Personality and Individual Dierences 28 (2000) 299±307
www.elsevier.com/locate/paid
Assessing Chinese adolescents' social support: the
multidimensional scale of perceived social support
Kee-Lee Chou
Department of Social Work and Social Administration, The University of Hong Kong, Pokfulam Road, Hong Kong
Received 25 September 1998; received in revised form 24 February 1999; accepted 18 March 1999
Abstract
The psychometric and factor-analytic properties of the Chinese version of the Multidimensional Scale
of Perceived Social Support (MSPSS-C) were investigated in a Hong Kong Chinese adolescent sample.
Two factors were extracted from this sample and they are Friend and Family subscales. The reliability,
concurrent validity, and construct validity of the MSPSS-C were also established. The MSPSS-C
demonstrated excellent internal consistency including the two subscales. In terms of construct validity,
the MSPSS-C correlated negatively with depression and anxiety assessed by the General Health
Questionnaire. The concurrent validity of the MSPSS-C was demonstrated by the positive association
between the MSPSS-C and the Lubben Social Network Scale. # 1999 Elsevier Science Ltd. All rights
reserved.
Keywords: Adolescents; Social support; Scale validation; Chinese
1. Introduction
In recent years, eorts have been directed toward understanding the role of social support,
as a coping resource in the relation between stress and psychological or physical symptoms
(Andrews, Tennant, Hewson & Vaillant, 1978; Lin, Simeone, Ensel & Kuo, 1979; Monroe,
Imho, Wise & Harris, 1983; Sarason, Sarason, Potter & Antoni, 1985; Wilcox, 1981).
However, some researchers argue that many studies have been conducted without including
adequate measures of social support (Vaux, Riedel & Stewart, 1987). Social support is a multi-
dimensional construct, and many researchers agree that indicators of social support include
size of social network, frequency of contact with members in the social network, instrumental
support, emotional support, quality of social support, and reciprocal helping of others
(Barrera, 1986; George, 1989; Tardy, 1988).
0191-8869/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 9 1 - 8 8 6 9 ( 9 9 ) 0 0 0 9 8 - 7
300 K.-L. Chou / Personality and Individual Dierences 28 (2000) 299±307
Recently, there has been growing interest in the assessment of social support; (for reviews of
these measures, see Bruhn & Philips, 1984; House & Kahn, 1985; Tardy, 1988). One
instrument, the Multidimensional Scale of Perceived Social Support (MSPSS) (Zimet, Dahlem,
Zimet & Farley, 1988), which is a self-administered measure of social support, would seem to
be useful for study in a Chinese society because it has a number of advantages. First of all, the
MSPSS speci®cally addresses the subjective assessment of social support adequacy. The
perceived adequacy of social support or satisfaction with the amount or quality of social
support appears to have a strong eect on depressive symptoms (Dean, Kolody & Wood,
1990; Vaux & Harrison, 1985); therefore it is an important component of social support.
Secondly, the MSPSS was designed to assess perceptions of social support adequacy from three
dierent sources: family, friends, and signi®cant others. Thirdly, the MSPSS contains only 12
items so it is simple to use and can be quickly administered and scored. As a result, the
MSPSS is an ideal measuring instrument for use when time is limited.
The MSPSS measures perceived social support from dierent sources, namely, family
members, friends, and signi®cant others and three factor scores have been described for this
instrument for these three sources of social support. The MSPSS scale has been found to have
psychometric stability, high internal and test-retest reliability, concurrent validity, construct
validity and discriminant validity (Kazarian & McCabe, 1991; Zimet et al., 1988). Using this
scale, Zimet et al. (1988) found that the factor scores of the MSPSS were negatively related to
depression and anxiety symptomatology as measured by the Hopkins Symptom Checklist in
university undergraduates. The factor scores of the MSPSS were also correlated with the Social
Support Behaviors scale and the self-concept scale (Kazarian & McCabe, 1991).
