Depictions of mental illness in adult media are associated with violence or danger and convey the impression that mental illness constitutes a threat (R1—R6). Although there appears to have been a reduction in the frequency of depictions that associate danger with mental illness, such depictions are still common (R7—R9). Other commonly used negative attributions in constructions of mental disease include unpredictability (R6, R10, R11) and social incompetence (R1, R4).
One reasonable assumption is that similar stigmatizing portrayals dominate children’s media because those portrayals are created by adults acculturated to the same stigmatizing values. If this is the case, then mental illness in children’s media might be associated with violence, unpredictability, and incompetence, resulting in those identified as mentally ill to be ostracized. This effect would more likely be the case when negative images are not balanced by real-life positive alternatives among family members or friends. Such an imbalance could encourage children, or even adults, to reject, bully, or become fearful of children who are viewed as mentally or behaviorally different. To break this cycle, we need to prevent the development of prejudicial attitudes toward those with mental illness, especially in young children.
There is, however, a dearth of information upon which to judge the need for interventions that limit stigmatizing depictions of mental illness in children’s media (R12—R16). One goal of this article is to review these studies in the light of their methodological strengths and weaknesses. A second goal is to identify key questions for further research and some of the challenges likely to be confronted in that work. Children’s media may be less appreciated as a potential source of stigmatization of those with mental illness, and we aim to educate and encourage readers, as prospective researchers, to seek to redress the gaps in this public health area.
The mass media have been shown to be the public’s most important source of information regarding mental illnesses (R17, R18). It is tempting, particularly in areas where there is less research, to presume a one-to-one correspondence between media depictions and attitudes. That presumption must be tested against research, although children’s attitudes toward mental illness show parallels with those of adults (R19, R20). For example, de Rosa asked Italian children of various ages to draw "madmen" and "normal people" (R19). One common form of drawing was jolly or jester-like, evoking a sense of freedom and creative opportunity, while another was monstrous with a threatening mien evoking fear. Although the research appears biased by the request to draw "madmen," the repertoire of images was found to be consistent with historical stereotypes of "madmen" as objects of social prejudice (R19). In a comprehensive review of studies investigating children’s knowledge about and attitudes toward mental illness, some children, even at an early age, had already acquired negative attitudes toward mental illness (R20).
We know less, however, about the varied sources of information that children might use in forming an understanding of mental illness. It seems intuitively reasonable that younger children will make relatively less use of print media compared to adults. To this end, it is important to understand how the growing accessibility of electronic media might influence children’s attitudes toward mental disease. For this purpose, electronic media include the television and radio as well as more recent computer games and the World Wide Web. In one recent cross-sectional telephone survey of nearly 1,500 parents of children with an average age of 5 years, television use averaged nearly 1.5 hours per day, while the use of videos and computer games averaged 1.1 hours and 0.5 hours, respectively (R21). One national telephone survey of approximately 1,000 parents of children ages 6 months to 6 years (response rate 40%) found that many of the children were active computer users, and even the very youngest children were highly exposed to television and other screen media (R22).
Central to any research in this area must be the recognition that children, like adults, actively engage with and interpret, rather than passively receive media of interest to them. For example, the meaning of a text, or what it is understood to say, is not equated to the dictionary meanings of individual words or phrases. Instead we assume that children draw on their own experiences of speech and text to interpret media. This is consistent with discourse analysis used in studies of adult media (R10, R23—R25). In discourse analysis, attention is directed to the language, images, storylines, layout, and interrelationships between texts. Through such processes reality is actively constructed as social and situated (R26, R27).
Discourse analysis enables access to the shared discursive resources—words, images, and storylines—utilized by the creators of factual and fictional depictions as well as the readers who engage with those stories. Such research is a useful adjunct to the more familiar content analyses that provide descriptions and frequencies of occurrence of relevant categories such as, gender, danger to others, etiology of condition, and treatment or prognosis. Each method of analysis has important limitations. First, discourse analysis is dependent on the researcher’s interpretation of the materials, and adult researchers’ interpretations are not necessarily "generalizable" to those of children. Second, content analysis tends to reduce the richness of the materials to predetermined categories, in which all items coded in the same category are treated as being identical.
