Journal of Psychopathology and Behavioral Assessment, Vol. 22, No.

3, 2000

Development, Reliability, and Validity of the

Beliefs Toward Mental Illness Scale
Michiyo Hirai1,2 and George A. Clum1

The present study describes the development of a new scale to measure Beliefs
toward Mental Illness (BMI), which was designed to measure cross-cultural
differences in such beliefs as well as to predict treatment-seeking behavior
among different cultural groups. A total of 216 students participated in this
study (114 Asian students and 102 American students). A series of factor
analyses examining construct validity for the BMI revealed that the BMI has
three dimensions, including dangerousness, poor social and interpersonal
skills, and incurability. Examination of the reliability estimates for each factor
revealed moderate to high internal consistency of the BMI. Comparisons
between American and Asian students revealed the expected cultural differ-
ences in their beliefs toward mental illness and treatment preference.
KEY WORDS: cross-cultural assessment; treatment-seeking behavior; beliefs toward psycho-
logical disorder.

INTRODUCTION

The present study describes the development of a new scale to measure

Beliefs toward Mental Illness (BMI). The BMI was designed to measure
cross-cultural differences in such beliefs as well as to predict treatment-
seeking behavior among different cultural groups. Non-western individuals
in general, including Asian people in particular, have been underrepre-
sented in psychological treatment facilities which provide western-style
psychological treatment. One recent survey by the National Institute of
Mental Health (Matsuoka, Breaux, & Ryujin, 1997) reported that for all
1
Virginia Polytechnic Institute and State University.
2
To whom correspondence should be addressed at Department of Psychology, Virginia Tech,
Blacksburg, Virginia 24061.

221

0882-2689/00/0900-0221$18.00/0  2000 Plenum Publishing Corporation

222 Hirai and Clum

types of services (e.g., inpatients or outpatients, etc.) across all types of

facilities (e.g., hospitals psychiatric services, mental clinics, community ser-
vices, etc.), Asian American/Pacific Islanders are much less likely than their
Euro-American counterparts to make use of mental health services. Several
authors (Enrique, 1993; Fujii, Fukushima & Yamamoto, 1993; Gaw, 1993;
Kim, 1993; Zhang, Snowden & Sue, 1998) described a lower rate of utiliza-
tion of mental health services among the Asian population including Chi-
nese, Japanese, Korean and other Southeast Asians in the United States,
and Asians’ reluctance to disclose their mental health issues (Zhang et al.,
1998). This trend may reflect beliefs toward mental illness formed in their
countries of origin as Asian people have been reported to seek treatment
consistent with their cultural beliefs (Gaw, 1993; Hatfield, Mohamad,
Rahim, & Tanweer, 1996; Kim, 1993; Lam & Kavanagh, 1996; Wintersteen,
Wintersteen, Mupedziswa, & Cheah, 1997).
Recent research studies have investigated the relationship between
Asians’ acculturation level and their treatment seeking behavior with incon-
sistent findings. For example, Leong, Wagner, and Kim (1995) reported
that Asian Americans’ positive attitudes toward psychological treatment are
related to their acculturation level, whereas Atkinson, Lowe, and Matthews
(1995) reported that Asian Americans’ willingness to seek psychological
treatment is not influenced by their acculturation level.
A variety of research studies have examined the beliefs of Asians
toward mental illness. These beliefs include that mental illness is inherent,
chronic, incurable, and shameful (Fabrega, 1991), that people with mental
illness are dangerous (Whaley, 1997), and that having depression is socially
disadvantageous (Raguram, Weiss, Channabasavanna, & Devins, 1996).
Whaley reported that Asians and Hispanics perceived mentally ill patients
as significantly more dangerous than did white respondents, regardless of
socioeconomic status or level of contact with persons with mental illness.
Other recent studies (Enrique, 1993; Fabrega, 1991; Fujii et al., 1993; Gaw,
1993; Kim, 1993; Ng, 1997) concluded that in Asian countries such as Korea,
China and Japan, mental illness brings shame upon the entire family and
raises concern about the appropriateness of sufferers for such social institu-
tions as marriage, business, and education. Two studies conducted in Hong
Kong (Chou & Mak, 1998; Chou, Mak, Chang, & Ho, 1996) reported a
low social acceptance of mentally ill people (e.g., unwillingness to live near
mental facilities, strict view on discharging mental patients from hospitals).
A study conducted in Australia reported that Asian immigrants were more
likely than Anglo-Australians to think that mentally-ill people are inferior
to normal people and should be restricted to protect the society (Fan,
1999). Johnson and Orrell (1995) stated that Asian Americans were more
likely to see mentally ill people as different from others. Based on these
studies, it seems likely that Asian people see mental illness as being danger-
Beliefs Toward Mental Illness Scale 223

ous, incurable, and shameful, and that it leads to social untrustworthiness.