With the growing interest in the social support of dierent cultural groups, a need has
emerged for translated versions of scales measuring social support. Although some studies have
shown that subscales derived from factor analysis of the MSPSS have been regarded as
relatively stable across situations and populations (Eker & Arkar, 1995; Kazarian & McCabe,
1991; Zimet et al., 1988), there were still few studies that speci®cally addressed the
reproducibility of these scales in a dierent cultural setting. To the best of our knowledge, the
reliability and validity of a Chinese version of MSPSS (MSPSS-C) has never been extensively
analyzed. Thus, this study aimed to examine the psychometric properties of the MSPSS-C
amongst a Chinese population. Speci®cally, reliability (internal consistency), factorial structure,
concurrent validity and construct validity of the MSPSS-C were examined in the current study.
2. Method
2.1. Participants
The participants in this study were 475 12th grade high school students who participated in
the survey in the 1996±97 school year. Nine schools were randomly selected to participate and
one class of 12th grade in each school was randomly selected. Approximately 50 students were
selected in each school. The school response rate was 75% (9 schools participated of 12
sampled) and the participation rate for students was approximately 89%. Out of 500 students
who were administered the questionnaires, 12 had excessive missing data (more than 20%
K.-L. Chou / Personality and Individual Dierences 28 (2000) 299±307 301
missing responses), and 13 appeared to have responded randomly to the questionnaire (for
criteria, see Farrell, Danish & Howard, 1991). Excluding those who met one or more of these
criteria resulted in a sample size of 475. Among the respondents, 410 (86%) were male and 65
(14 %) were female. Approximately ninety percent (88.5%) came from the 17±18 age group
and the mean of their age was 17.5 (SD=0.7).
2.2. Procedures
Using written instruments, research sta administered the survey to participating students in
the selected classes. The questionnaire took approximately 20 min to administer. Active
personal consent and passive parental consent were sought for all respondents to participate in
this study. Letters describing the study were written to the parents and the students, signed
consent forms were obtained from students and parents could sign `refusal to participate'
forms if they did not wish their child to participate in the study. To increase the validity of the
responses, we made eorts to protect students' anonymity; students answered the survey and
returned the completed questionnaires in sealed envelopes.
2.3. Measures
The MSPSS (Zimet et al., 1988) is a 12-item scale with a seven point scale (from 1=strongly
disagree to 7=strongly agree) measuring three sources of support, namely, Family, Friends,
and Signi®cant Other. Zimet et al. (1988) have reported excellent psychometric properties,
especially considering the number of items in the scale. In terms of internal reliability, an alpha
coecient of 0.88 for the total scale has been reported. Subscale reliability estimates are quite
high, with reported alphas of 0.81, 0.85, and 0.91 for Family, Friend, and Signi®cant Other
subscales, respectively. Additionally, test±retest reliability of 0.85 has also been obtained.
Finally, construct validity has been established by an inverse correlation with depression scores
(r=ÿ0.25, P < 0.01).
The researcher translated the 12-item version of the MSPSS into Chinese from the English
version (Zimet et al., 1988) and this was backtranslated by a bilingual professional translator.
Because the ®rst backtranslation of the MSPSS items indicated the need for minor revisions,
some items were modi®ed further. All items were then backtranslated by another professional
translator. This translation and backtranslation process was repeated before the ®nal form was
established.
The General Health Questionnaire (GHQ-30) was administered to assess psychological
symptoms in the previous 3 weeks. The GHQ-30 has thirty items, with a four-point scale (from
1 to 4) used to assess the psychological distress symptoms of an individual. It was originally
designed by Goldberg (1972) to enable family doctors to detect current non-psychotic
disturbances in general practice patients. The GHQ-30 has been translated into Chinese, and
has been extensively validated in various studies of students (Chan & Chan, 1983; Chan, 1993;
Shek, 1987). The internal consistency measures (Cronbach's alpha) of the Chinese version of
the GHQ ranged from 0.85 to 0.88 (Chan & Chan, 1983; Shek, 1987). Scores on this 30-item
scale of feelings and behaviors related to mental health ranged from 30 to 120 with high scores
indicating higher levels of distress. In this sample, the GHQ-30 yields scores on four primary
302 K.-L. Chou / Personality and Individual Dierences 28 (2000) 299±307
symptom dimensions including anxiety (a=0.86), depression (a=0.76), social dysfunction
(a=0.71) and inadequate coping (a=0.75) (Shek, 1987). Only anxiety and depression subscales
were used in the current study because previous studies have found negative association
between these two scales and social support (Shek, 1987).
