Abstract
The public's primary source of information on mental illness is the media [1–3]. Since Nunnally's studies [4, 5] of popular conceptions of mental health, several researchers have built on his work describing print media representations of mental illnesses [6–10]. One review of such work [11] shows that depictions of mental illnesses are relatively common in the media, that these depictions are biased towards the more severe forms of mental illness, and attend disproportionately to bizarre behaviours. Of greatest concern to mental health professionals is that many depictions convey the impression of a strong association between mental illness and violence [6, 8, 11].
To date, there are a number of methodological limitations in this area of work. First, there are problems with the representativeness of the samples obtained [4, 5, 9, 10]. Second, many samples were compiled retrospectively [6, 8, 9]. Third, items were selected from a previously existing database which meant that the authors were unable to operate with their own definition of mental disorder, or to evaluate how important mental illness was to the events reported [6, 8]. And fourth, none of the samples can claim to be a prospective national sample. Other problems with these studies are that they may not have adequately looked for positive elements in the depictions of individuals with mental illness. Further, there appears to have been little attention given to who is providing the information about mental illness in the articles. It seems reasonable to suggest that media depictions would be more positive should individuals with a mental illness be directly quoted or enabled to present their own stories.
With these limitations in mind, we set out to prospectively collect a comprehensive national sample of print media that depicted mental illness. Our primary goals were (i) to search for any positive representations of mental illness, (ii) to search for other (neutral or negative) representations of mental illness, and (iii) to analyse how mental illness was depicted or represented within each item. For this purpose we collected a sample of depictions from the New Zealand print media over four weeks.
Methods
A commercial clipping bureau was contracted to provide cuttings of all items with any mental health or illness aspect published throughout New Zealand between February 3rd and March 2nd 1997. The contract required the inclusion of all items about mental health or mental illness, persons with a mental disorder, psychiatric patients, treatments or practitioners and mental health services. Also included were terms associated with mental illness (e.g. ‘mad’, ‘madness’, ‘insane’) and colloquial phrases (e.g. ‘out to lunch’, ‘go completely bananas’). Prior to the contract being placed, one of the authors (RN) corresponded with the bureau and provided detailed definitions of these inclusion criteria.
Each collected item was categorized on seven variables. These were; the nature of the paper in which the article appeared (daily publication vs other), the circulation (< 50 000 vs > 50 000), the type of article (news, letter to the editor, cartoon, photo, advertisement), the page on which it appeared (first vs other), the size of the item (one or two sentences, one to four paragraphs, four paragraphs to half page, or larger), the gender of the person identified as mentally ill, and whether or not that person was given voice or directly quoted within the item.
One of our primary goals was to find positive representations of persons with mental illness. Prior to analysing individual items we developed a list of potentially positive and negative depictions. This list was generated from earlier studies of characters with a mental disorder in film [12, 13] and newspapers [6–8], as well as from our own clinical experiences. Each item was critically read and particular themes, whether positive or negative, were identified and classified. A theme, also referred to as a pattern [14] or interpretive repertoire [15, 16] is a cluster of mutually coherent discursive resources, words, images, and storylines. Each theme was instantiated by what the person(s) identified as having a mental illness said or how that person was reported to have behaved, together with comments, responses or descriptions provided by others.
We also sought to classify how mental illness was specified or implied within each item. Generic or undifferentiated terms such as ‘psychiatric patient’ or ‘mental illness’ were separated from other undifferentiated metaphorical or colloquial terms (e.g. ‘loco’, ‘crackpot’). When an individual was also referred to as having a specific disorder such as depression, schizophrenia or paedophilia within the same item, the generic or colloquial term was not recorded.
Only items where an individual was described as having a mental illness are included in this analysis. For example, paedophilic behaviour, that was not referred to as paedophilia or generically as mental illness was omitted from analysis. However, within the collection period one event concerned an individual who shot several people in a small settlement (Raurimu) killing six and injuring others. In a minority of the relevant items (n = 37 of 161, 23%) this individual was not referred to as mentally ill. However, because all of the items shared a strong common connection, we included these 37 in the analysis.
