Sun et al.

BMC Public Health (2024) 24:483 BMC Public Health

https://doi.org/10.1186/s12889-024-17795-5

RESEARCH Open Access

Social anxiety and loneliness among older

adults: a moderated mediation model
Shuting Sun1†, Yawen Wang1,2†, Lilu Wang1†, Jinjin Lu1†, Huihui Li3†, Jiahui Zhu1, Suzhen Qian1, Lianlian Zhu1* and
Hongbo Xu1*

Abstract
Background Few studies have clarified the mechanisms linking social anxiety and loneliness in older populations.
The study aimed to explore how social network mediate the relationship between social anxiety and loneliness in
older adults, with perceived social support playing a moderating role.
Methods A total of 454 older patients completed the Social Avoidance and Distress Scale, Lubben Social Network
Scale-6, Chinese version of the Short Loneliness Scale and Perceived Social Support Scale. Bootstrap and simple slope
methods were used to test the moderated mediation model.
Results Social anxiety had a significant positive predictive effect on loneliness and social network partially mediated
this relationship. The relationship between social anxiety and social network, as well as the relationship between
social network and loneliness, was moderated by perceived social support. Specifically, perceived social support
buffered the effects of social anxiety on social network, but the buffering effect diminished with increasing levels
of social anxiety. On the social network and loneliness pathway, the social network of older persons with higher
perceived social support has a stronger prediction of loneliness.
Conclusions The study found that social anxiety can contribute to loneliness by narrowing older adults’ social
network. High perceived social support can buffer this process, but do not overstate its protective effects. Thus,
interventions to reduce social anxiety and improve social network and social support may help prevent and alleviate
loneliness in older adults.
Keywords Social anxiety, Social network, Perceived social support, Loneliness, Older adults

Background
In the context of rapid population aging, loneliness has
†
Shuting Sun, Yawen Wang, Lilu Wang, Jinjin Lu and Huihui Li are become an important topic in public policy and public
contributed equally to this work.
health. Compared to younger adults, older adults are at
*Correspondence: a higher risk of social isolation and participate in fewer
Lianlian Zhu
zhll@wmu.edu.cn social activities, which makes them vulnerable to loneli-
Hongbo Xu ness with declining physical and cognitive abilities, shift-
xhb@wmu.edu.cn ing social roles, and worsening social adaptability [1–3].
1
School of Nursing, Wenzhou Medical University, University Town,
Wenzhou, China This subjective emotional experience arises when older
2
Department of Nursing, Jiangsu Provincial Xuzhou Pharmaceutical adults feel that the quantity and quality of their social
Vocational College, Xuzhou, China relationships do not actually match their expectations
3
School of Public Health and Management, Wenzhou Medical University,
Wenzhou, China [4]. Recently, it was reported that 11.9% of older adults

© The Author(s) 2024. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use,
sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and
the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included
in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The
Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available
in this article, unless otherwise stated in a credit line to the data.
Sun et al. BMC Public Health (2024) 24:483 Page 2 of 11