In analyzing children’s media, the issue of the adult reader is rendered more complex by the ongoing development of children’s cognitive abilities and methods for learning about their world. Children learn languages, social relationships, and ways of structuring the world through listening to, engaging with, and participating in portrayals of their world (R28). For example, according to Piaget, children display concrete thinking until around the age of 11 (R29). Additionally, studies on the acquisition of young children’s language, use of labels, and inductive reasoning facilitate understanding of their capacity to generalize or to overgeneralize (R30).
We found four relevant studies and a set of book reviews (R12—R16) pertinent to our objectives. Three of the studies concerned mental illness depictions in film (R13, R15, R16) and one on television (R13). This focus on cinematic depictions is in contrast to studies of adult media where there is a more even distribution of television, cinema, and print media, perhaps reflecting the accessibility of the materials for researchers.
The first published study in this corpus (R12) described portrayals of mental illness in four Walt Disney films: Dumbo, 1941; Alice in Wonderland, 1950; Mary Poppins, 1964; and Beauty and the Beast, 1992. Beveridge found that the featured characters in three of these films, in reacting strongly to their situation, were vilified for being mad when actually sane. For example, when Dumbo’s mother assails those teasing her son, she is judged to be mad by the other circus performers and locked away.
The second study by Wahl et al. (R15) analyzed films that had been publicly released the previous year and were identified by searching the Motion Picture Association of America web site. While it was not clear how many films were located in the search, 49 of those rated G or PG were available at local video stores. The films were rated independently by three raters trained for this purpose, who were instructed to identify any character having a mental illness. While the interrater reliability was high, the basis of judgment of mental illness was unclear. Raters also assessed the demographic characteristics, their physical attributes, violent behaviors, responses by others, treatment, and overall impressions as positive or negative. Twelve of the films (24%) contained at least one character labeled as having a mental illness (e.g., the Grinch in How the Grinch Stole Christmas, Margena in The Little Mermaid, and Cruella in 102 Dalmations). Many of these characters had positive attributes evoking sympathy, although the same characters also tended to threaten or frighten others.
Subsequently, Lawson et al. (R16) performed content analyses on 34 of the 40 full-length animated feature films that were produced by the Walt Disney Company between 1937 and 2001. The six excluded films did not have a consistent storyline, were not available on video, or were primarily educational. Coding was conducted using operational definitions that were piloted with non-Disney films, and interrater coding reliability was high. The vast majority of films (85%) contained references to characters with mental illness, with an average number of references per film being 5. The authors did not seek to apply a set of diagnostic criteria based on behavior and appearance. Instead, ascriptions of mental illness made by characters of themselves or by others were counted. The three most common single word references to mental illness were "crazy," "mad" or "madness," and "nut" or "nutty." By this criterion, 21% of all principal characters were judged to have a mental illness on at least one occasion. The authors concluded that these ascriptions served to segregate, alienate, and denigrate the target who served as objects of derision, fear, or amusement.
To our knowledge, the only reporting on depictions of mental illness in children’s television is one conducted by Wilson et al. and is a 1-week nationally collected prospective sample that used discourse analytic techniques (R13). The nearly 58 hours analyzed included 128 series episodes: 69 cartoon animations, 12 non-cartoon animations, and 47 real-life films. The broadcasters identified these as aimed at children under 10 years old. Nearly one-half (46%) of the episodes contain one or more references to mental illness, predominantly in cartoons. Additionally, six characters, five of whom were not humans, were defined as having a mental illness when they were consistently referred to by others as such. These characters were either evil or "comic insane" and continuously engaged in illogical and irrational actions such as hitting their head against a swinging light and laughing, sucking paper up through the nose, and handcuffing the jury in court. The characters, whether evil or comic, were stereotypically and blatantly portrayed negatively and were referred to as objects of amusement, derision, or fear. In the Wilson et al. study, the reliability of ratings was not determined, nor was the association of mental illness with violence to others assessed. Further, there was a failure to recognize one potentially positive consequence of being "crazy" as being merry or relatively free from social constraints.