As Johnson and Orrell (1995) suggest, these stigmas attached to mental
illness are likely to influence peoples’ willingness to acknowledge that they
have a mental illness. Asian people who have such negative beliefs may
be unwilling to acknowledge that they have a mental illness, and subse-
quently avoid mental health resources.
Although the above literature reviews of recent research studies indi-
cated possible relationships between beliefs towards mental illness and
treatment-seeking behavior, comprehensive studies in this area have not
been reported. This deficiency may be attributable to the lack of a compre-
hensive questionnaire to evaluate beliefs towards mental illness. Research
studies to date have measured beliefs toward mental illness using interviews
(Raguram et al., 1996), or telephone surveys (Chou & Mak, 1996; Chou et
al., 1998; Whaley, 1997) with one study using a self-report questionnaire
(Fan, 1999) developed by Cohen and Struening in 1962. The Cohen and
Struening scale contains items measuring such constructs as authoritarian-
ism, benevolence, and the desire to restrict mentally ill people, that are
different from those examined by the BMI.
The present study describes the development of a scale—the BMI—
that measures beliefs toward mental illness, reflective of the studies which
have examined cultural differences in such beliefs. While prior studies
related such beliefs to the avoidance of psychological services these studies
have ignored the relationships of these beliefs to other possible intervention
modalities, including folk medicine and family support. In the present study
we examined whether students who have more negative beliefs about men-
tal illness (e.g., dangerous, shameful) report less willingness to seek psycho-
logical treatment, while students who have more negative beliefs about
psychological disorders are more likely to report willingness to seek cultur-
ally appropriate (for Asian) treatment methods (e.g., folk medicine, reli-
gious healers, etc.). In addition, using the BMI, the present study examined
whether or not foreign students from Asian countries have more negative
beliefs toward psychological disorder than do American students.

METHOD

Subjects

Participants were recruited via flyers posted on campus, email, and

the Internet newsgroups of a large state university in southwest Virginia.
Email was sent to Asian students who were on the international student
list obtained from an International Student Center at the university. Stu-
224 Hirai and Clum

dents who responded to the email, or saw the flyers or the newsgroups and
responded to the investigator were informed that if they were taking either
Introductory Psychology or other psychology course, they would receive
one extra credit for participating in this study, and that if they completed
all the questionnaires, they could win a money prize up to $100. A sign-
up sheet for Introductory Psychology was also used for the recruitment
procedure, informing students of the extra credit and the money prize
opportunities.
A total of 216 students participated in this study. One hundred and
fourteen international students were born in Asia and held student visas,
5 were international students born in other western countries, and 97 were
American students including Caucasian (n ⫽ 89), African American (n ⫽
2), Hispanic (n ⫽ 4), Asian American (n ⫽ 5), and others (n ⫽ 1). ‘‘Asian
countries’’ were those included in the classification of Asian countries by
the World Almanac and Book of Facts (1996).
The majority (n ⫽ 91) of Asian students had been in the United States
less than 5 years, while 11 students had been in the United States between
5 to 7 years, and 12 students had been in the United States more than 7
years. Their countries of origin included Bangladesh (n ⫽ 1), China (n ⫽
31), Hong Kong (n ⫽ 3), India (n ⫽ 22), Indonesia (n ⫽ 8), Japan (n ⫽
5), Macao (n ⫽ 1), Malaysia (n ⫽ 5), Nepal (n ⫽ 5), Pakistan (n ⫽ 2),
Philippines (n ⫽ 1), South Korea (n ⫽ 16), Taiwan (n ⫽ 6), Thailand
(n ⫽ 7), and Vietnam (n ⫽ 1). Eighty five were male and 29 were female,
and 86 were single and 28 were married. Their age range was from 18 to
39 (M ⫽ 25.3, SD ⫽ 5.1). Ten were freshmen, 8 were sophomores, 6 were
juniors, 12 were seniors, 74 were graduate students, and 4 were language
intensive course students. Eight participants were enrolled in the Introduc-
tory Psychology or other psychology courses such as social psychology, and
were given an extra credit. Among American and western students their
countries of origin included Columbia (n ⫽ 1), England (n ⫽ 1), Puerto
Rico (n ⫽ 1), Spain (n ⫽ 2), and the United States (n ⫽ 97). Forty five
were male and 57 were female, and all of the participants were single. Their
age range was from 18 to 28 (M ⫽ 19.6, SD ⫽ 1.6). Thirty two were
freshmen, 26 were sophomores, 26 were juniors, 17 were seniors, and 1
was a graduate student. Ninety students were enrolled in the Introductory
Psychology or other psychology course, and were given an extra credit.

