The subjective well-being (SWB) has been an important area of research in positive psychology, which reflects individuals’
affective and cognitive evaluations of their lives. SWB contains a cognitive and an affective component. The affective component
of SWB refers to the positive affect and negative affect which their levels are used to indicate the level of SWB. The cognitive
component of SWB refers to life satisfaction that is a subjective evaluation of overall quality of life (Diener and Diener
1995). Individuals will report high life satisfaction if their perceived life circumstances are in line with their own standard.
According to Suldo and Huebner (2006), individuals with high levels of SWB, have higher life satisfaction and experience more positive than negative affect.
As a crucial indicator of SWB, life satisfaction has been thought to play a key role in the lives of adolescents and emerging
adults (Greenspoon and Saklofske 2001; Gilman and Huebner 2003). Numerous studies have shown that personality variables including social isolation, self-esteem, optimism, and other positive
emotional experiences (e.g., Diener 1996; Diener et al. 1999; Emmons and Diener 1985) and demographic variables e.g., age, gender and race (e.g., Diener 1984; Huebner et al. 1998, 2004) could significantly predict one’s level of satisfaction with life. Accordingly, individuals with high life satisfaction
reported less social stress and more positive relationships with others (Gilman and Huebner 2006) and received more social support from their friends and family (Suldo and Huebner 2006). Individuals with high life satisfaction reported higher emotional and social self-efficacy than those with low satisfaction
((Pinquart et al. 2004, Suldo and Huebner 2006), thus they have less anxiety and depression and high levels of hope and personal control (Gilman and Huebner 2006) and have fewer behavioral and emotional problems (Suldo and Huebner 2006). Therefore, life satisfaction is an important psychological variable in the adolescence period and an important indicator
of individuals’ psychological development.
There is a long history in psychological research which has examined the relationships between social support and life satisfaction.
Numerous studies have provided evidence for the positive relationship between social support and life satisfaction, even if
researchers control covariates such as personality (e.g., Argyle and Lu 1990; Cooper et al. 1992; Diener and Seligman 2002; Skok et al. 2006). In addition to such direct associations, a series of studies have further identified mediators to explain underlying mechanisms
of the link between social support and life satisfaction. Mediators refer to variables that establish how or why one variable
predicts or causes an outcome variable (Frazier et al. 2004). In other words, a mediator is the mechanism through which a predictor influences an outcome variable (Baron and Kenny 1986), providing useful information regarding psychological intervention for practicing psychologists.
The previous literature has corroborated self-esteem and loneliness as two promising mediators between social support and
life satisfaction (Yarcheski et al. 2001; Stroebe et al. 1996). Self-esteem is hypothesized as an important mediator because individuals with extremely low social support would rarely
be satisfied with themselves and more likely result in low self-esteem, which would significantly exacerbate their psychological
well-being. From the studies by Yarcheski and his associates (Yarcheski et al. 2001), self-esteem was found to fully mediate the link between social support and life satisfaction. Loneliness is also hypothesized
as a mediator based on the dual-path model of attachment theory (Stroebe et al. 1996). The model holds that marital status and social support influence well-being by different pathways, with the impact of marital
status mediated by emotional loneliness and the impact of social support mediated by social loneliness in adjustment to loss,
e.g., bereavement. In addition, Civitci and Civitci (2009) found that self-esteem could partially mediate the relationship between loneliness and life satisfaction in adolescents.
Although previous research on mediation effects of self-esteem and loneliness have a clear understanding of underlying mechanisms
to elucidate the relationship between social support and life satisfaction, further investigation can be explored. Firstly,
testing the concurrent mediating effects of self-esteem and loneliness, which has been examined separately in the previous
studies, would extend our insight into the mechanism the link between social support and life satisfaction. Secondly, only
depressive symptomatology and somatic complaints were used as criterion variables in the studies testing the mediating role
of loneliness, therefore, life satisfaction was utilized as an outcome variable in this study, providing meaningful information.
Finally, an important limitation in the previous literature is that the majority of the studies were conducted in western
countries. Researchers have agreed that it is necessary to examine the generalizability of the findings to other cultures,
such as Asian cultures, especially Chinese culture (e.g., Li and Liang 2007; Zhang and Leung 2002). Compared to collectivistic cultures (e.g., Chinese culture), individualistic cultures place more emphasis on self-esteem,
so this correlation between self-esteem and global life satisfaction is stronger among western countries. In collectivistic
cultures, social relationships are often included in the self-concept, so social relationship harmony is more strongly related
to global life satisfaction in these cultures (Tam et al. 2010). In summary, the aim of this study is to test the mediation effects of both self-esteem and loneliness between social support
and life satisfaction in Chinese college students, providing meaningful evidence for the external validity.2 Method2.1 Participants and Procedure
Data from 389 undergraduates from two universities in China were used in the current study. The participants consisted of
260 women and 129 men. The students’ ages ranged in age from 17 to 25, with an average of 20.39 (SD = 1.84).