None of these studies sought to employ a formal psychiatric categorization of the depicted individuals. In one study of adult television that used such a categorization (R31), there was sufficient evidence provided in the portrayal for all but a minority of cases to reach a diagnosis, which was most commonly of a psychotic disorder. Further, characters portrayed as having a mental illness were depicted performing violent and criminal acts disproportionately, with the calculated rates sometimes being more than 10 times those found in epidemiological studies of persons living with mental illnesses (R31). Viewer age was not treated as a relevant variable because programs lacked clear, consistent demarcations between materials aimed at different age groups.
The set of book reviews that we found included books that were frequently recommended for review or because they filled a gap in the overall mental health field (R14). In each case reviewers included authors, readers, clinical academics, and service users or consumers. A generally sympathetic view of mental health problems was provided in this set (R14). The five books for younger readers that were reviewed included Edward the Crazy Man (schizophrenia); Subtle Knife (schizophrenia); Cinderella’s Bum (distorted body image); House of the Scorpion (effects of trauma); and Dustbin Baby (foster care and suicide of a foster mother).
+
Recommendations for Future Research
Given the limitations in the research to date, there is an urgent need both for replication studies, particularly of children’s television, and of studies that delineate the patterns of portrayals and their use in media other than those studied to date. Central to such research must be the issue of definition. While experience with the analysis of children’s media suggests that the depictions are unlikely to supply the kinds and amount of detail a psychiatrist would desire before making a diagnosis, we suggest that a single verbal descriptor is not an adequate basis for classification. There is a need to distinguish references to madness and mental illness from the identification that a particular character is being portrayed as having a mental illness.
There is also a need to further determine the association in children’s media between mental illness and violence to others and between mental illness and antisocial or criminal behaviors. Echoing the recommendation of Lawson et al. (R16), we emphasize a need to assess the behavior, including verbal and physical aggression directed at those labeled as having mental illness, as an indicator of structural discrimination (R32). On the other hand, it is also important to determine possible positive connotations of depictions, which, to our knowledge, has not been an explicit goal of the studies to date. Additionally, it is important to assess the relative balance of positive to negative connotations. To this end, discourse analytic methods have high utility in that they enable an assessment of how stories and characters are created by words, actions, appearance, and other technical aspects of production. In one example of this, the potentially humanizing theme of respect due a psychiatric patient, particularly his right to privacy, was found to be undermined and even negated by how the journalist told the story (R33).
Our most important recommendation for research in this area is to conduct reception studies that explore how children engage with such materials, how they generalize or "particularize" knowledge about mental illnesses (R34), and whether that knowledge has implications for the children’s interaction with their social world. Relevant factors for investigation in this research include the age (e.g., whether adult or child) and status of the portrayed person and the developmental stage of the viewer or reader. Such reception studies can be undertaken using a focus group methodology. Relevant variables to be considered in constituting groups include social class, age, gender, and ethnicity. Purposive sampling should aim to cover a range of potentially relevant perspectives (R35). Models for this form of research are provided by research on children’s awareness of cigarette advertising and brand identities (R36).
There is a concern that negative attitudes toward mental illnesses acquired during childhood may persist into adulthood (R13, R37), although this issue has not been addressed in longitudinal studies. A related question is whether children’s knowledge, attitudes, and related behaviors can be modified by interventions. Such interventions might include working with media (R38—R40), assertive campaigns in schools using books (R14), or providing children with positive role models of persons who live successfully with mental illness (R25). Alternatively, regulations or standards for production of children’s media could be developed. We would caution, however, that there remains a lack of information to support production or broadcasting standards for children in this area. Wertham’s lurid campaign against comic books (R41) was limited because it lacked any respect for the readers’ understanding of the materials. In the absence of reception studies, it is simply not possible to understand how children attribute meaning to the media with which they engage.
The evidence suggests that there is a preponderance of negative depictions of mental illnesses in children’s media. In this research, the words used to describe mental illnesses were commonly pejorative, and the characters identified as having a mental illness were commonly stigmatized. More research is needed, however, because of the methodological limitations of the work to date. In particular, we have identified a need for studies of how children interpret the available materials. As previously argued by Wahl (R37), the concern is that we will face a future generation who are fully indoctrinated with the negative images and views of our own generation unless we target children with "destigmatizing" programs. This last consideration adds to the urgency for meeting the research challenges identified in this article.