Assessment Instruments

The Beliefs toward Mental Illness (BMI) scale was designed to assess
negative stereotypical views of mental illness, including: (1) that mental
Beliefs Toward Mental Illness Scale 225

illness is incurable, (2) that suffering from mental illness is shameful, (3)
that mentally ill people are dangerous, and (4) that individuals with mental
illness are socially untrustworthy. These dimensions are based on descrip-
tive constructs in the studies by Enrique (1993), Fujii et al., (1993), Gaw
(1993), Fabrega (1997), Johnson and Orrell (1995), Kim (1993), Ng (1991)
and Raguram et al. (1996), but that had not been integrated into a standard-
ized assessment instrument. Initially, a pool of items was developed that
reflected constructs examined in the above studies as well as from the
investigators’ experience with Asian cultures. Several iterations were con-
ducted in which the pool of items was reduced by eliminating redundant
and irrelevant items. The final BMI was composed of 24 statements, in
which each of the above four dimensions is represented by six items each.
The respondents were asked to evaluate their level of agreement with each
statement describing stereotypical views toward mental illness. The items
are rated on a 6-point Likert scale ranging from completely disagree (0)
to completely agree (5). Higher scores reflect more negative beliefs about
mental illness.
The Suinn–Lew Asian Self-Identity Acculturation Scale (SL-ASIA,
Suinn, Figueroa, Lew, & Vigal, 1987) was administered to measure Asian
students’ acculturation level. Because a number of Asian students had been
students in the United States for several or more years it was necessary to
establish that the Asian student sample continue to have Asian cultural
values. The SL-ASIA consists of 26 items covering topics such as language
preferences, ethnic identity, friendship choice, behaviors, generational/geo-
graphic history, and attitudes toward one’s ethnic group. Twenty one items
(item 1 to item 21) were used in the present study based on the instructions
provided by Suinn (personal contact, 1998). Each item is scored on a
continuum ranging from 1.00 which is indicative of low acculturation (or
very Asian) to 5.00 which is indicative of high acculturation (or very angli-
cized). A reliability estimate using Cronbach’s alpha of .91 was obtained
by Suinn, Ahuna, & Khoo (1992). A reliability estimate using Cronbach’s
alpha in the present study was .82 (n ⫽ 100).
The Treatment-Seeking Behavior (TSB) scale is a newly developed
questionnaire to assess treatment-seeking behavior. (Additional informa-
tion on the construction of this scale and on its reliability and validity is
available from the authors.) The questionnaire consists of 16 statements,
which examine preferences for different treatment approaches, including
western psychological treatment, an Asian-based culturally-appropriate
treatment (e.g., folk medicine, religious healers, etc.), medical treatment,
no-treatment seeking, and family care. All items in the TSB were originally
developed from constructs reported in the cross-cultural literature and
from the investigators’ clinical experience. Fourteen items were used in
226 Hirai and Clum

the present study based on results of a factor analysis conducted in this

study. The respondents were required to evaluate their level of agreement
with each statement if they were to experience psychological symptoms.
The items are rated on a 6-point Likert scale ranging from completely
disagree (0) to completely agree (5). Higher scores on each item reflect an
increased likelihood the respondents would seek the designated treatment.
A series of factor analyses with the combined samples (N ⫽ 216), examining
construct validity for the TSB, identified five treatment preference factors
consistent with the items described above. Examination of the reliability
estimates for each factor with the exception of the no-treatment factor
(움 ⫽ .65) revealed moderate to high reliability (움 ⬎ .70). Similar reliability
estimates were found among both American and Asian groups.

Procedures

The BMI, TSB, and two other measures were administered to all
participants. Results from the other measures will be reported in a separate
study. The SL-ASIA was given only to Asian participants to estimate their
level of acculturation and, if needed, to control for acculturation level in
relating beliefs to treatment preference. Prior to answering these question-
naires, the participating individuals were given an informed consent and
the identifying information sheet covering general demographic questions.
The consent form detailed the research purpose, procedure, freedom of
withdrawal, risks and benefits, and compensation. After reading the consent
form individuals who agreed to participate signed it. The individuals were
asked to answer the questions in the same order honestly. The investigator
verbally answered Asian participants’ questions regarding English words
and allowed them to use the dictionary.
All participants were informed that they could request a copy of the
results of this study. Sixteen participants were chosen by lottery to be given
a money prize for their full participation. Specifically, the prizes included
$100 for 2 participants, $50 for 4 participants and $10 for 10 participants.
The winners were notified by email after all data were collected.