Participants completed the questionnaires in the classroom environment. Informed consent was obtained by all participants
before completing the measures. Students did not place their names on the measures and were assured of the confidentiality
of their responses. Instruments took approximately 20 min to complete.2.2 Instruments
2.2.1 Satisfaction with Life Scale
The Satisfaction with Life Scale (SWLS) measures a person’s subjective evaluation of global life. The SWLS developed by Diener
et al. (1985) consists of five items. Each item is answered on a 7-point Likert type scale ranging from 1 = strongly disagree to 7 = strongly
agree. It includes items such as, “I am satisfied with my life” and “In most ways my life is close to my ideal”. The SWLS
has good psychometric properties (Pavot and Diener 1993). In this study, the Cronbach alpha coefficient for the SWLS was 0.78.
2.2.2 Social and Emotional Loneliness Scale
The Social and Emotional Loneliness Scale (SELS) was developed by Wittenberg (cited in Shaver and Brennan, 1991). It consists of 10 items designed to assess social loneliness (SL) and emotional loneliness (EL). Responses to each item
are given on a five-point Likert-type scale ranging from 1 (never) to 5 (very often). It includes items such as, “1 have a
really nice set of friends”; “I have friends and acquaintances with who I like to be together”; “I feel lonely even when I
am with other people”. Scale scores are the sum of items with reverse coding of relevant items. In this study, the Cronbach
alpha coefficient for the SELS was 0.67.
2.2.3 Chinese Social Support Rating Scale
The Chinese Social Support Rating Scale (CSSRS) was developed by Xiao (cited in Wang et al. 1999), based on the unique environmental and cultural conditions in China. The scale consists of 10 items designed to assess subjective
support, objective support and utilization of support. The total score for the 10 items was used as a measure of current total
social support status. The scale has been widely used in Chinese populations (Tan et al. 2004) and proved to have good validity and reliability (Wang et al. 1999). In this study, the Cronbach alpha coefficient for the CSSRS was 0.71.
2.2.4 Rosenberg Self-Esteem Scale
The Rosenberg Self-Esteem Scale (RSES), developed by Rosenberg (1965), is a 10-item self-report measure of global self-esteem. Items are rated from 1 (strongly disagree) to 4 (strongly agree).
The scores can range from 10 (low level of self-esteem) to 40 (high level of self-esteem). Examples of items include: “I am
able to do things as well as most other people.”; “I take a positive attitude toward myself”. Scale scores are the sum of
items with reverse coding of relevant items. The psychometric properties of the RSES have been reported (Corwyn 2000). In this study, the Cronbach alpha coefficient for the RSES was 0.83.3 Results3.1 Analysis Strategy
The analysis of the mediation effects adopted the two-step procedure (Anderson and Gerbing 1988). The measurement model first was tested to assess the extent to which each of the latent variables was represented by its
indicators. If the confirmatory measurement model was accepted, test the structural model using the maximum likelihood method
in LISREL 8.7 (Jöreskog and Sörbom 1996). In order to control inflated measurement errors due to multiple items for the latent variable, three item parcels were
created for the self-esteem factor. The following four indices were utilized to evaluate the goodness of fit of the model
(see Hu and Bentler 1999; Quintana and Maxwell 1999): (a) chi-quare statistics; (b) root-mean-square error of approximation (RMSEA) of 0.08 or less; (c) standardized root-mean-square
residual (SRMR) of 0.08 or less; and (d) comparative fit index (CFI) of 0.95 or more.