The social support network was assessed by the Lubben Social Network Scale (LSNS;
Lubben, 1988) which is a 10-item scale measuring ®ve aspects of social networks: family
network, friend networks, helping others, con®dant relationships, and living arrangements.
LSNS assesses two aspects of social network and they are the size of social networks and
frequency of contact with members in the social network. Cronbach's alpha in our sample for
this measure was 0.72. The Chinese version of the Lubben Social Network Scale has been
widely used in Chinese samples (Chi, 1995).
3. Results
3.1. Factor analysis
The item responses of the 475 subjects to the 12-item 4-point MSPSS-C were recoded and
scored in the direction of high social support. The researcher conducted an exploratory factor
analysis of the MSPPS-C items following the procedures recommended by Ben-Porath, Waller,
and Butcher (1991) and DeVellis (1991). A principal component factor analysis of the MSPSS-
C also generated two dimensions with eigenvalues exceeding unity. To avoid underextraction,
each of the varimax-rotated factor solutions from 2- to 4-factor solutions were examined for
simple structure and interpretability. Among the factor solutions, the varimax-rotated two-
factor solution was most interpretable. Solutions with more than two factors would include
inconsequential factors that were invariably de®ned by only one or two salient items and were
Table 1
Summary of varimax-rotated two-factor solution of the MSPSS (N = 475) (P < 0.01; P < 0.05)
Item Item-to-factor r Loading
Factor 1 (Friend factor)
1. There is a special person who is around when I am in need 0.80 0.79
2. There is a special person with whom I can share my joys and sorrows 0.86 0.86
3. I have a special person who is a real source of comfort to me 0.83 0.84
4. My friends really try to help me 0.86 0.83
5. I can count on my friends when things go wrong 0.85 0.83
6. I have friends with whom I can share my joys and sorrows 0.84 0.84
7. There is a special person in my life who cares about my feelings 0.84 0.85
8. I can talk about my problems with my friends 0.85 0.79
Factor 2 (Family factor)
9. My family really tries to help me 0.85 0.81
10. I get the emotional help and support I need from my family 0.86 0.87
11. I can talk about my problems with my family 0.83 0.85
12. My family is willing to help me make decisions 0.83 0.82
K.-L. Chou / Personality and Individual Dierences 28 (2000) 299±307 303
not readily interpretable. In addition, the incremental magnitude of the Tucker±Lewis
reliability coecient was relatively low beyond the two-factor solution. Therefore, the varimax-
rotated two-factor solution (Friend and Family Factor) as shown in Table 1 was chosen for
interpretation. Factor 1 (eigenvalue=5.86) was de®ned by a relatively heterogeneous array of
items relating to support from friends and signi®cant others and it accounted for 48.8% of the
total variance. Factor 2 (eigenvalue=2.58) was loaded by salient items re¯ecting social support
from family members and accounted for 21.5 of the total variance. In short, the factor
structure of the MSPSS-C is quite similar to that of the MSPSS except that the Friend and
Signi®cant Others Factors of the original MSPSS were combined in the MSPSS-C.
3.2. Descriptive statistics
Table 2 shows the distribution of responses, and the means and standard deviations of each
item for the total sample. As regards the intercorrelations among subscales, the Friend Factor
was found to be moderately correlated with Family Factor (r = 0.23; P < 0.01).
3.3. Reliability
The internal consistency coecient (Cronbach's alpha) for the MSPSS-C scale was 0.89 in
the current sample. No item could be excluded because removal of any item would have a
negative eect on alpha. The Cronbach's coecient alpha was obtained for Friend and Family
subscales which were 0.94 and 0.86, respectively. Results have shown that the reliabilities were
quite good for the MSPSS-C.
