WD5879.

289-302 7/31/97 1:55 PM Page 289

A RT I C L E

THE PORTRAYAL
OF MENTAL ILLNESS
ON PRIME-TIME TELEVISION
Donald L. Diefenbach
University of North Carolina at Asheville
■

In this content analysis of television, the portrayal of persons with mental

disorders was highly correlated with the portrayal of violent crime. The
mentally ill were found to be nearly 10 times more violent than the general
population of television characters, and 10 to 20 times more violent (during
a two week sample) than the mentally ill in the U.S. population (over the
course of an entire year). The mentally ill on television were also judged to
have a negative impact on society and a negative quality of life. © 1997
John Wiley & Sons, Inc.

INTRODUCTION
Studies reveal that the mass media are a primary source of public information about
mental illness (Daniel Yanklovich Group, 1990; U.S. President’s Commission on Mental
Health, 1978), and that media portrayals of mental illness are false and negative (Berlin
& Malin, 1991; Gerbner, 1980; Nunnally, 1957; Wahl & Harman, 1989). Experiments have
shown a link between media portrayals and negative attitudes toward mental illness
(Domino, 1983; Thornton & Wahl, 1996; Wahl & Lefkowits, 1989).
The body of literature examining media portrayals of mental illness is small. The
present research updates our understanding of contemporary television portrayals and
adds new methodological elements. It is the first analysis of television content since 1985,
and the first study in this research track to break down portrayals into specific diagnos-
tic categories, as suggested by Wahl (1992). Mental disorders are operationalized ac-
cording to the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric
Association, 1994), and violent crime is operationalized according to the U.S. Depart-
ment of Justice (USDJ, 1994) to allow comparisons of television reality to social reality.

Requests for reprints should be sent to Donald L. Diefenbach, Department of Mass Communication, Univer-
sity of North Carolina at Asheville, Asheville, NC 28804.

JOURNAL OF COMMUNITY PSYCHOLOGY, Vol. 25, No. 3, 289–302 (1997)

© 1997 John Wiley & Sons, Inc. CCC 0090-4392/97/030289-14
WD5879.289-302 7/31/97 1:55 PM Page 290

290 • Journal of Community Psychology, May 1997

MENTAL ILLNESS AND TELEVISION CONTENT

The systematic, scientific investigation of portrayals of mental illness in the media began
in the late 1950s (Gerbner, 1959; Taylor, 1957). Nunnally (1957) compared the views of
mental health experts, the general public, and the mass media. The study found that the
view of mental illness expressed by the mass media was even further removed from health
professionals than that of the lay public.
Goldstein (1979) examined prime-time police and detective dramas and found a
greater number of negative labels were attributed to mentally disturbed deviants than to
deviants whose wrongdoings were attributable to causes other than mental disorders, and
that mentally ill criminals were presented as more dangerous than those not labeled as
mentally ill. Wahl and Roth (1982) examined prime-time television content and found
that “not only did mental illness appear to be a common theme in both news and en-
tertainment media, but the depiction of the mentally ill in these media was found to be
decidedly negative” (p. 600). Fruth and Padderud (1985) concluded, “daytime serials
perpetuate a negative image of mental illness . . .” (p. 387).
Signorielli (1989) used a data set of 17 week-long program samples collected from
1969–1985, but the operationalization of mental illness was narrow—“Mental illness is
only coded when it is specifically mentioned in the story line” (p. 326). Signorielli found
that 72.1% of prime-time dramatic adult characters portrayed as mentally ill hurt or
killed others, and that 75.7% of all mentally ill characters were victims of violence. Only
2.7% of the mentally ill were portrayed in light or comic roles, while 83.5% were por-
trayed in serious roles. The mentally ill also were the characters most likely to be rated
as a “bad” character type and the least likely to be rated as a “good” character type. The
literature of television research, and of print media research (Day & Page, 1986; Matas,
Guebaly, Harper, Green, & Peterkin, 1986; Shain & Phillips; 1991) indicates that the mass
media portray a false, negative, and violent image of mental illness.