worldwide suffer from loneliness [5]. A study performed dysfunctional interpersonal loop has the potential for
by Wei et al. based on data from the 2008/2009 wave adverse social effects, such as loneliness [20]. On this
from the Chinese Longitudinal Healthy Longevity Survey basis, in order to establish a more comprehensive model
found that 33.3% of older adults feel lonely [4]. Loneliness that enhances the understanding of the relationship
is also strongly associated with negative physical and psy- between SA and loneliness, the study incorporates two
chological outcomes in older adults. Previous research variables (social network and perceived social support
has shown that higher levels of loneliness may predict (PSS)). Social network concerns the number and fre-
frailty, depression, cognitive decline, an increased risk quency of relationships a person establishes and main-
of cardiovascular disease, suicide, and mortality [6–8]. tains (objective structural aspects of social relationships),
Therefore, it has become a focal point of mental health consisting of family members, close friends, neighbors,
issues for older adults on how to alleviate loneliness. and others individuals in their social circle [22, 23]. Previ-
People with social anxiety (SA) often exhibit signs of ous studies have shown that SA is negatively associated
anxiety, fear, or discomfort in social interactions and even with social network, and those with high SA typically
avoid social situations to evade negative comments from have smaller social networks [24, 25]. The core feature of
others [9, 10]. The prevalence of SA often decreases with SA is avoiding social interactions [16], thus anticipating
age. Bai et al.‘s assessment of older persons with chronic a reduction in the size and frequency of social network
disease using the Social Avoidance and Distress Scale in socially anxious individuals. On the other hand, the
revealed a lower score compared to a study of a sample idea that positive social network is beneficial to an indi-
of college students [11]. Nevertheless, the 12-month inci- vidual’s physical and mental health is widely supported
dence of diagnosed SA disorders in persons aged 65 and by theoretical and empirical research [26–28]. Loneliness
older was reported to be 2.3%, which is the second most is significantly impacted by social network, with the lat-
frequent anxiety disorder in this age group [12]. Previ- ter acting as a negative predictor of loneliness [28]. In the
ous studies have demonstrated a substantial correlation Amsterdam Longitudinal Study of Aging, Domènech et
between SA and loneliness in later years. For example, a al. [29] followed participants over 75 years of age for 11
longitudinal study guided by Lim et al. discovered that years and found that decreasing social network size leads
SA is only significant predictor of future loneliness, with to higher levels of loneliness over time. If a person is iso-
earlier SA positively predicting future loneliness [13]. lated from family or friends, their perception of quality
Hoffman et al. [14] found that older adults with high SA of life may decline and they are more likely to experience
were more likely to experience intimate loneliness (one loneliness when functional and emotional needs are not
of the characteristics of loneliness, i.e., lower quantity or adequately met [30]. Conversely, positive social contacts
quality of intimate companionship) than younger adults. and wider networks showed the capacity for social adap-
People who feel lonely psychologically actually expect tation, generating more social resources and support,
to construct social connections with others [15]. How- allowing older adults to maintain a positive attitude and
ever, avoidance of social situations and concerns about a sense of belonging [31]. Collectively, weak social net-
the threat of social situations may prevent these con- work was likely to be accompanied by SA and loneliness.
nections from forming, leading to decreased satisfac- Thus, based on the interpersonal model of SA and prior
tion with interpersonal relationships, a lack of intimacy, evidence, we hypothesized that SA may exacerbate older
and increased loneliness in older adults [14, 16]. Cur- adults’ dissatisfaction with family and friend network
rently, the association between SA and loneliness has (loneliness) by reducing the size of these ties and the fre-
been widely studied in children, adolescents, and young quency of contact.
adults [9, 17, 18], while a limited number of studies have Social support is categorized into received social sup-
been conducted in older adults, most of which use a port and PSS [32]. PSS is broadly defined as individuals’
wide range of ages [13, 19]. Besides, relevant pathways perceptions of the availability of social support in their
or internal mechanisms between these variables in older networks, emphasizing the subjective emotional expe-
adults have not been completely explored, and further rience and satisfaction of an individual when they feel
investigation is needed to facilitate prevention and inter- respected, supported and understood in society (func-
ventions for loneliness in older populations. tional aspects of social relationships) [23, 32]. Compared
The interpersonal model of SA proposes that when to received social support, PSS has a stronger relationship
social situations may trigger SA, people will adopt self- with a person’s mental health [33]. According to Hobfoll’s
protective strategies that result in low density and dys- conservation of resources theory, people are susceptible
functional social relationships [20]. Whether through to resource loss when faced with stressful circumstances,
active, passive, or a combination of both avoidance yet protecting and maintaining resources can mitigate
strategies, individuals tend to have lower expectations the potential negative consequences of stress [34]. There-
of the outcome of their social interactions [21]. This fore, PSS may act as an adjustable resource to alleviate
Sun et al. BMC Public Health (2024) 24:483 Page 3 of 11