RESULTS

Participants

T-tests and Pearson chi square tests were performed to detect any
possible distribution differences on demographic variables between foreign
Beliefs Toward Mental Illness Scale 227

students from Asian countries and American students. These comparisons

revealed a significant distribution difference between the two samples on
age (t ⫽ 10.71, p ⬍ .01), gender (␹2 ⫽ 8.35, p ⬍ .01), marital status (␹2 ⫽
28.77, p ⬍ .01), and class year (␹2 ⫽ 105.28, p ⬍ .01). Within the Asian
group, no difference on any demographic variable was found between
students living in the United States less than 5 years and those living in
the United States 5 or more years. Comparing acculturation level using the
SL-ASIA for Asian students living in the United States less than 5 years
(M ⫽ 2.11, SD ⫽ .26, n ⫽ 85), to Asians students living in the United
States 5 or more years (M ⫽ 2.44, SD ⫽ .51, n ⫽ 15) yielded a significant
difference (t (100) ⫽ 3.90, p ⬍ .001). However, these scores indicated that
both Asian groups identified themselves as Asians and believed in Asian
values more than in American values. No significant relationships between
Asians’ acculturation levels and their treatment preference were found.
Among American students their subsamples were small; therefore all ethnic
groups within the American sample were combined. Only one graduate
student was found in the American group. This individual was eliminated
from the data when demographic variables were used for analyses among
American students because of his unique status among the American stu-
dent sample. Based on these results, Asian and American participants’
demographic variables were analyzed as possible factors related to the
measures of interest.

Construct Validity

A series of exploratory factor analyses were conducted to examine

construct validity in the BMI scale. The analyses were conducted separately
for the Asian and the American and western students followed by a factor
analysis of the combined group. The principal components method and
varimax with Kaiser normalization rotation were employed in a series of
factor analyses on each measure. The analyses were conducted both by
specifying four as the number of factors to be extracted and without speci-
fying the number of factors. In each process, a factor loading level of .40
was used as the cutoff criterion for factor membership as well as was an
examination of the face validity of each item. The factors produced by the
analyses with the combined group were given first priority because the
combined group had the largest sample size and produced the highest
reliability estimates. Items that loaded on more than one factor were placed
with the factor determined most appropriate after examining both factor
loading and face validity. Items were retained in the factor where they
had a lower factor loading only when the factor reliability for the newly
228 Hirai and Clum

constituted factor did not change significantly from the original one. Based
on this procedure, meaningful factors on the BMI were identified and titled.
The 4-factor solution produced the most meaningful interpretation of
the BMI. One of the four factors whose items had little face validity was
eliminated from the final solution. The three remaining factors were titled:
(1) Dangerousness, (2) Poor social and interpersonal skills, and (3) Incur-
ability. The excluded factor was composed of items expected to factor on
two separate factors, and included 5 items. These items were moved to one
of the three retained factors or eliminated using the rules specified above.
Twenty-one items of the original 24 items were included in the final solution.
Items in each factor for this scale are presented in Table I along with their
factor loadings.
The inter-factor correlations were obtained from each of the subgroups
as well as the combined group. The patterns of the inter-factor correlations
were similar in all three groups. Within the combined group the correlations
ranged from .51 to .66 (p ⬍ .01).
Cronbach’s alpha reliability for each factor of the developed scales
was obtained for Asian students, American students, and the combined
group. These results are shown in Table II. Examination of the reliability
estimates shown in Table II revealed moderate to high internal consistency
of the BMI, with comparable results in both the Asian and American
groups. Item-total correlations for each subscale were obtained for the
combined group and each of the subgroups to further examine the reliability
of the scales. The results from this analysis demonstrated significant item-
total correlations (.22 ⬍ r ⬍ .72, p ⬍ .01) for all items within the com-
bined group.