PUBLIC INFORMATION AND PUBLIC ATTITUDES

In the report on chronic mental illness conducted by the Daniel Yanklovich Group, Inc.
(1990) seven of every eight survey respondents (87%) cited television and news pro-
grams as a source of information about mental illness. Newspapers were cited by 76%;
radio news, 75%; magazines, 74%; while family and friends were cited by only 51% of
the respondents.
Lopez (1991) surveyed adolescents (age 14–18) because “they will soon be the adults
exerting a force in determining public policy . . .” (p. 271). Adolescents reported their
sources of information about mental illness used to base their attitudes. General read-
ing not related to instruction or school (including books, magazines, newspapers) and
television were ranked first or second by 54.1% of the subjects.
The U.S. President’s Commission on Mental Health (1978) concluded that attitudes
toward mental illness are significantly influenced by mass media sources. Wahl and Har-
man (1989) surveyed members of the National Alliance for the Mentally Ill about their
experience with the stigma of mental illness. The respondents consistently cited media
sources as the perpetuators of mental illness stereotypes and stigma.
Thornton & Wahl (1996) found that subjects exposed to an article portraying a men-
tal patient as a violent criminal demonstrated attitudes significantly more negative to-
ward the mentally ill than did control subjects. Domino (1983) administered attitude
questionnaires to subjects prior to the release of the film, One Flew Over the Cuckoo’s Nest.
WD5879.289-302 7/31/97 1:55 PM Page 291

Portrayal of Mental Illness • 291

After the film was released the questionnaire was readministered. Attitudes toward men-
tal illness changed substantially in a negative direction for subjects who viewed the film.
Wahl and Lefkowits (1989) exposed subjects to a made-for-television movie entitled, Mur-
der by Reason of Insanity. After viewing the film, subjects were administered a standardized
questionnaire (Taylor & Dear, 1981) measuring attitudes toward mental illness. The ex-
perimental groups exposed to the target film showed more negative attitudes than the
control group. Subjects who watched the control film scored similarly to random sub-
jects in the community.

MENTAL ILLNESS AND VIOLENT CRIME

Determining the real world association between mental illness and violent crime is a dif-
ficult task which requires extrapolation from limited data. The body of literature paints
a complicated portrait, but one which allows the formation of an appropriate compari-
son statistic for the present research.
Teplin (1985) analyzed over 1,000 police-citizen encounters. She found “persons ex-
hibiting signs of serious mental disorder were not suspected of serious crimes at a rate
disproportionate to their numbers in the population. The patterns of crime for mental-
ly disordered persons and for non-mentally disordered persons were substantially simi-
lar. These data help dispel the myth that the mentally ill constitute a dangerous group
prone to violent crime” (p. 593).
Cocozza, Melick, and Steadman (1978) specifically examined the relationship be-
tween violent crime and mental illness through the study of two samples of patients re-
leased from New York State Psychiatric Centers in 1968 and 1974–75. They found that
the absolute arrest rates for violent crime (including murder, manslaughter, and assault)
were quite low: 0.9% of the 1968 sample, and 1.7% of the 1974–75 sample in the year-
and-a-half following release. “These absolute numbers are important to counteract the
attitudes and beliefs developed from the media which would suggest that the vast ma-
jority of mental patients are dangerous” (p. 321). The rate of arrest for violent crime in
patients with no prior arrests was less than that of the general population. Cocozza et al.
found that former hospitalization was not a predictor of future arrest. The former pa-
tients subsequently arrested for violent crime were found to possess the same demo-
graphic characteristics which are associated with future arrest in the general population,
for example, socioeconomic status and prior arrest history.
The present research uses the highest estimates in the ranges of violent crime rates
predicted for ex-mental patients for each offense reported by Cocozza et al. to form an
artificially high expected frequency of mentally ill violent criminals. The present as-
sumption for hypothesis testing is that the overall risk of the mentally ill committing a
violent crime (murder, rape, robbery, or non-simple assault), based on relative frequen-
cy of offenses, is 2.43 times as high as the general population. The present research uses
this factor and compares a two-week violent crime rate on television to a fifty-two-week
violent crime rate in the real world, without weighting, to make support of the hypoth-
esis more difficult. This conservative approach helps insure that significance is a result
of true difference between the real world and television portrayals and not an artifact of
variation in sampling methodologies between the U.S. Department of Justice and the
present study. If this study can provide evidence that the mentally ill on television (in just
a two-week sample) are portrayed as significantly more violent than the mentally ill in
the real world (over the course of an entire year) even when making the most violent as-
WD5879.289-302 7/31/97 1:55 PM Page 292

292 • Journal of Community Psychology, May 1997

sumptions about the mentally ill in the real world, then television faces a notable criti-
cism indeed.
Estimated totals of personal victimizations, not merely reported offenses, were used
to assess the real-world offender rate for rape, robbery, and assault (USDJ, 1994, p. 252),
and an estimated number of offenses known to police was used to assess the rate of mur-
der and non-negligent manslaughter (USDJ, 1994, p. 353). If a zero rate of repeat of-
fenders is assumed for an artificially inflated per capita offender rate, then the rate of
Americans who commit a violent crime is about 1.5% per year. Therefore, the expected
frequency of mentally ill violent criminals on television for testing Hypothesis One is
1.5% 3 2.43 (factor) 5 3.65% rate of violent crime.