stress and promote individual mental health. Ren et al. age ≥ 60 years; (2) choose to age in place; (3) clear con-
discovered that PSS had a negative association with SA sciousness; (4) have the ability to read or speak, and com-
and moderated the link between physical activity and SA municate with investigators without difficulty. All older
in left-behind children [35]. Among older adults, PSS was adults combined with serious organic disease or mental
a protective factor against loneliness, and the predictive illness were excluded.
effect of chronic diseases on loneliness was more signifi- A priori power analysis with GPower 3.1.9.7 [37]
cant with low levels of PSS [36]. However, to our knowl- revealed that a linear regression with a significance level
edge, no thorough investigation has been made into the of 0.05, power of 0.80, and medium effect size of 0.15 for
mechanisms of the moderating role of PSS in the triadic a maximum of 15 variables (including demographic vari-
interaction between SA, social network, and loneliness ables) required a minimum sample size of 139. Besides,
in older persons. A high level of PSS can make older studies suggested that sample sizes are generally 5 ~ 10
persons feel valued and motivated to socially interact, times the number of variables [38]. Thus, this study
expanding their social networks when they suffer anxiety yielded a sample size of 150 based on 10 times the num-
due to socially uncertain situations and negative evalua- ber of variables. Considering a 20% attrition rate, the final
tions. Thus, based on the conservation of resources the- sample size was estimated to be 174 ~ 188. Ultimately, 454
ory and existing research evidence, we hypothesized that participants were actually included in the study.
PSS moderates the mediating effect of the social network
on the association between SA and loneliness. Data collection
One of the most well-liked psychosocial models cur- All investigators received standardized training before
rently explores how social network and social support conducting the investigation, ensuring familiarity with
buffer the effects of life events or changes on health [22]. the investigation methods and techniques. During the
Social network highlights the quantity of social relation- formal survey, the investigators firstly explained the
ships, whereas social support emphasizes the function research objectives and response requirements to older
of social relationships. They explain different aspects of people. Only after obtaining the informed consent did,
interpersonal relationships. By incorporating both social they distribute the questionnaire on a one-to-one basis.
network and PSS into the model, this study aims to pro- Older adults with literacy skills could fill in the ques-
vide a more comprehensive understanding of the effects tionnaire by themselves. For illiterate older persons, the
of social relationships on SA and loneliness in older investigators would ask questions one by one. If there
adults, enrich research on loneliness in older adults, and was any difficulty in comprehension, investigators make
provide a new empirical basis for reducing loneliness in neutral and accurate word explanations and record
older adults. The research hypothesis are as follows: according to their original answers. After completing the
questionnaire, a second investigator carefully checked
H1 SA is positively associated with loneliness in older the completeness of the data. Any missing items were
adults. filled in promptly and verified on the spot. In this survey,
483 older people were interviewed and they filled out the
H2 Social network mediates the relationship between SA questionnaires. However, 29 participants did not com-
and loneliness in older adults. plete all the items due to missing information or opting
out of the program, so they were excluded. Finally, 454
H3 PSS moderates the relationship between SA and valid questionnaires were acquired for data analysis. The
social network and the relationship between social net- effective response rate of the questionnaire was 93.9%.
work and loneliness.
Measures
Methods Social-demographic information
Participants Socio-demographic data were collected using our own
Using the quota sampling method, a total of 214 older designed short questionnaire, which included age, gen-
adults in 20 nursing homes of different sizes and 240 der, marriage status, education background, self-reported
older adults living in the community in 5 cities in Zhe- economic status, degree of filial piety of children, self-
jiang Province were surveyed from January to May 2021. rated health and religious belief.
Inclusion criteria for the elderly in nursing homes com-
prised the following: (1) age ≥ 60 years; (2) live in a nurs- Social anxiety
ing home ≥ 1 month; (3) clear consciousness; and (4) The Social Avoidance and Distress Scale (SAD), devel-
have the ability to read or speak, and communicate with oped by Watson et al. and translated by Wang et al., was
investigators without difficulty. Inclusion criteria for the used to assess the SA of the older population [10, 39]. It
older adults living in the community were as follows: (1) comprises two dimensions: social avoidance (e.g., “I tend
Sun et al. BMC Public Health (2024) 24:483 Page 4 of 11