To compare two or more models, Akaike Information Criterion (AIC: Akaike 1987) of smaller values representing a better fit of the hypothesized model (Byrne 2001) and Expected Cross-Validation Index (ECVI: Browne and Cudeck 1993) of smallest values exhibiting the greatest potential for replication (Byrne 2001) were additionally examined to determine the best model.3.2 Measurement ModelThe measurement model included four latent factors (social support, loneliness, self-esteem and life satisfaction) and 13
observed variables. An initial test of the measurement model revealed a very satisfactory fit to the data: χ2 (59, N = 389) = 95.28, p < 0.001; RMSEA = 0.040; SRMR = 0.041; and CFI = 0.98. All the factor loadings for the indicators on the latent variables
were significant (p < 0.001), indicating that all the latent factors were well represented by their respective indicators. In addition, all the
latent factors from the measurement model were significantly correlated (Table 1).Table 1 Inter-correlations between four latent variables
1N = 389. All correlation coefficients are significant at p < 0.0013.3 Structural Model
In order to find the best model, we assessed five alternative models (Models 1–5). First, a partially-mediated model (Model
1) with two mediators and a direct path from social support to life satisfaction showed a good fit to the data: χ
(60, N = 389) = 115.94, p < 0.001; RMSEA = 0.049; SRMR = 0.047; and CFI = 0.97. However, the standardized path coefficient from social support to life
satisfaction became non-significant (β = –0.08). Thus, a fully-mediated model (Model 2) was tested subsequently with this path constrained to zero, which revealed
a good fit to the data: χ
(61, N = 389) = 116.63, p < 0.001; RMSEA = 0.0048; SRMR = 0.047; and CFI = 0.97. In term of the fit indices, Model 2 was obviously better than Model
1.Next, a path from loneliness to self-esteem was added to the fully-mediated model (Model 3) and the results showed a very
good fit to the data (Table 2). The significant chi-square difference When comparing Model 2 to Model 3, Δχ
(1, N = 389) = 20.94, p < 0.001, indicated that this additional path significantly contributed to the model. The standardized path coefficient from
loneliness to self-esteem was statistically significant (β = −0.52), however, the path from social support to self-esteem became non-significant (β = −0.23). Thus, this path was removed and the model was retested (Model 4). The results also revealed a very good fit to
the data (Table 2). Although there was no great difference between Model 3 and Model 4 according to the fit indices, the parsimony of Model
4 suggested that the fit of Model 4 were was more satisfactory (Table 2).
Table 2 Fit indices among competing models
CI for ECVI
0.35, 0.48N = 389. Boldface type represents the best modelRMSEA root-mean-square error of approximation, CI confidence interval, SRMR standardized root-mean-square residual, CFI comparative fit index, AIC Akaike information criterion, ECVI expected cross-validation indexIn addition, Model 5 was tested by adding a path from self-esteem to loneliness to Model 2, which showed a good fit to the
data (Table 2). Both the standardized path coefficients from self-esteem to loneliness (β = −0.39), from social support to loneliness (β = −0.60) and from social support to self-esteem (β = 0.28) were statistically significant. Although there was no great difference between Model 4 and Model 5 in terms of the
fit indices, a slightly smaller AIC value suggested that Model 5 was better than Model 4. Therefore, Model 5 was selected
as the best model (Fig. 1).Fig. 1 The finalized structural model (N = 389). Factor loadings are standardized. SS social support, SE self-esteem, LS life satisfaction. SL social loneliness, EL emotional loneliness, LS1–LS5 five parcels of life satisfaction, SE1–SE3 three parcels from self-esteem; S–SS subjective support, O–SS objective support, U–SS utilization of support. **p < 0.01; ***p < 0.001
In the final model, loneliness and self-esteem fully mediated the link between social support and life satisfaction (z = 5.09,
p < 0.001).The partial mediation effect of loneliness in the link between self-esteem and life satisfaction (z = 2.76, p < 0.001) was significant. Moreover, self-esteem partially mediated the link between social support and loneliness (z = −2.98,
p < 0.001). Notably, the path of social support → self-esteem → loneliness → life satisfaction was significant, indicating
that individuals with high social support are more likely to engage in self-esteem, which may lower their loneliness and result
in high life satisfaction.
Finally, we used multi-group analysis to explore whether the path coefficients were moderated by gender. The gender differences
were examined by comparing the first model, which allows the structural paths to vary across sexes, with the second model
which constrains the structural paths across sexes to be equal. All the factor loadings, error variances and structure covariances
were constrained to be equal. The following fit indices were generated by a SEM analysis. The non-significant chi-square differences
between the two models, Δχ
(14, N = 389) = 10.80, p > 0.05, as well as the slightly smaller AIC value suggested that the structural paths of the final model were not found to
differ by gender, lending preliminary support to its robustness.4 Discussion
The current study was designed to examine the mediation effects of both loneliness and self-esteem for the relationship between
social support and life satisfaction with a sample of Chinese college students. As expected, the best model from the current
study supports the mediation effect of self-esteem between social support and life satisfaction. This result is consistent
with earlier studies reporting to self-esteem as a mediator between social support and general well-being (e.g., Yarcheski
et al. 2001), and generates to Chinese college students. The previously reported mediating role of loneliness (Stroebe et al. 1996) is also robust with a different cultural sample (Chinese college students) and different indices of well-being (life satisfaction).