3.4. Construct and concurrent validity
One of the hypotheses underlying the development of this instrument was that perceived
Table 2
Percentage distribution of responses, mean, and standard deviation for all items of the MSPSS (N = 475)
1=strongly disagree 2 3 4 5 6 7=strongly agree Mean SD
Item
1. 4.1% 8.7% 14.3% 22.3% 17.8% 17.1% 15.6% 4.55 1.66
2. 2.2% 8.0% 9.9% 17.9% 16.8% 19.8% 25.4% 5.00 1.67
3. 3.0% 5.6% 12.0% 18.5% 18.3% 21.1% 21.5% 4.78 1.63
4. 7.7% 12.3% 17.2% 22.2% 18.1% 12.9% 9.7% 4.86 1.47
5. 2.4% 8.2% 12.9% 18.5% 21.3% 17.2% 19.6% 4.52 1.58
6. 2.5% 4.7% 10.5% 21.4% 24.6% 20.1% 16.3% 5.05 1.55
7. 4.3% 7.8% 13.2% 21.4% 24.7% 16.7% 11.9% 5.03 1.64
8. 9.7% 14.6% 16.8% 19.1% 14.6% 15.7% 9.5% 5.16 1.50
9. 1.1% 7.4% 8.9% 16.7% 22.3% 21.5% 22.1% 4.93 1.63
10. 2.2% 8.0% 9.5% 15.5% 17.7% 24.1% 23.1% 4.08 1.69
11. 5.2% 8.0% 13.5% 22.8% 17.8% 18.9% 13.8% 3.99 1.79
12. 1.5% 3.9% 10.1% 17.2% 18.8% 26.1% 22.4% 4.52 1.66
304 K.-L. Chou / Personality and Individual Dierences 28 (2000) 299±307
social support would be negatively related to reported anxiety and depression symptoms.
Support for this prediction was demonstrated by correlations between MSPSS subscales and
the Depression and Anxiety subscales of the GHQ. As can be seen in Table 3, perceived
support from Family was signi®cantly inversely related to depression and marginally related to
anxiety. Moreover, perceived support from Friend was inversely associated to anxiety but
marginally to depression. Respondents were categorized as the most anxious (mean scores=3
to 4) and the least anxious (mean scores=1 to 2) and it was found that there were no
signi®cant dierences in perceived support from Friend and Family. Respondents were
categorized as the most depressed (mean=3 to 4) and the least depressed (mean=1 to 2) and it
was found that there were no signi®cant dierences in perceived support from Friend and
Family.
The scores obtained from both the MSPSS-C and the Lubben Social Network Scale were
correlated to assess the concurrent validity of the MSPSS-C. Table 3 shows that both Family
and Friend subscales were positively associated with the Lubben Social Network Scale. It
should be noted that the Lubben Social Network Scale correlated more highly with the Family
subscale of the MSPSS-C than the Friend subscale of the MSPSS-C.
3.5. Analyses of gender and age dierences
To examine the eects of gender and age group, multivariate analysis of variance
(MANOVA) was performed on the set of two MSPSS-C subscales. The results indicated that
there was no signi®cant eect for an overall Gender Age Group Interaction, overall gender
eect, and overall age eect.
4. Discussion
It is also important to evaluate assessment of social support in dierent cultures. It is agreed
that this translated version of the MSPSS-C validly and reliably assesses perceived quality of
social support amongst Hong Kong Chinese. Content validity was examined by comparing the
backtranslated and original items of the MSPSS. Backtranslation seems to satisfactorily
replicate MSPSS items, suggesting that these items are translated into Chinese with some
degree of con®dence.
Table 3
Intercorrelations of the Family and Friend subscales of the MSPSS-C, depression and anxiety subscales of the
GHQ, and the Lubben Social Network Scale (P < 0.01; P < 0.05)
1. 2. 3. 4.