to withdraw from people”) and social distress (e.g., “It’s Ethical considerations
easy for me to relax when I am with strangers”). The scale This study was approved by the Institutional Review
has 28 items, 14 of which are rated from 0 (False) to 1 Board of Wenzhou Medical University (approval number:
(True), and the remaining 14 items are reverse scored. 2021-011). First, the principle of informed consent was
The total score varied from 0 to 28 with higher scores abided by before administering the survey. Second, the
indicating a higher degree of avoidance or distress. SAD identity information of the participants was kept strictly
has been validated and used in the older population, confidential and not disclosed to members outside the
although few articles are available [11, 40]. The Cron- research team. Third, Participants had the right to decide
bach’s α of the scale in this study was 0.90. whether to participate or not, and can withdraw from the
investigation at any time. All methods were performed in
Loneliness accordance with the relevant rules and regulations of the
The 6-item De Jong Gierveld Loneliness Scale (DJG-6), Declaration of Helsinki.
developed by deJong Gierveld et al. and translated into
Chinese by Leung et al., was applied to assess the level of Data analysis
loneliness [41, 42]. The scale includes two dimensions of SPSS 26.0 was applied for statistical analysis. The
emotional loneliness (e.g., “I experience a general sense of numeric data in this study were identified as non-nor-
emptiness”) and social loneliness (e.g., “There are enough mal distribution after the Kolmogorov-Smirnov test, so
people I feel close to”). Each item has answer categories the non-normal distributed data were described by the
of no, more or less and yes, where the emotional dimen- median and interquartile range. Categorical data were
sion is 0 (no), more or less (1) and yes (1) and social expressed in frequency and percentage. The Mann-
loneliness dimension is scored using a reverse scoring Whitney U test and the Kruskal-Wallis test were used
method. The higher the score, the higher the level of to assess differences in loneliness across demographic
loneliness. The scale has been verified as a tool with high characteristics. Spearman correlation was used for cor-
reliability and validity for measuring loneliness in large relation analysis between variables. Considering that the
surveys of older adults [42]. In this study, the Cronbach’s children’s filial piety and self-rated health status may be
α of the scale is 0.73. related to loneliness, they were included as control vari-
ables [46, 47]. Model 4 and 58 in the SPSS 26.0 macros
Social network program PROCESS compiled by Hayes [48], was used
The Chinese version of the Lubben Social Network to construct the moderated mediation model with 5,000
Scale-6 (LSNS-6) was used to measure social network bootstrap sampling. To further explore the moderating
of older adults, demonstrating high reliability and valid- effect of PSS, PSS was divided into two groups, namely,
ity [43, 44]. The scale is a 6-item self-report measure high and low PSS by adding and subtracting a standard
with two domains: family network (e.g., “How many rela- deviation by mean. Subsequently, the simple slope analy-
tives do you see or hear from at least once a month?”), sis was carried out. P value < 0.05 was considered statisti-
and friend network (e.g., “How many of your friends do cally significant. To check for the possibility of common
you see or hear from at least once a month?”). Each item method deviation, Harman’s single factor test was used.
is scored from 0 (none) to 5 (9 or more). The total score It is based on exploratory factor analysis to determine
ranges from 0 to 30, with a high score indicating a better the number of factors needed to explain the variance of
social network. In this study, the Cronbach’s α was 0.80. a variable. The bias is more likely if a single factor pre-
cipitates or explains a greater variance [49]. The variance
Perceived social support explained by a single factor is generally considered to be
The Perceived Social Support Scale (PSSS) was developed less than 40% [50].
by Zimet et al. and revised by Jiang [39, 45]. This scale
consists of 12 items organized into 3 dimensions: family Results
support (e.g., “My family can help me concretely”), friend Common method bias
support (e.g., “My friends can share happiness and sad- The results showed that 12 factors with eigenvalue
ness with me”) and other support (e.g., “Some people greater than 1 were co-precipitated, and the variance
(relatives, colleagues, neighbors) in my life care about my explained by the first factor was 20.03%, which was less
feelings”). Each item uses a seven-level scoring method than the critical standard of 40%, indicating that the com-
from 1(totally disagree) to 7 (completely agree). The total mon method deviation of this study was not significant.
score ranges from 12 to 84, and a higher score indicates
a higher level of PSS. The Cronbach’s α in this study was Demographic characteristics
0.91. Table 1 lists the demographic characteristics. The average
age of the 454 older adults (214 older people in nursing
Sun et al. BMC Public Health (2024) 24:483 Page 5 of 11