These results strongly suggest that social support influences well-being by two pathways, with the impact of social support
mediated by loneliness and the impact of social support mediated by self-esteem. In other words, individuals with low levels
of social support are likely to engage in loneliness and low self-esteem, which result in low life satisfaction.
Another finding of the study shows the path of social support → self-esteem → loneliness → life satisfaction is significant.
This path indicates that individuals with high social support are apt to engaging in high self-esteem, which may lower their
loneliness and in turn, lead to high life satisfaction. That is, loneliness is a mediator between self-esteem and life satisfaction
while self-esteem partially mediates the relationship between social support and loneliness. Here, the partial mediating role
of loneliness provides a new insight into the relationships among social support, self-esteem and loneliness. Thus, the mediating
role of self-esteem is inconsistent with recently reported self-esteem as a partial mediator in the relationship between loneliness
and life satisfaction in adolescents (Civitci and Civitci 2009). A possible cause of the inconsistent findings is that they did not explore whether self-esteem may predict loneliness in
the study by Civitci and Civitci (2009).
On the issue of whether loneliness may predict self-esteem or vice versa, the previous literature has reported conflicting
findings. For example, numerous studies on the relationship between self-esteem and loneliness have shown that self-esteem
significantly predicted loneliness (Mahon et al. 2006; McWhirter et al. 2002; Neto 2002). Similarly, Sedikides et al. (2004) found that self-esteem was a significant mediator between narcissism and psychological health (i.e., depression, loneliness,
subjective well-being, anxiety). On the contrary, Cacioppo et al. (2006) reported that loneliness significantly predicted self-esteem. In view of these conflicting findings, the present study reveals
a more complicated relationship among social support, loneliness, self-esteem, and life satisfaction. Notably, the current
study supported a fully mediating of self-esteem and a fully mediating of lonliness between social support and life satisfaction
and a significant path of social support → self-esteem → loneliness → life satisfaction. Thus, the results contribute to explain
the mixed findings in the literature regarding the relationships between loneliness and self-esteem.
Some limitations in the current study should be addressed. First, this study has a cross-sectional structure, so it is difficult
to draw causal relationships. Future researchers are advised to test the final model of using longitudinal and experimental
studies, providing a better understanding of these variables from a developmental perspective. Second, the data in this study
collected only through self-report scales could be a threat to internal validity. The use of multiple methods for evaluation
(e.g., parent, peer reports) may minimize the influence of subjectivity. Third, in addition to the gender differences in the
structural model that were tested in this study, other variables, such as age and race, should be examined. Fourth, SWB contains
two components, and testing the model reflecting various aspects of SWB is recommended. Finally, the study group was composed
of college students in Chinese culture, which limits the generalizability of the findings of the current study.
Despite its limitations, the current study considerably extended our insight into underlying mechanisms between social support
and life satisfaction in Chinese college students. The research findings provided meaningful evidence for external validity
for the loneliness and self-esteem mediated model in China. Moreover, the significant path from social support through self-esteem
and loneliness to life satisfaction further threw light on complex relationships among these variables. In consideration of
the probable mechanisms, it can be stated that self-esteem and loneliness improvement programs may have a preventive function
if implemented by counseling services to college students.
Akaike, H. (1987). Factor analysis and AIC. Psychometrika,
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach.
Argyle, M., & Lu, L. (1990). The happiness of extroverts. Personality and Individual Differences,
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual,
strategic and statistical considerations. Journal of Personality and Social Psychology,
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. Newbury Park, CA: Sage. Byrne, B. M. (2001). Structural equation modeling with AMOS: Basic concepts, applications, and programming. Mahwah, NJ: Lawrence Erlbaum Associates. Cacioppo, J. T., Hawkley, L. C., Ernst, J. M., Burleson, M., Berntson, G. G., Nouriani, B., et al. (2006). Loneliness within
a nomological net: An evolutionary perspective. Journal of Research in Personality,
Civitci, N., & Civitci, A. (2009). Self-esteem as mediator and moderator of the relationship between loneliness and life satisfaction
in adolescents. Personality and Individual Differences,
Cooper, H., Okamura, L., & Gurka, V. (1992). Social activity and subjective well-being. Personality and Individual Differences,
Corwyn, R. F. (2000). The factor structure of global self-esteem among adolescents and adults. Journal of Research in Personality,
Diener, E. (1984). Subjective well-being. Psychological Bulletin,
Diener, E. (1996). Traits can be powerful, but are not enough: Lessons from subjective well-being. Journal of Research in Personality,
Diener, E., & Diener, M. (1995). Cross-cultural correlates of life satisfaction and self-esteem. Journal of Personality and Social Psychology,
Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment,
Diener, E., & Seligman, E. (2002). Very happy people. Psychological Science,
Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being three decades of progress. Psychological Bulletin,
Emmons, R. A., & Diener, E. (1985). Personality correlates of subjective well-being. Personality and Social Psychology Bulletin,
Frazier, P. A., Tix, A. P., & Barron, K. E. (2004). Testing moderator and mediator effects in counseling psychology research.