HYPOTHESES
This study was based on four hypotheses:

Hypothesis One: Mentally ill characters on television are more likely to be violent
criminals than are the mentally ill in the U.S. population.
Hypothesis Two: In the world of television, mentally ill characters are more likely to
be violent criminals than the general population of characters.
Hypothesis Three: In the world of television, mentally ill characters will have a qual-
ity of personal life which is negative in value compared to an assumed population
mean of neutral.
Hypothesis Four: In the world of television, mentally ill characters will have an im-
pact/effect on society which is negative in value compared to an assumed popu-
lation mean of neutral.

To establish that television portrays the mentally ill as more violent than they are in
reality only addresses part of the issue. Hypothesis Two compares the crime rate of the
mentally ill in the world of television to the crime rate of the general population of tele-
vision characters within the same sample. This test only includes speaking characters so
that violent crime offender rates for mentally ill characters and all characters can be es-
tablished.

METHODOLOGY
Sample
The sample universe includes all network, prime-time (8–11 PM) programming broad-
cast between September 1 and November 30, 1994. From this universe, a stratified ran-
dom sampling method was applied to select the equivalent of two weeks of prime-time
programming from four major broadcast network affiliates: ABC, CBS, FOX, and NBC
in the Syracuse, New York market. The first-level sampling unit was the month to be
videotaped. September was randomly chosen. September captures a time of transition
when the summer season is ending and the new fall season is beginning. It is a time of
specials, reruns, and made-for-television-movies, as well as new fall shows. September,
therefore, gives a broad representation of television fare around the calendar, but it is less
representative of the new fall season than October or November. To reduce the sample
to the equivalent of two weeks of programming, uninterrupted blocks of programming
WD5879.289-302 7/31/97 1:55 PM Page 293

Portrayal of Mental Illness • 293

from 8–11 PM for each of the networks were drawn so each network was represented by
two programming blocks for each night of the calendar week. Therefore, CBS has two
Sunday evenings represented in the sample, two Mondays, and so on. The sample for
analysis contained 168 hours of television programming—184 programs. The four ma-
jor broadcast networks constitute approximately a 71% share of all prime-time viewers
during this time frame (Nielson Media Research, 1994).

Procedure
Four coders were recruited to code the sample of programming. To promote uniform
judgment among the coders, all coders were given training in the classification of men-
tal disorders according to the American Psychiatric Association (1994). The attribution
of clinical labels, colloquial labels, and character behavior portrayed or described with-
in a television program were all considered in the assessment of mental illness. Coders
were trained in the identification of violent crime according to the U.S. Department of
Justice (1994) for murder, rape, robbery, and non-simple assault, and the New York State
penal code for other violent offenses including involuntary manslaughter, abuse, kid-
napping, unlawful imprisonment, reckless endangerment, extortion, and intimidation.
The sample of programming was randomly distributed across the four coders. Ten per-
cent of the sampling block was given to multiple coders for reliability checks. Intercoder
reliability was .76 for mental illness classification; .81 for violent criminal classification;
and .84 for victim classification using Krippendorff’s alpha (1980).

Global Assessment
The same four coders made two global assessments for each character portrayed as men-
tally ill, a violent crime victim, or a violent criminal. Coders were directed to rate the
overall portrayal of the character’s quality of personal life based on socioeconomic status,
happiness, strength of interpersonal relationships, balance of family and career, work
and play. Coders also judged the overall portrayal of the character’s impact/effect on soci-
ety based on pro-social and anti-social behavior, altruism and fraud, employment status,
and nature of employment. Coders used a five-point scale to make these judgments: 1 5
very negative; 2 5 somewhat negative; 3 5 neutral (or not enough information); 4 5
somewhat positive; 5 5 very positive. Coders were instructed that the neutral-point on
the global assessment scale (3.0 on a five-point scale) represents the assumed population
mean of all characters. Intercoder reliability for global assessments was .82 using Krip-
pendorff’s alpha (1980).