Table 1 Social-demographic characteristics and comparison of Table 2 Correlation coefficient of SA, social network, loneliness
loneliness scores in different groups (n = 454) and PSS
Variable N (%) Loneliness Z/H Median 1. SA 2. Social 3. Loneliness 4.
(median (IQR) network PSS
(IQR)) 1. SA 7.00 1
Age (years) (10.25)
60 ~ 74 193 (42.5) 3.00 (4.00) 0.66 2. Social 15.00 − 0.292** 1
75 ~ 89 235 (51.8) 3.00 (4.00) network (10.00)
≥ 90 26 (5.7) 2.00 (3.00) 3. Loneliness 3.00 (4.00) 0.484** − 0.416** 1
Gender 4. PSS 64.00 − 0.298** 0.558** − 0.524** 1
Male 204 (44.9) 3.00 (4.00) −1.71 (18.00)
Abbreviations: IQR: Interquartile Range; SA: Social anxiety; PSS: Perceived social
Female 250 (55.1) 3.00 (3.00)
support
Marital status
Note: **p < 0.01. Spearman correlation was used for correlation analysis
Married 232 (51.1) 2.00 (3.00) −1.81
Single/divorced/widowed 222 (48.9) 3.00 (4.00)
Education background their children, loneliness scores were lower among older
Illiterate 138 (30.4) 3.00 (3.00) 1.42 adults who perceived their children to be filial. Besides,
Graduated primary school 174 (38.3) 3.00 (3.25) older adults with good self-rated health had lower lone-
Graduated junior high school 142 (31.3) 3.00 (4.00) liness scores compared to those with poorer self-rated
and above health.
Self-reported economic status
Enough and excess 221 (48.7) 3.00 (4.00) 4.85 Correlations among the main variables
Roughly enough 174 (38.3) 3.00 (4.00) In this study, the correlations among the four variables
difficulties 59 (13.0) 2.00 (3.00) of SA, social network, loneliness and PSS were analyzed.
Degree of filial piety of children
The correlation matrix was presented in Table 2. The
Filial 350 (77.1) 2.00 (3.00) 54.53***
results showed that: (1) SA was significantly negatively
General/unfilial 88 (19.4) 5.00 (1.25)
correlated with social network and PSS, but significantly
Childless 16 (3.5) 3.00 (3.75)
positively correlated with loneliness; (2) Social network
Self-rated health
was significantly negatively correlated with loneliness,
Good 254 (55.9) 2.00 (3.00) 33.30***
but significantly positively correlated with PSS; (3) Lone-
Moderate 141 (31.1) 4.00 (3.00)
liness was significantly negatively correlated with PSS. All
Poor 59 (13.0) 4.00 (3.00)
data were standardized before further analysis.
Religious belief
Strong 103 (22.7) 3.00 (4.00) 0.17
Mediation model
General 172 (37.9) 3.00 (4.00)
No 179 (39.4) 3.00 (4.00)
As shown in Table 3, after controlling for filial piety and
Abbreviations: IQR: Interquartile Range self-rated health, SA was positively related to loneliness
Note: ***P < 0.001 (Β = 0.40, t = 9.68, P < 0.001). After adding social network
as a mediating variable, SA (Β = 0.33, t = 8.07, P < 0.001)
and social network (Β = −0.27, t = −6.80, P < 0.001) were
homes and 240 in the community) was 76.3 ± 8.5 years positively and negatively correlated with loneliness,
old, with 193 (42.5%) people aged 60 ~ 74 years and 235 respectively. Besides, SA was found to significantly
(51.8%) people aged 75 ~ 89 years. 204 (55.1%) were negatively predict social network (Β = −0.26, t = −5.62,
women and 232 (51.1%) were married. In terms of educa- P < 0.001). In the bias-corrected percentile bootstrap
tional background, only 31.3% of the participants gradu- analysis, the mediating effect of social network on the
ated from junior high school or above. A total of 395 relationship between SA and loneliness was significant
participants (87%) self-reported their economic status (ab = 0.07, 95% CI [0.04, 0.11]), accounting for 17.74% of
was roughly enough and enough and excess. Most par- the total effect. Regarding the direct effect of SA on lone-
ticipants (77.1%)perceived their children to be filial. Fur- liness, the 95% CI ([0.25, 0.41]) did not contain 0, indi-
thermore, 55.9% of the elderly rated their health as good cating social network partially mediated the relationship
and 39.4% had no religious beliefs. Comparative analy- between SA and loneliness.
ses of loneliness based on demographic characteristics
showed no statistically significant differences in loneli- Moderated mediation model
ness scores with respect to age, gender, marital status, The moderated mediation model testing result was dis-
education background, self-reported economic status played in Table 4. SA has a significant negative predictive
and religious belief. Regarding the degree of filial piety of effect on social network (Β = −0.15, t = −3.59, P < 0.001).
Sun et al. BMC Public Health (2024) 24:483 Page 6 of 11