1. Family subscale of the MSPSS
2. Friend/signi®cant others subscale of the MSPSS 0.23
3. Anxiety subscale of the GHQ ÿ0.11 ÿ0.14
4. Depression subscale of the GHQ ÿ0.16 ÿ0.12 0.51
5. The Lubben scale 0.41 0.25 ÿ0.10 ÿ0.17
K.-L. Chou / Personality and Individual Dierences 28 (2000) 299±307 305
Results of the factor analyses did not provide evidence for the construct validity of the three
social support subscales for this sample of Chinese respondents. Rather, two factors emerged,
re¯ecting two relatively independent dimensions of social support. In this study, it seems that
the Family subscale of the original three-factor solution is retained, whereas the original Friend
and Signi®cant Others subscales are combined into one Friend subscale in this sample. The
question arises as to why two factors emerged in the present analysis, instead of three factors
found in previous studies conducted in other cultures. One could speculate that translation,
method or cultural dierences, or some combination of these factors, may account for the
discrepancy. The other reason is that the two-factor solution may be due to the sample we
used in the current study. In this study, respondents are Chinese adolescents in Hong Kong
high schools. The samples were university undergraduate students in the original MSPSS
validation study. In our sample, the signi®cant others of adolescents are more likely to be their
peers. Therefore, items related to support from friends and signi®cant others are highly
associated. However, in the sample of the original MSPSS validation study, those respondents
might engage in romantic relationships. Therefore, their close friends in romantic relationships
would most likely be their signi®cant others.
The factor analysis demonstrates that the youths clearly dierentiated between two sources
of perceived social support from peers and family members. This ®nding con®rms and extends
Procidano and Heller's (1983) demonstration of friend and family as independent and
internally consistent sources of social support. In addition, it suggests that studies which fail to
consider the source of support may lose important information. In this sample of high school
students, perceived social support from family was more strongly inversely related to
depression than was perceived support from friends whereas perceived social support from
friends was more strongly negatively associated to anxiety than was perceived support from the
family.
Results also support the construct validity of the MSPSS-C. The pattern of a negative
association between perceived social support and psychological distress (depression and
anxiety) found in the present study clearly demonstrates the construct validity of the MSPSS-
C. Consistent with these ®ndings, Zimet et al. (1988) found a negative relationship between
social support and depression. Finally, our results are unequivocal in demonstrating the
concurrent validity of this translated version. The MSPSS-C is positively associated to size of
social network and frequency of contacts among members in their social network. These
®ndings are consistent with a previous study by Kazarian and McCabe (1991).
Although evidence indicates that the MSPSS-C is a sound research instrument, there are two
limitations of the current study. Firstly, the samples used in the current study are adolescents.
It would be important, therefore, to investigate the psychometric properties and factorial
structure of the MSPSS-C with participants drawn from the adult populations. Secondly, it is
important to note that the causal direction of social support and depression cannot be de®ned
in this correlational study and a longitudinal study should be conducted to clarify the causal
nature of this relationship.
To my knowledge, this study is the ®rst to provide an extensive examination of the reliability
and validity of the MSPSS for an Asian population. The results of the present study indicate
that the MSPSS-C displays adequate psychometric properties in terms of internal consistency,
content validity, concurrent validity, and construct validity, at least for Hong Kong Chinese
306 K.-L. Chou / Personality and Individual Dierences 28 (2000) 299±307
young people. The availability of the MSPSS as a reliable, valid, and easy-to-administer scale
should facilitate the examination of the causal mechanism involved in the link between
perceived social support and mental health among Chinese adolescents.
References
Andrews, G., Tennant, C., Hewson, D. M., & Vaillant, G. E. (1978). Life event stress, social support, coping style,
and risk of psychological impairment. Journal of Nervous and Mental Disease, 166, 307±316.
Barrera, M. (1986). Distinctions between social support concepts, measures, and models. American Journal of
Community Psychology, 14, 413±445.
Ben-Porath, Y. S., Waller, N. G., & Butcher, J. N. (1991). Assessment of coping: an empirical illustration of the
problem of inapplicable items. Journal of Personality Assessment, 57, 162±176.
