case was coded as it was described within the content of the story. The categories used for situation coding were ‘‘death’’ (n ¼ 71), ‘‘kidnapping’’ (n ¼ 79), and ‘‘other= general missing’’ (n ¼ 35). Due to the overriding violent nature of stories involving death, cases that involved both kidnapping and death were coded as ‘‘death.’’ The analyses for RQ1 and RQ2 were done through difference of proportion tests using the Z distribution. The test, also known as the Z test for independent  212 S.-J. Min & J. C. Feaster proportions, is often used to see whether a particular category is deviating from its expected value more than just sampling error would tend to produce by chance. The expected proportions used for calculation were those of the known population provided in the missing children FBI dataset. Z score was defined as follows: O
E Z ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; ½Oð1
OÞ=n1 þ ½Eð1
EÞ=n2 where O was the observed proportion of the incidents (e.g., for RQ1, it was the proportion of African American missing children in news reports), and E was the expected proportion of the incidents (for RQ1, it was the proportion of African American missing children in the official FBI statistics). The total samples or populations from which the observed and expected came from were represented by n1 and n2, respectively. The analysis conducted for RQ3 was done so using cross-tabulations and the Downloaded By: [Feaster, John] At: 21:58 27 July 2010 Pearson chi-square test. RQ3 was answered considering only the news coverage stat- istics, not comparing them to the real-world statistics. This was because the FBI data- set did not provide situational details. Moreover, the dataset provided only the numbers of cases that involved certain conditions, such as the number of cases involv- ing a child of a certain ethnicity or the number of cases involving female children. As such, it was not possible to calculate proportions of cases with overlapping conditions, such as the percentage of cases involving children who were both girls and Black. Results RQ1 asked whether African American missing children cases are underrepresented in national television news compared to their actual rates of incidence. When the proportion of African American missing children cases covered in the news (19.5%) was compared to the actual proportion of reported incidents in the FBI data- set (33.2%), African American missing children cases were significantly underrepre- sented in television news (Z ¼
4.71, p < .001). RQ2 asked whether female missing children cases are overrepresented in television news when compared to their actual rates of incidence. When comparing the proportion of female missing children cases covered in the news (37.8%) to the actual proportion of reported incidents (57.4%), female missing children were significantly underrepresented in the coverage (Z ¼
5.50, p < .001). Information regarding the comparisons made for RQ1 and RQ2 can be found in Table 1. RQ3 asked if there is any relationship between the coverage of race and gender groups and the types of missing children cases covered in news reports. When the four demographic categories (African American boy, African American girl, non- African American boy, and non-African American girl) from the news sample were cross-tabulated with the three missing children categories (death, kidnapping, and general missing), a significant relationship was found: v2(9, N ¼ 185) ¼ 48.45, p < .001. In examining the standardized residuals, the coverage of kidnapping cases  Communication Research Reports 213 Table 1 Comparisons of Proportions of Missing Children Cases in News Reports and Official Reports, 2005–2007 A: Television B: Official Variable News Sample FBI Dataset Difference (A
B) Race African American 19.5% (n ¼ 36) 33.2% (N ¼ 212,770)a –13.7% Non-African American 80.5% (n ¼ 149) 66.8% (N ¼ 428,463)a þ13.7% Gender Male 62.2% (n ¼ 115) 42.6% (N ¼ 279,635)a þ19.6% Female 37.8% (n ¼ 70) 57.4% (N ¼ 376,374)a –19.6% a The expected values were the average of the values from the three analyzed years. of non-African American girls was significantly greater than expected (Z ¼ 2.10, Downloaded By: [Feaster, John] At: 21:58 27 July 2010 p < .05), and the coverage of death cases for African American boys was significantly greater than expected (Z ¼ 3.80, p < .01). Discussion The results of this study suggest that the coverage of missing children cases may be disproportional based on race and gender. In accordance with previous research (e.g., Keever et al., 1997), which has found disproportional news coverage for minorities, African American missing children cases were significantly underrepresented when compared to national statistics. This suggests that judgments of newsworthiness for missing children cases in this domain may indeed be affected by a predominantly White cultural perspective in news organizations (Tuchman, 1978; van Dijk, 1993) or the commercial drive to attract the attention of a predominantly White audience (Abrams & Giles, 2007; Fuhrman, 2009). However, contrary to past research arguing that women are overrepresented as victims (Benedict, 1992; Meyers, 1994), the results of our study suggest that girls were proportionally underrepresented relative to actual rates of incidence. When analyzed by situation, however, coverage of non-African American female kidnapping cases were greater than expected. Although the findings of this study provide some valuable insight regarding the coverage of missing children based on race and gender, there are a number of limita- tions to address and directions for future research to explore. First, this study aimed to merely assess if any disproportions in the coverage of race and gender in missing children cases were present; but, aside from type of case, moderating factors associa- ted with coverage were not considered. For example, does the attractiveness of the child have an impact on coverage? Does the social class or affluence of the child’s family lead to a greater likelihood in a case receiving coverage? Missing children who are considered ‘‘cute’’ or come from affluent families may have higher chances of receiving coverage than those who are unattractive or less affluent. Another factor that may affect the coverage of missing children is the availability of the pictures.  214 S.