Table 3 Testing the mediation effects of social network in the relation between SA and loneliness
Regression equation Global fit index Significance of regression coefficient
Outcome variable Predictor variable R R2 F Β (95% CI) t
Loneliness Filial piety 0.53 0.28 58.01 0.43 (0.28, 0.59) 5.53***
Self-rated health 0.13 (0.04, 0.22) 2.92**
SA 0.40 (0.32, 0.49) 9.68***
Social network Filial piety 0.33 0.11 18.59 −0.32 (−0.50, −0.15) −3.74***
Self-rated health 0.03 (−0.12, 0.07) 0.51
SA −0.26 (−0.35, −0.17) −5.62***
Loneliness Filial piety 0.59 0.35 59.44 0.34 (0.19, 0.49) 4.53***
Self-rated health 0.12 (0.04, 0.21) 2.90**
SA 0.33 (0.25, 0.41) 8.07***
Social network −0.27 (−0.35, −0.20) −6.80***
Abbreviations: SA: Social anxiety
Note: **P < 0.01, ***P < 0.001. Filial piety and self-rated health were analyzed as control variable

Table 4 Testing the moderated mediation effect in the relation between SA and loneliness
Regression equation Global fit index Significance of regression coefficient
Outcome variable Predictor variable R R2 F Β (95% CI) t
Social network Filial piety 0.59 0.35 47.42 −0.01 (−0.16, 0.15) −0.12
Self-rated health 0.02 (−0.06, 0.11) 0.55
SA −0.15 (−0.23, −0.07) −3.59***
PSS 0.55 (0.46, 0.63) 12.70***
SA×PSS −0.08 (−0.16, −0.01) −2.13*
Loneliness Filial piety 0.65 0.42 53.25 0.25 (0.10, 0.39) 3.25**
Self-rated health 0.08 (0.00, 0.16) 1.85
SA 0.31 (0.23, 0.39) 7.89***
Social network −0.10 (−0.19, −0.01) −2.26*
PSS −0.35 (−0.45, −0.26) −7.28***
Social network × PSS −0.11 (−0.18, −0.04) −3.07**
Abbreviations: SA: Social anxiety; PSS: Perceived social support
Note: *P < 0.05, **P < 0.01, ***P < 0.001. Filial piety and self-rated health were analyzed as control variable

Table 5 The moderating effect of different levels of PSS To further analyze the moderating effect of PSS in the
between SA and social network mediation model, we regarded the mean of PSS plus
PSS Effect size SE Boot LLCI Boot ULCI one standard deviation (M + SD) as the high-level group,
M-SD −0.06 0.06 −0.17 0.05 the mean (M) as the medium-level group, and the mean
M −0.15*** 0.04 −0.23 −0.07 minus one standard deviation (M-SD) as the low-level
M + SD −0.23*** 0.06 −0.34 −0.12 group for simple slope analysis. The moderating effect
Abbreviations: SA: Social anxiety; PSS: Perceived social support; LL: Low Limit; UL: of different levels of PSS between SA and social network
Upper Limit
was shown in Table 5 and Fig. 1. The results revealed
Note: ***P < 0.001
that the predictive effect of SA on the social network of
the elderly with medium PSS (M) was lower than that of
SA (Β = 0.31, t = 7.89, P < 0.001) and social network the elderly with high PSS (Βmedium PSS = −0.15, p < 0.001;
(Β = −0.10, t = −2.26, P < 0.05) were positively and nega- Βhigh PSS = −0.23, p < 0.001), whereas the relationship
tively related to loneliness, respectively. PSS was signifi- between SA and loneliness was not found in low PSS
cantly negatively predicted loneliness (Β = −0.35, t = −7.28, (Βlow PSS = −0.06, p > 0.05). It indicated that the predictive
P < 0.001). The interaction effect of SA and PSS on social effect of SA on social network increased gradually with
network was significant (Β = −0.08, t = −2.13, P < 0.05), the improvement of PSS of the elderly.
indicating that the pathway of “SA → social network” The moderating effect of different levels of PSS
was moderated by PSS. Moreover, the interaction effect between social network and loneliness was shown in
of social network and PSS on loneliness was significant Table 6 and Fig. 2. The results revealed that social net-
(Β = −0.11, t = −3.07, P < 0.01) indicating that the pathway work of older adults with medium PSS were less effective
of “social network → loneliness” was moderated by PSS. in predicting loneliness than those with high PSS (Βmedium
PSS = −0.10, p < 0.05; Βhigh PSS = −0.21, p < 0.001), whereas
Sun et al. BMC Public Health (2024) 24:483 Page 7 of 11