Concurrent Validity

First, scores on the TSB subscales were examined for each student
group. An average score per subscale was calculated for each group. The
American and Asian groups were compared on the average scores within
subscales. No significant difference was found and the order of preference
was the same for both groups. The most frequently endorsed approach was
family care, followed by psychological intervention, medical intervention,
folk medicine intervention, and no treatment. Next, correlations were com-
puted between scores of each factor for the BMI and treatment preference
scales. A summary of these correlations is shown in Table III. Correlations
between demographic variables and scores of each factor for the BMI and
the treatment preference scale were also conducted to investigate whether
or not there were any significant relationships between the demographic
Beliefs Toward Mental Illness Scale 229

Table I. Extracted Factors and Their Factor Loadings of the BMI for the Combined Group
Factor % of
Items loadings variance
Factor 1: Dangerousness 54.9
1. A mentally ill person is more likely to harm others than .80 34.5
a normal person.
2. Mental disorder would require a much longer period of .46 8.6
time to be cured than would other general diseases.
3. It may be a good idea to stay away from people who .69 5.3
have psychological disorder because their behavior is
dangerous.
6. Mentally-ill people are more likely to be criminals. .68 4.1
13. I am afraid of people who are suffering from psycholog- .46 2.4
ical disorder because they may harm me.
Factor 2: Poor interpersonal and social skills. 26.3
4. The term ‘‘Psychological disorder’’ makes me feel em- .60 5.0
barrassed.
5. A person with psychological disorder should have a job .70 4.5
with minor responsibilities.
8. I am afraid of what my boss, friends, and others would .52 3.5
think if I were diagnosed as having a psychological dis-
order.
11. It might be difficult for mentally-ill people to follow social .52 2.8
rules such as being punctual or keeping promises.
12. I would be embarrassed if people knew that I dated a .61 2.6
person who once received psychological treatment.
14. A person with psychological disorder is less likely to .40 2.1
function well as a parent.
15. I would be embarrassed if a person in my family .39 2.0
became mentally ill.
19. Mentally-ill people are unlikely to be able to live .54 1.5
by themselves because they are unable to assume respon-
sibilities.
20. Most people would not knowingly be friends with a .63 1.5
mentally-ill person.
24. I would not trust the work of a mentally-ill person .46 .8
assigned to my work team.
Factor 3: Incurability 13.9
7. Psychological disorder is recurrent. .64 3.6
9. Individuals diagnosed as mentally ill will suffer from .80 3.2
its symptoms throughout their life.
10. People who have once received psychological treatment are .74 3.0
likely to need further treatment in the future.
18. I do not believe that psychological disorder is ever .72 1.6
completely cured.
21. The behavior of people who have psychological disorders .47 1.3
is unpredictable.
22. Psychological disorder is unlikely to be cured regardless .59 1.2
of treatment.
230 Hirai and Clum

Table II. Reliability Estimates of Total Score and Factors Analyzed Subscales of the BMI
All students Asians Americans
Factors 움* n 움* n 움* n
BMI total 0.91 210 0.91 112 0.89 98
BMI subscales
Factor 1: Dangerousness 0.75 216 0.80 114 0.77 102
Factor 2: Poor social skills 0.84 214 0.82 113 0.74 101
Factor 3: Incurability 0.82 211 0.81 112 0.85 99
*Cronbach’s Alpha.

variables and the BMI scale and treatment preferences. Language intensive
students were not included in analyses with education level since their
grade placement was undetermined.
Several significant correlations were found between the BMI scale and
subscales and treatment preferences, although these relationships differ
depending on which group of students is being examined. Preferences for
psychological treatment and family care were not correlated with any BMI
subscale. Interestingly, these were the two treatment options endorsed most
frequently by both student groups. Asian students who believe that mentally
ill people are dangerous and that psychological disorder is incurable were
less likely to prefer the medical treatment option. Asian students who
believe that mentally ill people possess poor social and interpersonal skills
were more likely to choose folk medicine remedies. Among American
students, all BMI factors were positively correlated with preference for

Table III. Correlations Between BMI Subscale and Treatment Preferences

TSB Factors
No Folk Psychological Medical Family
BMI Factors treatment medicine treatment treatment care
All participants
Dangerousness .17* .21** .03 ⫺.09 ⫺.04
Poor social skills .24** .27** .03 ⫺.09 ⫺.05
Incurability .13 .13 ⫺.09 ⫺.05 .04
All Asians
Dangerousness .15 .16 ⫺.01 .23* ⫺.08
Poor social skills .13 .19* .03 ⫺.15 ⫺.03
Incurability .14 .10 ⫺.11 ⫺.21* ⫺.02
All Americans
Dangerousness .14 .15 .06 .22* .03
Poor social skills .31** .17 .01 .25* ⫺.02
Incurability .10 .11 .12 .22* .12
*Correlation is significant at the 0.05 level (2-tailed).
**Correlation is significant at the 0.01 level (2-tailed).
Beliefs Toward Mental Illness Scale 231