FINDINGS
Coders identified 96 of 4362 speaking characters to be portrayed as mentally ill (2.2%).
In total, 127 characters were identified to be portrayed as mentally ill (speaking and non-
speaking). Of these 127 characters, 43 (33.9%) were portrayed as perpetrators of a mur-
der, rape, robbery, and/or non-simple assault. When only speaking mentally ill charac-
ters are considered, the offense rate was 29 of 96 (30.2%).
Adding involuntary manslaughter, abuse, kidnapping, unlawful imprisonment, reck-
less endangerment, extortion, and intimidation within the definition of violent crime,
the number of mentally ill characters portrayed committing one or more violent offenses
WD5879.289-302 7/31/97 1:55 PM Page 294

294 • Journal of Community Psychology, May 1997

TABLE 1. Dominant Disorders of Characters Portrayed as Mentally Ill—Random

Sample of Prime-Time Television (September 1994)

Diagnosis Frequency %

Not definable 15 11.8

Psychosis 15 11.8
Paraphilia 12 9.4
Drug abuse 10 7.9
Alcohol abuse 9 7.1
Mental retardation 9 7.1
Depressive disorder 8 6.3
Obsessive compulsive disorder/personality 6 4.7
Attention deficit disorder/hyperactivity 5 3.9
Intermittent explosive disorder 5 3.9
Personality disorder (undifferentiated) 4 3.1
Phobia 4 3.1
Alzheimer’s 3 2.4
Anxiety disorder 3 2.4
Conduct disorder 2 1.6
Dependent personality disorder 2 1.6
Narcissistic personality disorder 2 1.6
Adjustment disorder 1 .8
Amnesia 1 .8
Antisocial personality disorder 1 .8
Dissociative disorder 1 .8
Elective mutism 1 .8
Factitious disorder 1 .8
Hypochondria 1 .8
Kleptomania 1 .8
Multiple personality disorder 1 .8
Paranoia 1 .8
Schizophrenia 1 .8

Total 127 100

is 56 of 127 (44.1%). When only speaking mentally ill characters are considered, the of-
fense rate was 37 of 96 (38.5%).
Psychosis was the most frequently portrayed definable disorder on television. Table
1 provides the distribution of mental disorders portrayed on television.
Of the 184 programs in the sample, 58 programs (32%) contained at least one men-
tally ill character. Genre does affect the frequency and tone of portrayals of mental ill-
ness. In crime dramas, reality-based shows, news magazines and movies, the mentally ill
violent offender rate is over 50% in each category. In other dramas, the violent offend-
er rate is only 11.8% for the mentally ill, and in situation comedies none of the charac-
ters identified as mentally ill were portrayed as violent criminals (Table 2).
Hypothesis One is supported by the data. A one-way chi-square “goodness of fit” test
found a significant difference between the violent crime offender rate among the men-
tally ill in the United States (,1.5% – 3.65% per year) and the mentally ill on television
(33.9% in a two-week sample) (Table 3).
Hypothesis Two is supported by the data. The offender rate for the general popula-
tion of speaking characters on television for murder, rape, robbery, and assault is 3.2%.
WD5879.289-302 7/31/97 1:55 PM Page 295

Portrayal of Mental Illness • 295

TABLE 2. Mentally Ill Violent Criminals by Genre—Random Sample

of Prime-Time Television (September 1994)

Genre Mentally ill Mentally ill criminals

(Hours) characters n %

Crime drama 25 16 (64%)

26 (15.5%)
Realty-based 9 5 (55.6%)
2.5 (7.4%)
Other drama 17 2 (11.8%)
29 (17.3%)
Sitcom 21 0 (0.0%)
37 (22.0%)
News magazine 40 22 (55.0%)
4.5 (8.6%)
Movies 15 11 (73.3%)
34 (20.2%)
Sports 0 0 (0.0%)
8 (4.8%)
Special event 0 0 (0.0%)
7 (4.2%)

Total 127 56 (44.1%)

168 (100%)

TABLE 3. One-Way Chi-Square to Compare Sample and Estimated

U.S. Population Frequency of Mentally Ill Violent Criminals
(Murder, Rape, Robbery, and Assault)—Random Sample
of Prime-Time Television (September 1994)

Television
observed Expected
frequency frequencya

Violent criminal (Murder, 43 (33.9%) 5 (3.9%)

rape, robbery, assault)
Not a violent criminal (Murder, 84 (66.1%) 122 (96.1%)
rape, robbery, assault)