-J. Min & J. C. Feaster Television requires certain images or videos, and the ease of obtaining visual materials may determine which story is aired. Although these factors could not be explored in the current study due to the setup of the FBI dataset, they are potentially important avenues for future investigations. Second, it was our original intention to analyze the specific proportion of each race–gender group (White boy, White girl, African American boy, African American girl, Hispanic boy, Hispanic girl, etc.); however, as previously discussed, the structure of the FBI dataset made this impossible. If more detailed, real-world statistics become available in the future, such an analysis would yield important insight regarding the coverage of missing children. Third, beyond determining if there are differences in the number of stories covering missing minority children, future studies may investigate if there are differ- ences in how they are covered. Are people of different relationships interviewed? What attention is given to potential suspects? How much time is devoted to each story? Is there a difference in story framing? A better understanding of such dynamics Downloaded By: [Feaster, John] At: 21:58 27 July 2010 would yield valuable insight not only for the coverage of missing children, but for all forms of news coverage. Conclusion News coverage is a vital component of society that demands attention due to its potential to affect the worldviews of audience members. The findings of this study suggest that television news has some biases concerning the coverage of missing chil- dren cases with regard to race and gender. In particular, it was found that, although a relatively large number of African American children are actually missing, they are significantly underrepresented in television news. Of course, it cannot be argued that bias toward any group is entirely, or even mostly, attributable to how that group is covered in the news. However, media coverage may be symptomatic of higher-order issues. The evidence found in this study, and in those similar to it, should be taken as a reminder of the necessity for societal level self-evaluation. Greater attention must be placed on the practices regarding judgments of news- worthiness. As institutions that exist to generate profit, the culture of news organizations may have tendencies to report stories that will primarily serve those interests. However, because coverage of missing children is important to their being found, news organizations have an ethical obligation to strive for at least pro- portional coverage, regardless of profit. Missing children from minority groups are no less important than are those from the majority and, as such, have just as much right to have a face as opposed to being a mere statistic. Notes [1] Due to the setup of the dataset we used, we posit underrepresentation of African American missing children rather than overrepresentation of White missing children. This issue is further discussed in the Method and Discussion sections.  Communication Research Reports 215 [2] Hurricane Katrina, which hit the southern part of the United States in 2005, predominantly affected the African American population, making it possible that more African American children went missing in 2005. However, it was learned that the Federal Bureau of Investiga- tions dataset contained only dozens of Katrina-related cases out of more than 660,000 children reported missing in 2005. Given this small number, we decided to remove Katrina-related cases from our sample. Besides, the news clips in our sample that featured missing children due to Katrina were unusual in that (a) they were all from CNN, and (b) they were simultaneous messages that accompanied regular news programs—that is, the missing children messages were displayed as a ticker on one side of the screen with a very short amount of time allotted to each case. References Abrams, J. R., & Giles, H. (2007). Ethnic identity gratifications selection and avoidance by African Americans: A group vitality and social identity gratifications perspective. Media Psychology, 9, 115–134. Benedict, H. (1992). Virgin or vamp: How the press covers sex crimes. New York: Oxford University Downloaded By: [Feaster, John] At: 21:58 27 July 2010 Press. Byerly, C. M., & Ross, K. (2006). Women and media: A critical introduction. Malden, MA: Blackwell. Cable News Network. (2002, June 19). CNN Evening News [Television broadcast]. Atlanta, GA: Author. Chermak, S. M. (1995). Victims in the news: Crime and the American news media. Boulder, CO: Westview. Dixon, T. L., & Linz, D. (2000). Race and the misrepresentation of victimization on local television news. Communication Research, 27, 547–573. Entman, R. (1994). Representation and reality in the portrayal of Blacks on network television news. Journalism Quarterly, 71, 509–520. Fuhrman, M. (2009). The murder business: How the media turns crime into entertainment and subverts justice. Washington, DC: Regnery Press. Hartley, J. (1998). Juvenation: News, girls, and power. In C. Carter, G. Branston, & S. Allan (Eds.), News, gender, and power (pp. 47–70). New York: Routledge. Harwood, J. (1997). Viewing age: Lifespan identity and television viewing choices. Journal of Broadcasting & Electronic Media, 41, 203–213. Keever, B., Martindale, C., & Weston, M. A. (Eds.). (1997). U.S. news coverage of racial minorities: A sourcebook, 1934–1996. Westport, CT: Greenwood. Kunkel, D., & Smith, S. (1999). The news media’s picture of children in the United States. In C. von Feilitzen & U. Carlsson (Eds.), Children and media: Image, education, participation (pp. 79–86). Go¨tenberg, Sweden: United Nations Educational, Scientific, and Cultural Organization. Lens-Rios, M., Rodgers, S., Thorson, E., & Yoon, D. (2005). Representation of women in news and photos: Comparing content to perceptions. Journal of Communication, 55, 152–167. Mathews, A., & Mackintosh, B. (1998). A cognitive model of selective processing in anxiety. Cognitive Therapy and Research, 22, 539–560. Meyers, M. (1994). News of battering. Journal of Communication, 44, 47–64. Pritchard, D., & Hughes, K. D. (1997). Patterns of deviance in crime news. Journal of Communi- cation, 47, 49–67. Romer, D., Jamieson, K. H., & de Coteau, N. J. (1998). The treatment of persons of color in local television news: Ethnic blame discourse of realistic group conflict? Communication Research, 25, 286–305. Sedlak, A., Finkelhor, D., Hammer, H., & Schultz, D. (2002). National estimates of missing children: An overview. National incidence studies of missing, abducted, runaway, and  216 S.-J. Min & J. C. Feaster thrownaway children. Washington, DC: Office of Juvenile Justice and Delinquency Preven- tion, U.S. Department of Justice. Tuchman, G. (1978). Making news: A study in the construction of reality. New York: Free Press. van Dijk, T. A. (1993). Elite discourse and racism. London: Sage. Yiend, J., & Mathews, A. (2001). Anxiety and attention to threatening pictures. Quarterly Journal of Experimental Psychology, 54A, 665–681. Downloaded By: [Feaster, John] At: 21:58 27 July 2010