medical treatment. American students who believe that individuals with

mental illness have poor social and interpersonal skills were more likely
to prefer the no-treatment option. For the combined sample, participants
who believe that mentally ill people are dangerous and have poor social
and interpersonal skills were more likely to choose both the no-treatment
and folk medicine options.
Next examined were the correlations between demographic variables
and the BMI and treatment preference variables. For the combined group
males were more likely to have negative beliefs that psychologically ill
people are dangerous (r ⫽ .23, p ⬍ .01) and have poor social and interper-
sonal skills (r ⫽ .35, p ⬍ .01). Males were also more likely to prefer the
no-treatment option (r ⫽ .38, p ⬍ .01) and folk medicine option (r ⫽ .19,
p ⬍ .01). Older students were more likely to believe that mentally ill people
have poor social and interpersonal skills (r ⫽ .34, p ⬍ .01). They were also
more likely to endorse folk medicine remedies (r ⫽ .17, p ⬍ .05) and less
likely to endorse the medical treatment option (r ⫽ ⫺.15, p ⬍ .05). Among
Asian students, males were more likely to believe that mentally ill people
have poor social and interpersonal skills (r ⫽ .21, p ⬍ .05). They were also
more likely to choose the no-treatment option (r ⫽ .32, p ⬍ .01). Older
Asian students were more likely to choose the psychological treatment
modality (r ⫽ .21, p ⬍ .05). Among American students males were more
likely to hold negative beliefs that mentally ill people are dangerous (r ⫽
.26, p ⬍ .01) and have poor interpersonal and social skills (r ⫽ .33, p ⬍
.01). They were also more likely to prefer no-treatment (r ⫽ .37, p ⬍ .01)
and folk medicine remedies (r ⫽ .33, p ⬍ .01). Thus, age and gender were
consistently found to be related to both beliefs toward mental illness and
treatment preference.

BMI Differences Between Asian and American Students

T-tests comparing Asian and American students were conducted on

each factor and total score of the BMI. A summary of these results is
presented in Table IV. As Table IV indicates, Asian students are more
likely to think that people suffering from psychological disorder are danger-
ous and poor at social and interpersonal relationships. No difference be-
tween groups was found on the incurability factor.
Because significant differences of gender and age existed between
American and Asian students, hierarchical regression analyses were con-
ducted separately for the three BMI subscales to control these demographic
variables in order to determine whether the two ethnic groups had different
beliefs about mental illness (coded as American ⫽ 0 and Asian ⫽ 1). These
232 Hirai and Clum

Table IV. Comparisons of American and Asian Students on all BMI Measures
Test
Asian students American students statistics
Factors M SD n M SD n t
BMI total 56.09 19.22 112 43.13 15.33 98 5.35*
BMI subscales
Dangerousness 12.55 5.04 114 10.16 4.53 102 3.66*
Poor social skills 24.57 9.07 113 16.72 7.09 101 6.99*
Incurability 14.98 5.76 112 13.87 5.50 99 1.43
*p ⬍ .001.

results are presented in Table V. As can be seen from these analyses,

cultural differences between American and Asian students were predicted
by their beliefs that mentally ill people are dangerous and have deficits in
social and interpersonal skills after controlling for demographic differences.
Further comparisons revealed no significant relationships between
Asians’ acculturation levels and the BMI factors were found. However,
considerable variability among Asian students in the relationships between
their nationality and the BMI factors and treatment preferences. Students
from China, India, and South Korea were compared in separate analyses
to other Asian students because of the relatively large number of students

Table V. Regression Analyses Predicting Cultural Membership

from Beliefs Toward Mental Illness Controlling for Gender and
Age Variables
Variables 웁 R2change Fchange
Step 1 .37 62.37**
Age .55**
Gender .15*
Step 2 .02 7.02**
Age .54**
Gender .12*
Dangerousness .15**
Step 1 .37 61.54**
Age .55**
Gender .14*
Step 2 .05 15.81**
Age .49**
Gender .08
Dangerousness .23**
Note: Predictor variables which did not predict ethnic difference
were not reported. Gender indicates male. n ⫽ 213 for the danger-
ousness factor and n ⫽ 211 for the social skill deficits factor.
*p ⬍ .05.
**p ⬍ .01.
Beliefs Toward Mental Illness Scale 233

from these countries. Chinese students are more likely to think that people
suffering from psychological disorder are dangerous than other Asian stu-
dents (t(112) ⫽ 2.85, p ⬍ .01). Indian students are more likely to think that
mental illness is incurable than other Asian students (t(110) ⫽ 2.34, p ⬍
.05). Students from South Korea are more likely to choose family care for
psychological disorder than other Asian students (t(112) ⫽ 2.24 p ⬍ .05).