Total 127 (100%) 127 (100%)

Note. Violent crime offender rate for U.S. general population is ,1.5% annually. Mentally ill
characters identified in study: n 5 127. df 5 1. Chi-square 5 300.6. Probability , .0001.
a
Expected frequencies are based on U.S. Department of Justice rates of violent crime (USDJ,
1994) multiplied by a factor extrapolated from Cocozza et al. (1978) to make support of the
hypothesis more difficult.
WD5879.289-302 7/31/97 1:55 PM Page 296

296 • Journal of Community Psychology, May 1997

TABLE 4. One-Way Chi-Square to Compare Sample and Expected

Frequency of Speaking Mentally Ill TV Characters by Violent Criminal
Activity (Murder, Rape, Robbery, and Assault)—Random Sample of
Prime-Time Television (September 1994)

Observed Expected
frequency of frequency of
mentally ill mentally ill
on TV on TV a

Violent criminal (Murder, 29 (30.2%) 7 (7.3%)

rape, robbery, assault)
Not a violent criminal (Murder, 67 (69.8%) 89 (92.7%)
rape, robbery, assault)

Total 96 (100%) 96 (100%)

Note. Violent crime offenders, general population of speaking television characters 5 3.2%.
Speaking mentally ill characters identified in study: n 5 96. df 5 1. Chi-square 5 74.5. Proba-
bility 5 ,.0001.
aExpected frequencies are based on observed offenders rates for all television characters

(3.2%) multiplied by a factor extrapolated from Cocozza et al. (1978) to make support of the
hypothesis more difficult.

The observed offender rate on television for the speaking mentally ill characters was
30.2%, nearly ten times higher than the 3.2% offender rate of the general population
of television characters (Table 4).
Available data also permit the calculation of a television offender rate for all violent
offenses and the testing of Hypothesis Two based on murder, rape, robbery, nonsimple
assault, involuntary manslaughter, abuse, kidnapping, unlawful imprisonment, reckless
endangerment, extortion, and intimidation. The offender rate increases on television for
both the general population of speaking characters and the speaking mentally ill char-
acters. The dramatic spread between them remains, however, with the overall television
offender rate increasing to 3.8% and the offender rate for the mentally ill increasing to
38.5%, remaining approximately ten times higher than the general population of tele-
vision characters (Table 5).
The present study found 24.4% of the mentally ill on television were female and
70.1% were male (5.5% were characters whose gender could not be determined). Part
of the explanation for the under-representation of females as mentally ill on television
might be attributed to the fact that mental illness is highly correlated with violent crime
on television, and it is men who are portrayed on television as responsible for the great-
est share of violent crime. Men account for 87.5% of violent crimes in the U.S. and
women account for 12.5% (USDJ, 1994). This is very close to the ratio portrayed on tele-
vision with men responsible for 82% of violent crime, women responsible for 12.6%, and
5.4% of violent crime portrayed was attributed to criminals whose gender could not be
determined.
The number of characters portrayed as mentally ill, as a violent criminal, or as a vic-
tim of violent crime was 524. These characters were divided into seven exhaustive and
exclusive categories for a rank-order comparison (Table 6).
Mentally ill characters, whatever their status, are at the bottom of the list ranking
WD5879.289-302 7/31/97 1:55 PM Page 297

Portrayal of Mental Illness • 297

TABLE 5. One-Way Chi-Square to Compare Sample and Expected

Frequency of Speaking Mentally Ill TV Characters by Violent
Criminal Activity (All Violent Offenses)—Random Sample
of Prime-Time Television (September 1994)

Observed Expected
frequency of frequency of
mentally ill mentally ill
on TV on TV a

Violent criminal 37 (38.5%) 8 (8.3%)

(all offenses)
Not a violent criminal 59 (61.5%) 88 (91.7%)
(all offenses)

Total 96 (100%) 96 (100%)

Note. Violent crime offenders, General Population of speaking television characters 5 3.8% (All
violent ofenses). Speaking mentally ill characters identified in study: n 5 96. df 5 1. Chi-square
5 114.5. Probability 5 ,.0001.
aExpected frequencies are based on observed offender rates for all television characters (3.8%)

multiplied by a factor extrapolated from Cocozza et al. (1978) to make support of the hy-
pothesis more difficult.