Bruhn, J. G., & Philips, B. U. (1984). Measuring social support: a synthesis of current approaches. Journal of
Behavioral Medicine, 7, 151±169.
Chan, D. W. (1993). The Chinese General Health Questionnaire in a psychiatric setting: the development of the
Chinese scaled version. Social Psychiatry and Psychiatric Epidemiology, 28, 124±129.
Chan, D. W., & Chan, T. S. C. (1983). Reliability, validity and the structure of the General Health Questionnaire in
a Chinese context. Psychological Medicine, 13, 363±371.
Chi, I. (1995). Mental health of the old-old in Hong Kong. Clinical Gerontologist, 15, 31±44.
Dean, A., Kolody, B., & Wood, P. (1990). Eects of social support from various sources on depressive in elderly
persons. Journal of Health and Social Behavior, 31, 148±161.
DeVellis, R. F. (1991). Scale development: theory and application. London: Sage.
Eker, D., & Arkar, H. (1995). Perceived social support: psychometric properties of the MSPSS in normal and patho-
logical groups in a developing country. Social Psychiatry and Psychiatric Epidemiology, 30, 121±126.
Farrell, A. D., Danish, S. J., & Howard, C. W. (1991). Evaluation of data screening methods in surveys of adoles-
cents' drug use. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 3, 295±298.
George, L. (1989). Social and economic factors. In E. W. Busse, & D. G. Blazer, Geriatric Psychiatry (pp. 203±234).
Washington, DC: American Psychiatric Press.
Goldberg, D. P. (1972). The detection of psychiatric illness by questionnaire. London: Oxford University Press.
House, J. S., & Kahn, R. L. (1985). Measures and concepts of social support. In S. Cohen, & S. L. Syme, Social
support and health (pp. 83±108). Orlando, FL: Academic.
Kazarian, S. S., & McCabe, S. B. (1991). Dimensions of social support in the MSPSS: factorial structure, reliability,
and theoretical implications. Journal of Community Psychology, 19, 150±160.
Lin, N., Simeone, R. S., Ensel, W. M., & Kuo, W. (1979). Social support, stressful life events and illness: a model
and an empirical test. Journal of Health and Social Behavior, 20, 108±119.
Lubben, J. E. (1988). Assessing social networks among elderly populations. Journal of Family and Community
Health, 11, 42±52.
Monroe, S. M., Imho, D. F., Wise, B. D., & Harris, J. E. (1983). Prediction of psychological symptoms under
high-risk psychosocial circumstances: life events, social support, and symptom speci®city. Journal of Abnormal
Psychology, 92, 338±350.
Procidano, M. E., & Heller, K. (1983). Measures of perceived social support from friends and from family: three
validation studies. American Journal of Community Psychology, 11, 1±24.
Sarason, I. G., Sarason, B. R., Potter, E. H., & Antoni, M. H. (1985). Life events, social support, and illness.
Psychosomatic Medicine, 47, 156±163.
Shek, D. T. L. (1987). Reliability and factorial structure of the Chinese version of the General Health
Questionnaire. Journal of Clinical Psychology, 57, 683±691.
Tardy, C. H. (1988). Social support measurement. American Journal of Community Psychology, 13, 187±202.
Vaux, A., & Harrison, D. (1985). Support network characteristics associated with support satisfaction and perceived
support. American Journal of Community Psychology, 13, 245±268.
K.-L. Chou / Personality and Individual Dierences 28 (2000) 299±307 307
Vaux, A., Riedel, S., & Stewart, D. (1987). Modes of social support: the Social Support Behaviors (SS-B) Scale.
American Journal of Community Psychology, 15, 209±237.
Wilcox, B. L. (1981). Social support, life stress, and psychological adjustment: a test of the buering hypothesis.
American Journal of Community Psychology, 9, 371±386.
Zimet, G. D., Dahlem, N. W., Zimet, S. G., & Farley, G. K. (1988). The multidimensional scale of perceived social
support. Journal of Personality Assessment, 52, 30±41.