DISCUSSION

The present study describes the development of a self-report question-

naire, assessing beliefs toward mental illness. The results from this study
indicate that the BMI is a valid measure with satisfactory psychometric
properties. In developing the new scale the present study used American
students as a control group, in contrast to most of the recent research
studies examining cultural differences on beliefs toward mental illness or
treatment-seeking behavior (Chou & Mak, 1996; Chou et al., 1998; Hatfield
et al., 1996; Raguram et al., 1996; Wintersteen et al., 1997). Total scale score
and two of the subscales of the BMI were significantly different for Asian
and American students and were significantly related to treatment-seeking
behavior for both American and Asian students.
A series of factor analyses with the combined samples (N ⫽ 216),
examining construct validity for the BMI, revealed that the BMI has three
dimensions, including dangerousness, poor social and interpersonal skills,
and incurability. The factor structure of the BMI was as expected with
the exception that items reflecting both embarrassment and untrustworthy
feelings toward mentally ill people were integrated into the poor social
skills factor. In general, identified factors and item distributions produced
by the combined student group were consistent with expectations and reflect
beliefs toward mental illness discussed in general terms in the literature.
Examination of the reliability estimates for each factor revealed moder-
ate to high internal consistency of the BMI. Similar reliability estimates
were found among both American and Asian groups, indicating that the
measure can be utilized with either group. Item-total correlations revealed
moderate to high correlations. These results indicate that the BMI is a
relatively reliable measure for both Asian and American Students.
Comparing American and Asian students, Asian students were more
likely to believe that mentally ill individuals are dangerous and have poor
social and interpersonal skills. A series of regression analyses demonstrated
that these differences continued to exist after controlling for age and gender.
These results support the expectation that the BMI can successfully assess
cultural differences. The results of comparisons among Asian students indi-
234 Hirai and Clum

cate that there is some variability among Asian students from specific
countries in their beliefs toward mental illness and treatment preferences.
These findings argue for cross-cultural comparisons that utilize more homo-
geneous groupings of participants.
Some of the expected relationships between beliefs toward mental
illness and treatment preferences were found. On the other hand, no rela-
tionships existed between the BMI subscales and preference for psychologi-
cal treatment and family care. Inability to predict these two treatment
options by the BMI may reflect the general tendency of both groups of
students to frequently select each of these treatment options.
The BMI subscales reflective of beliefs that mentally ill people are
both dangerous and have poor social skills were more prominent among
Asian than American students. These beliefs also proved to be related to
preferences for no-treatment or folk medicine among all participants.
Within subgroups the belief that mentally ill people had poor social skills
was related to preference for folk medicine remedies among Asian students
and to a preference for the no-treatment option among American students.
Those subscales that differentiate Asian from American students also pre-
dict no-treatment or folk medicine remedies. Considered together these
results provided support for the validity of the BMI.
Interestingly, the two groups of students present different views with
regard to the relationships of negative beliefs (i.e., dangerousness, incurabil-
ity, and poor social skills) to the medical treatment option. Asian students
are likely to select medical treatment—a western style option—when they
have positive beliefs (i.e., curable, not dangerous) toward mental illness.
American students, on the other hand, select the medical option when they
have negative beliefs. It is possible that Asians view the medical option
positively while American students view it negatively.
The incurability subscale did not differentiate the two student groups
nor did it for the most part relate to treatment preferences. It may be that
incurability represents a more factually based attitude—you either know
mental illness is treatable or you do not. Dangerousness and social skill
deficits, on the other hand, may represent attitudes shaped by one’s culture
that are in turn related to culturally-based treatment preferences. Given
this conjecture, however, it is difficult to reconcile the incurability subscale’s
failure to correlate with the no-treatment option. It is also worth noting
that students from India were significantly more likely than other Asian
students to believe that mental illness is incurable.
There are several issues regarding the conclusions arrived at in the
present study. First, most of the Asian participants who select study in
the United States may be more familiar with western ideas than students
remaining in their countries of origin. Second, most of the previous studies
Beliefs Toward Mental Illness Scale 235

identifying Asians’ beliefs and attitudes toward psychological disorders and

mental illness were conducted with community samples (Chou & Mak,
1996; Chou et al., 1998; Hatfield et al., 1996; Lam & Kavanagh, 1996;
Matsuoka et al., 1997; Raguram et al., 1996; Whaley, 1997; Wintersteen et
al., 1996; Zhang et al., 1998), while only college students were used in
the current study. Community samples are much less likely to have the
sophistication of Asian student samples who are primarily graduate stu-
dents. Third, most of the previous research studies were conducted within
one ethnic group sample (Chou et al., 1996; Chou & Mak, 1998; Hatfield et
al., 1996; Raguram et al., 1996), while students from various Asian countries
participated in the present study. The findings in the present study suggest
that there is considerable variability among Asian students in negative
beliefs toward mental illness. Fourth, the current study did not use actual
treatment-seeking behavior as the criterion, but administered self-report
questionnaires to evaluate treatment-seeking preferences. It is less clear
that the prediction demonstrated in the current study reflects participants’
actual treatment-seeking behavior.