TABLE 6. Rank-Order of Quality of Personal Life Assessments—Random Sample

of Prime-Time Television (September 1994)

Category Rank M n SD

Criminals who are also victims of crime 1 3.11 18 0.67

Victims of crime 2 3.05 214 0.81
Violent criminals 3 2.57 165 0.81
Mentally ill criminals 4 2.52 50 0.93
Mentally ill 5 2.46 65 0.79
Mentally ill victims of crime 6 2.33 6 1.21
Mentally ill criminals also victims of crime 7 1.66 6 1.21

Rank-Order of Impact/Effect on Society Assessments—Random Sample

of Prime-Time Television (September 1994)

Category Rank M n SD

Victims of crime 1 3.09 214 0.74

Mentally ill 2 2.75 65 0.77
Mentally ill victims of crime 3 (Tie) 2.16 6 0.98
Criminals who are also victims of crime 3 (Tie) 2.16 18 1.09
Mentally ill criminals also victims of crime 5 1.66 6 1.03
Violent criminals 6 1.61 165 0.89
Mentally ill criminals 7 1.30 50 0.64
WD5879.289-302 7/31/97 1:55 PM Page 298

298 • Journal of Community Psychology, May 1997

TABLE 7. T- Tests to Compare Mentally Ill Television Characters with a Population Mean of
All Television Characters—Random Sample of Prime-Time Television (September 1994)

Sample Mean Population

Variables (SD) Mean t dt Significance

Quality of life 2.44 3.00 29.65 126 ,0.01

(0.89)
Impact on society 2.13 3.00 27.02 126 ,0.01
(1.01)

Note. Rated on a five-point scale. 1 5 “Very Negative” . . . 5 5 “Very Positive.” Population mean is assumed to be neutral, or
3.0. The population variance is unknown. The T-test for an assumed population mean (variance unknown) is in Kanji (1993).

quality of personal life. According to television, it is better to be a victim of violent crime,

or a violent criminal than mentally ill if one is to have a better quality of life. T-tests com-
paring all mentally ill characters with an assumed population mean of neutral support
Hypothesis Three and Hypothesis Four (Table 7).
The mentally ill are portrayed on television as victims of violent crime much less fre-
quently than as perpetrators. This is contrary to the findings of Signorielli (1989) who
used longitudinal data, a narrower operationalization of mental illness, and a broader
operationalization of violence. Signorielli found mentally ill characters were portrayed
slightly more often as victims than as perpetrators (75.7% and 72.1% respectively). The
victimization rate in the present study is 9.5% for all mentally ill characters and 10.4%
when considering only speaking mentally ill characters. The victimization rate for the
general population of speaking characters in the present study is 4.1%. Therefore, the
mentally ill are victimized at a rate more than twice that of the general population on
television (Table 8). In addition, the present data indicate that the mentally ill are por-
trayed as criminals four times more often than they are portrayed as victims. The Alliance
for the Mentally Ill of New York State (1989) reports that, “In fact, people with a men-
tal illness are much more likely to be victims of violence than its cause” (p. 5, emphasis
in original).
The preceding analysis includes all disorders portrayed on television classified in the
Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association,
1994). Wahl (1992) recommends, however, that “mental retardation, substance abuse,
and organic disorders associated with aging be treated separately from other psychiatric
disorders in examining media portrayals; these exclusions would yield a less contami-
nated picture of the kinds of conditions most mental health professionals and advocates
are thinking about when they express concerns about public views of mental illness. Bet-
ter still, would be a breakdown of portrayals into more specific diagnostic categories”
(p. 350).
Table 9 presents a complete breakdown of diagnostic categories with correlation to
violent crime on television to allow researchers to construct various classification para-
meters. Chi-square tests excluding drug and alcohol abuse portrayals also support Hy-
pothesis One (p , .0001) and Hypothesis Two (p , .0001) in the present study. Since
no characters with mental retardation or organic disorders associated with aging were
portrayed as violent criminals, exclusion of these classifications only strengthen the hy-
potheses.
WD5879.289-302 7/31/97 1:55 PM Page 299

Portrayal of Mental Illness • 299

TABLE 8. Crime Victimization Across Mental Disorders—Random Sample

of Prime-Time Television (September 1994)