We searched News Text and News Index databases for any additional newspaper items concerning one of the stories that achieved some prominence during the sample time period. For this story, which was on the topic of confidentiality, we found an additional 5 items. The number of stories located by the commercial clipping bureau on this topic was 50, suggesting that overall the rate of identification of relevant stories was at least 91%.
Results
We collected 600 items depicting a person or persons with a mental illness that were published during the study period. These most commonly consisted of news articles or editorial pieces (n = 562, 93.7%). Additionally, there were 25 letters (4.2%), 10 photos (1.7%), two cartoons and one advertisement. Most of the items (n = 530, 88.3%) were published in daily newspapers which had distributions ranging from 2000 to 100 000 (n = 421, 79.4%) or greater than 100 000 (n = 72, 13.6%; missing information (n = 37, 7.0%). Items ranged in size from news briefs of one or two sentences (n = 82, 13.7%), one to four paragraphs (n = 200, 33%), four paragraphs to half page (n = 220, 36.7%), or larger (n = 65, 10.8%; missing information n = 33, 5.5%).
Two events received extensive coverage throughout the collection period. The first occurred on day 5 (February 8th) when a young man who was identified as having schizophrenia shot several people in the small rural New Zealand settlement of Raurimu. There were 161 items related to this event. The second was initially reported on day 9 (February 10th). This event, reported in 55 items, was the release of a legal opinion by New Zealand's Privacy Commissioner on the disclosure, nearly three years earlier, of a patient's health information by a psychiatric nurse. The patient had been released from a secure unit under the Mental Health (Compulsory Assessment and Treatment) Act 1992. A psychiatric nurse, concerned about the psychiatric patient's potential danger to others, had disclosed health information to an opposition Member of Parliament. Subsequent to this disclosure the patient had kidnapped and attempted to sexually violate an 11-year-old boy [17, 18]. These 55 items were classified as ‘Confidentiality’.
Other items were classified by topic. There were 52 articles (8.7%) on suicide and 29 (4.8%) personal stories or in-depth accounts of a person's experiences with mental illness as told by the person concerned or by an associate. Other common topics included mental health services (n = 117, 19.5%), paedophilia (n = 51, 8.5%), and alcohol or substance use (n = 39, 6.5%). The remainder (n = 96, 16%) included diverse topics such as court reports, insurance coverage, and colloquial language (e.g. crazy or crackpot).
As can be seen in Table 1, the mental illness depictions were predominantly negative. More than half of all items depicted the mentally ill person as dangerous, including practically all Raurimu and all Confidentiality items. Other common forms of negative depictions were criminality and vulnerability. Vulnerability portrayed the patient as incompetent and unable to control their own life. Unpredictability, an inability to anticipate consequences or account for one's own behaviour, and dangerousness to self (self harm or suicidal behaviour) were less commonly depicted and occurred in slightly less than a quarter of all items.
Frequency of positive and negative depictions∗of mental illness for the total print sample (n = 600), and for some subsamples;‡ Raurimu killings (n = 161), confidentiality (n = 55), suicide (n = 52), and personal stories (n = 29)
Positive depictions occurred in slightly more than a quarter of all items (Table 1). All items on confidentiality and more than half of all personal stories had a positive element. Examples of positive elements in depictions included human rights themes, leadership (e.g. director of a commune), sporting prowess or educational accomplishments. We found only one example of an item that was neither clearly positive nor negative.