Journal of Counseling Psychology,
Gilman, R., & Huebner, E. S. (2003). A review of life satisfaction research with children and adolescents. School Psychology Quarterly,
Gilman, R., & Huebner, E. S. (2006). Characteristics of adolescents who report very high life satisfaction. Journal of Youth and Adolescence,
Greenspoon, P. J., & Saklofske, D. H. (2001). Toward an integration of subjective well-being and psychopathology. Social Indicators Research,
Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria
versus new alternatives. Structural Equation Modeling,
Huebner, E. S., Laughlin, J. E., Ash, C., & Gilman, R. (1998). Further validation of the multidimensional students’ life satisfaction
scale. Journal of Psychoeducational Assessment,
Huebner, E. S., Suldo, S. M., Valois, R. F., Drane, J. W., & Zullig, K. J. (2004). Brief multidimensional students’ life satisfaction
scale: Sex, race, and grade effects in a high school sample. Psychological Reports,
Jöreskog, K. G., & Sörbom, D. (1996). LISREL 8: User’s reference guide. Chicago, IL: Scientific Software International. Li, L. W., & Liang, J. (2007). Social exchanges and subjective well-being among older Chinese: Does age make a difference?
Psychology and Aging,
Mahon, N. E., Yarcheski, A., Yarcheski, T. J., Cannella, B. L., & Hanks, M. M. (2006). A meta-analytic study of predictors
for loneliness during adolescence. Nursing Research,
McWhirter, B. T., Tricia, M., Besett-Alesch, T. M., Horibata, J., & Gat, I. (2002). Loneliness in high risk adolescents: The
role of coping, self-esteem, and empathy. Journal of Youth Studies,
Neto, F. (2002). Loneliness and acculturation among adolescents from immigrant families in Portugal. Journal of Applied Social Psychology,
Pavot, W., & Diener, E. (1993). Review of the satisfaction with life scale. Psychological Assessment,
Pinquart, M., Silbereisen, R. K., & Juang, L. P. (2004). Moderating effects of adolescents’ self-efficacy beliefs on psychological
responses to social change. Journal of Adolescent Research,
Quintana, S. M., & Maxwell, S. E. (1999). Implications of recent development in structural equation modeling for counseling
psychology. The Counseling Psychologist,
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Sedikides, C., Rudich, E. A., Gregg, A. P., Kumashiro, M., & Rusbult, C. (2004). Are normal narcissists psychologically healthy?
Self-esteem matters. Journal of Personality and Social Psychology,
Shaver, P. R., & Brennan, K. A. (1991). Measures of depression and loneliness. San Diego, CA: Academic Press. Skok, A., Harvey, D., & Reddihough, D. (2006). Perceived stress, perceived social support, and wellbeing among mothers of
school-aged children with cerebral palsy. Journal of Intellectual and Developmental Disability,
Stroebe, W., Stroebe, M., Abakoumkin, G., & Schut, H. (1996). The role of loneliness and social support in adjustment to loss:
A test of attachment versus stress theory. Journal of Personality and Social Psychology,
Suldo, S. M., & Huebner, E. S. (2006). Is extremely high life satisfaction during adolescence advantageous? Social Indicators Research,
Tam, K. P., Lau, H. P., Jiang, D. (2010). Culture and subjective well-Being: A dynamic constructivist view. Journal of Cross–Cultural Psychology. doi:10.1177/0022022110388568. Tan, H. Z., Luo, Y. J., Wen, S. W., Liu, A. Z., Li, S. Q., & Yang, T. B. (2004). The effect of a disastrous flood on the quality
of life in Dongting Lake Area in China. Asia-Pacific Journal of Public Health,
Wang, X. D., Wang, X. L., & Ma, H. (1999). Rating scales for mental health: Social support rating scale. China, BJ: The Journal of Chinese Psychological Press. Yarcheski, A., Mahon, N. E., & Yarcheski, T. J. (2001). Social support and well-being in early adolescents: The role of mediating
variables. Clinical Nursing Research,
Zhang, L., & Leung, J. (2002). Moderating effects of gender and age on the relationship between self-esteem and life satisfaction
in mainland Chinese. International Journal of Psychology,