Frequency
Total of crime Victims per
Diagnosis frequency victims disorder

Not definable 15 1 (8.3%) 6.7%

Psychosis 15 1 (8.3%) 6.7%
Paraphilia 12 2 (16.6%) 16.6%
Drug abuse 10 1 (8.3%) 10.0%
Depressive disorder 8 1 (8.3%) 12.5%
Obsessive compulsive disorder/personality 6 1 (8.3%) 16.6%
Phobia 4 1 (8.3%) 25.0%
Adjustment disorder 1 1 (8.3%) 100.0%
Amnesia 1 1 (8.3%) 100.0%
Antisocial personality disorder 1 1 (8.3%) 100.0%
Anxiety disorder 3 1 (8.3%) 33.3%
All other disorders portrayed 58 0 (0.0%) 0.0%

Total 127 12 (100%)

TABLE 9. Violent Criminal Activity Across Mental Disorders—Random Sample

of Prime-Time Television (September 1994)

Frequency Violent
Total of violent criminals
Diagnosis frequency criminals per disorder

Not definable 15 8 (14.2%) 5.3%

Psychosis 15 13 (23.2%) 86.7%
Paraphilia 12 12 (21.4%) 100.0%
Drug abuse 10 5 (8.9%) 50.0%
Alcohol abuse 9 2 (3.6%) 22.2%
Obsessive compulsive disorder/personality 6 3 (5.4%) 50.0%
Intermittent explosive disorder 5 5 (8.9%) 100.0%
Personality disorder (undifferentiated) 4 1 (1.8%) 25.0%
Conduct disorder 2 2 (3.6%) 100.0%
Dependent personality disorder 2 1 (1.8%) 50.0%
Posttraumatic stress disorder 2 1 (1.8%) 50.0%
Adjustment disorder 1 1 (1.8%) 100.0%
Antisocial personality disorder 1 1 (1.8%) 100.0%
Anxiety disorder 3 1 (1.8%) 33.3%
All other disorders portrayed 40 0 (0.0%) 0.0%

Total 127 56 (100%)

WD5879.289-302 7/31/97 1:55 PM Page 300

300 • Journal of Community Psychology, May 1997

CONCLUSIONS

Television portrays the mentally ill as significantly more violent than other television
characters and significantly more violent than the mentally ill in the real world. In addi-
tion, prime-time television portrays the mentally ill as having a quality of personal life
which is negative in value and an impact/effect on society which is negative. In fact, the
mentally ill are portrayed on television as having a quality of personal life which is more
negative than violent criminals.
Each program and each network were weighted equally in the present study. This
was done to present the demographics of network prime-time television as a self-con-
tained society. Social effects of television, however, come principally from viewers watch-
ing programs, and the variation in viewership between networks and individual programs
must be considered in the social effects argument. Future research should not only seek
to replicate the trend of the present findings, but match content analysis data with view-
er survey data to allow correlation of viewer exposure to content categories and por-
trayals with viewer attitudes toward mental illness.
For future research, a segment of programming capturing some or all of October is
recommended as a purposive sampling frame to represent the fall season. During Sep-
tember the season is yet in transition, and November is confounded by sensational pro-
gramming during ratings “sweeps.”
Given the above findings, the author recommends a three-tiered policy approach
which seeks: (1) long-term, (2) intermediate-term, and (3) immediate movement toward
achieving more accurate media portrayals of mental illness. For the long-term, a national
conference on the subject of mental illness and the media, is needed to provide an in-
formation exchange between mental health professionals, scholars, and mass media
practitioners to heighten awareness of the issues and to formulate practical remedies. In-
termediate-term results may be facilitated by educators. Academic units in psychology,
journalism, and mass communication can sensitize future practitioners in mental health
care, news reporting, and fiction writing to the realities of mental illness, and the me-
dia’s inaccurate, and widely disseminated, version. This will allow psychologists to better
communicate with the media, and allow tomorrow’s media practitioners to better un-
derstand the larger implications of their portrayals.
The immediate step toward accurate portrayals, however, depends on the action of
current media practitioners. Relentlessly portraying the mentally ill as violent criminals,
and vice versa, not only has negative social implications, but it reflects poorly on the
artists perpetuating these portrays. Real creativity is expressed by transcending stereo-
types to capture an audience. The author challenges current media practitioners to
demonstrate their creativity, and fairness.