Clinical diagnostic categories most commonly employed were schizophrenia (n = 52, 8.7%) and depression not otherwise specified (n = 30, 5%). Alcohol or drug use (e.g. marijuana) related to 49 (8.2%) items. Generic terminology such as ‘psychiatric patient’ or ‘mentally ill’ was more common than specific diagnostic categories being present in almost half of all items (n = 284, 47.3%). Colloquial language was present in 51 items (8.5%). Gender, where specified, was most frequently male (n = 328, 54.7%), whereas only 43 items (7.2%) referred specifically to women. No references to children with mental illnesses were found. On only a few occasions (n = 29, 4.8%) was a person designated as having a mental illness also the speaker or given voice in the published item by being quoted directly either by a reporter or by someone else such as a family member, personal friend, or lawyer.
Discussion
Our findings show that the principal discursive resources for portraying mental illness in this nationwide sample of newsprint items were outstandingly negative. First, these discursive resources enabled depictions of mental illness as dangerous to others in more then half of all items. Dangerousness to others was realized in actions that led to police involvement or court appearances creating a strong association between serious crime and mental illness [6, 8, 11]. This is in marked contrast to research showing very limited associations between mental illness and violence [19]. Dangerousness to self as represented by harmful or suicidal behaviour was also conveyed in about one in every five items. Vulnerability and unpredictability can contribute to a view of sufferers of mental illness as exhibiting erratic, frightening, unreliable, or untrustworthy behaviours, or as unable to cope or control their lives therefore becoming subject to the will of others.
We also found character depictions that drew on positive discursive resources in more than one quarter of all items exemplified by human rights and qualities such as leadership and success in educational or sporting arenas. In particular, the human rights theme that asserts that those with a mental disorder are entitled to the same rights as any other person, was the one positive theme evident in all articles on confidentiality. However, as we have previously shown [17], the existence of this theme was not sufficient to ensure a positive portrayal of mental illness since it was both fragmented and outweighed by the more familiar negative depictions, and undermined by contrary argument. Positive themes predominated in personal stories. One hypothesis, which we are currently testing, is that when those identified as having a mental illness speak directly to the readers through the media, a substantially more positive and less stigmatizing sense of mental illness will be conveyed. This sample will also enable us to check a related finding that psychiatrists, when reported in print, make little positive difference to the negative depiction of mental illness and mental health services [20].
We also found that generic or undifferentiated classifications of mental illness such as ‘psychiatric patient’ or ‘mentally ill’, were predominant, which is consistent with findings from television dramas [13, 21]. In children's television however, more colloquial phrases such as ‘crazy’, ‘mad’ or ‘losing your mind’ predominated [22] and in one study of newspaper depictions a ‘composite’ patient was identified as a male individual with schizophrenia [6]. In the absence of a diagnosis, readers may generalize from a particular depiction to other persons with mental illness, a process that is enhanced by readers' pre-existing views on these conditions. This stigmatizing may be mitigated by efforts to convey the clinical symptoms related to the specific diagnosis and their relevance to the reported incidents.
Our findings must be considered in the light of several limitations. We cannot be certain of the completeness of the sample in that we had not conducted an independent search for other relevant items. A search of a news database suggested that around 10% of all items on the topic of confidentiality were missed and it is possible that a similar proportion of items were missed in other topic areas.
Further, the findings do not generalize to media outside of New Zealand. One factor that limits generalizability is the degree to which variables such as tabloid coverage or circulation vary. Also our prospective sample included two unusual events that received considerable attention. One, involving a multiple murder by a psychiatric patient may have skewed later reporting.
We recognize that many terms conveying the themes are subjectively defined and we have not assessed the reliability of judgements involved. Future studies should determine how readers construct depictions and examine whether judgements have an acceptable level of reliability.
Given these limitations, this study confirms how print media portrayals are negative, exaggerated, and do not reflect the reality of most people with a mental illness. We are concerned that these depictions will encourage people to shun patients and oppose facilities to help them [23]. We are also concerned that these resources will undermine efforts by patients, families and their carers to respond empathically. Our findings re-emphasize the challenge we face in attempting to reduce stigmatizing depictions of mental illness.