REFERENCES

Atkinson, D. R., Lowe, S., & Matthews, L. (1995). Asian-American acculturation, gender,
and willingness to seek counseling. Journal of Multicultural Counseling and Development,
23(3), 130–138.
Chou, K. L., & Mak, K. Y. (1998). Attitudes to mental patients among Hong Kong Chinese:
A trend study over two years. International Journal of Social Psychiatry, 44(3), 215–224.
Chou, K. L., Mak, K. Y., Chung, P. K., & Ho, K. (1996). Attitudes towards mental patients
in Hong Kong. International Journal of Social Psychiatry, 42(3), 231–219.
Enrique, G. A. Jr. (1993). Psychiatric care of Filipino Americans. In Gaw, A. C (Ed.) Culture,
ethnicity, and mental illness. Washington, DC: American Psychiatry Press, Inc.
Fabrega, H. Jr. (1991). Psychiatric stigma in non-Western societies. Comprehensive Psychiatry,
32, 534–551.
Fan, C. (1999). A comparison of attitudes towards mental illness and knowledge of mental
health services between Australian and Asian students. Community Mental Health Jour-
nal, 35(1), 47–56.
Fujii, J. S., Fukushima, S. N., & Yamamoto, J. (1993). Psychiatric care of Japanese Americans.
In Gaw, A. C (Ed.) Culture, ethnicity, and mental illness. Washington, DC: American
Psychiatry Press, Inc.
Gaw, A. C. (1993). Psychiatric care of Chinese Americans. In Gaw, A. C (Ed.) Culture,
ethnicity, and mental illness. Washington, DC: American Psychiatry Press, Inc.
Hatfield, B., Mohamad, H., Rahim, Z., & Tanweer, H. (1996). Mental health and the Asian
Communities: A local survey. British Journal of Social Work, 26, 315–336.
Johnson, S., & Orrell, M. (1995). Insight and psychosis: a social perspective. Psychological
Medicine, 25, 515–520.
Kim, L. I. C. (1993). Psychiatric care of Korean Americans. In Gaw, A. C (Ed.), Culture,
ethnicity, and mental illness. Washington, DC: American Psychiatry Press, Inc.
Lam, A. P., & Kavanagh, D. J. (1996). Help seeking by immigrant Indochinese psychiatric
patients in Sydney, Australia. Psychiatric Services , 47(9), 993–995.
Leong, F. T. L., Wagner, N. S., & Kim, H. H. (1995). Group counseling exceptions among
236 Hirai and Clum

Asian American students: the role of culture-specific factors. Journal of Counseling

Psychology, 42(2), 217–222.
Matsuoka, J. K., Breaux, C., & Ryujin, D. H. (1997). National utilization of mental health
services by Asian Americans/Pacific Islanders. Journal of Community Psychology,
25(2), 141–145.
Ng, C. H. (1997). The stigma of mental illness in the Asian cultures. Australian and New
Zealand Journal of Psychiatry, 31, 382–390.
Raguram, R., Weiss, M. G., Channabasavanna, S. M., & Devins, G. M. (1996). Stigma,
depression and somatization in South India. American Journal of Psychiatry, 153, 1043–
1049.
Suinn, R. M., Ahuna, C., & Khoo, G. (1992). The Suinn-Lew Asian Self-Identity Acculturation
Scale: concurrent and factorial validation. Educational and Psychological Measurement,
52, 1042–1046.
Suinn, R. M., Figueroa, K. R., Lew, S., & Vigal, P. (1987). The Suinn-Lew Asian Self-Identity
Acculturation Scale: an initial report. Educational and Psychological Measurement, 47,
401–407.
Whaley, A. L. (1997). Ethnic and racial differences in perceptions of dangerousness of persons
with mental illness. Psychiatric Services, 48(10), 1328–1330.
Wintersteen, R. T., Wintersteen, L. B., Mupedziswa, R., & Cheah, A. (1997). Families and
mental illness: Observations from two developing countries. International Social Work,
40(2), 191–207.
Zhang, A. Y., Snowden, L. R., & Sue, S. (1998). Differences between Asian and white
Americans’ help seeking and utilization patterns in the Los Angeles Area. Journal of
Community Psychology, 26(4), 317–326.