REFERENCES

Alliance for the Mentally Ill of New York State (1989). Understanding Mental Illness. Deer-
field, MA: Channing L. Bete.
American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders
(4th ed.). Washington, DC: Author.
Berlin, F. S., & Malin, H. M. (1991). Media distortion of the public’s perception of recidivism
and psychiatric rehabilitation. American Journal of Psychiatry, 148, 1572–1576.
WD5879.289-302 7/31/97 1:55 PM Page 301

Portrayal of Mental Illness • 301

Cocozza, J. J., Melick, M. E., & Steadman, H. J. (1978). Trends in violent crime among ex-men-
tal patients. Criminology, 16, 317–355.
Daniel Yanklovich Group, Inc. (1990). Public Attitudes Towards People with Chronic Mental Illness.
Boston, MA: Author.
Day, D. M., & Page, S. (1986). Portrayal of Mental Illness in Canadian Newspapers. Canadian Jour-
nal of Psychiatry, 31, 813–816.
Domino, G. (1983). The impact of the film, “One Flew Over the Cuckoo’s Nest,” on attitudes to-
ward mental illness. Psychological Reports, 53, 179–182.
Fruth, L., & Padderud, A. (1985). Portrayals of mental illness in daytime television serials. Jour-
nalism Quarterly, 62, 384–386, 449.
Gerbner, G. (1959). Mental illness on television: A study of censorship. Journal of Broadcasting, 3,
293–303.
Gerbner, G. (1980). Dreams that hurt: Mental illness in the mass media. In R. C. Baron, I. D. Rut-
man, & G. Klaczynska (Eds.), The Community Imperative (pp. 19–23). Philadelphia: Horizon
House Institute.
Goldstein, B. K. (1979). Television portrayals of mentally disturbed deviants in prime-time police/detective
dramas. University of New Mexico: Unpublished Thesis.
Kanji, G. K. (1993). 100 Statistical Tests. London: Sage.
Krippendorff, K. (1980). Content analysis: An introduction to its methodology. Thousand Oaks, CA:
Sage.
Lopez, L. R. (1991). Adolescents’ attitudes towards mental illness and perceived sources of their
attitudes: An examination of pilot data. Archives of Psychiatric Nursing, 5, 271–280.
Matas, M., Guebaly, N., Harper, D., Green, M., & Peterkin, A. (1986). Mental illness and the
media: Part II. Content analysis of press coverage of mental health topics. Canadian Journal of
Psychiatry, 31, 431–433.
Nielson Media Research (1994). Ratings: Week 8. Broadcasting & Cable, Nov. 21, 36.
Nunnally, J. (1957). The communication of mental health information: A comparison of the
opinions of experts and the public with mass media presentations. Behavioral Science, July,
222–230.
Shain, R., & Phillips, J. (1991). The stigma of mental illness: Labeling and stereotyping in the
news. In L. Wilkins & P. Patterson (Eds.), Risky business: Communicating issues of science, risk, and
public policy (pp. 61–74). Westport, CT: Greenwood Press.
Signorielli, N. (1989). The stigma of mental illness on television. Journal of Broadcasting and Elec-
tronic Media, 33, 325–331.
Taylor, S. M., & Dear, M. J. (1981). Scaling community attitudes toward the mentally ill. Schizo-
phrenia Bulletin, 7, 225–240.
Taylor, W. L. (1957). Gauging the mental health content of the mass media. Journalism Quarterly,
34, 191–201.
Teplin, L. A. (1985). The criminality of the mentally ill: A dangerous misconception. American Jour-
nal of Psychiatry, 142, 593–599.
Thornton, J. A., & Wahl, O. F. (1996). Impact of a newspaper article on attitudes toward mental
illness. Journal of Community Psychology, 24, 17–25.
United States Department of Justice (1994). The Sourcebook of Criminal Justice Statistics: 1993.
Washington, DC: Author.
United States President’s Commission on Mental Health (1978). Report to the President from
the President’s Commission on Mental Health. Washington, DC: U.S. Government Printing Office.
Wahl, O. F. (1992). Mass media images of mental illness: A review of the literature. Journal of Com-
munity Psychology, 20, 343–352.
WD5879.289-302 7/31/97 1:55 PM Page 302

302 • Journal of Community Psychology, May 1997

Wahl, O. F., Harman, C. R. (1989). Family views of stigma. Schizophrenia Bulletin, 15, 131–139.
Wahl, O. F., Lefkowits, J. Y. (1989). Impact of a television film on attitudes toward mental illness.
American Journal of Community Psychology, 17, 521–528.
Wahl, O. F., & Roth, R. (1982). Television images of mental illness: Results of a metropolitan
Washington media watch. Journal of Broadcasting, 26, 599–605.