Fig. 1 The moderating effect of perceived social support on social anxiety and social network

Table 6 The moderating effect of different levels of PSS

the relationship between social network and loneliness
between social network and loneliness
was not found in low PSS (Βlow PSS = 0.01, p > 0.05). This
PSS Effect size SE Boot LLCI Boot ULCI
showed that the predictive effect of social network on
M-SD 0.01 0.06 −0.11 0.13
loneliness gradually increased with the improvement of
M −0.10* 0.05 −0.19 −0.01
PSS for older adults.
M + SD −0.21*** 0.05 −0.31 −0.11
Abbreviations: PSS: Perceived social support; LL: Low Limit; UL: Upper Limit
Note: *P < 0.05, ***P < 0.001

Fig. 2 The moderating effect of perceived social support on social network and loneliness
Sun et al. BMC Public Health (2024) 24:483 Page 8 of 11

Discussion Altogether, the study of the mediating role of social

This study employed social network as the mediating network has enriched the understanding of the inter
variable and PSS as the moderating variable to construct mechanisms underlying the relationship between SA and
a moderated mediation model. In addition to elucidat- loneliness, highlighted the exploration of the antecedents
ing the processes by which SA affects loneliness in older of social network, and validated the interpersonal model
adults, this novel evidence provided important responses of SA. Future practice should pay attention to developing
to the conditions under which SA affects loneliness, and effective strategies that not only address the issue of SA
contributed to the development of effective targeted among older people, but also strive to strengthen social
interventions. Specifically, SA had an impact on older network in order to reduce the occurrence of loneliness.
adults’ loneliness through social network. Additionally,
PSS moderated the relationship between SA and social The moderating role of PSS
network as well as the relationship between social net- According to hypothesis 3, the results demonstrated that
work and loneliness. PSS moderated the mediation role of social network. In
the pathway of “SA → social network”, social network
The mediating role of social network was higher in the high PSS state at both low and high SA,
In this study, SA had a significant positive predictive suggesting a protective effect of PSS on social network
effect on loneliness, in line with findings from several reduction due to SA. However, this protective effect rap-
studies [14, 19]. The outcome highlights SA is a risk fac- idly diminished with higher SA, indicating that PSS is a
tor of becoming lonely, and that assessing and interven- stress-vulnerability factor in pathway 1. Although there
ing in SA in older adults can help identify and reduce was no support for stress-buffering in the result, there
loneliness. Furthermore, social network was further was also an indication of the protective effect of PSS.
found to play a mediating role between SA and loneli- Rueger et al.’s study on the association between PSS and
ness, confirming the validity of hypothesis 2. SA not only depression conceptualized protective factors’ buffering
affected feelings of loneliness directly, but also exerted an effect in two different ways: stress-buffering (effects of
indirect effect on loneliness through social network. Sev- social support are enhanced) and reverse stress-buffering
eral reasons contribute to this phenomenon: first, over (effects of social support are dampened) [55]. The find-
time, older adults may experience negative events, such ing of this study was consistent with reverse stress-buff-
as age-related physical limitation, retirement or widow- ering model, where the protective effect of PSS is limited.
hood, which lead to negative self-perceptions and feel- High level of PSS arose from interpersonal relationships,
ings of disconnection from an increasingly fast-paced which, to some extent, created part of the social network
and rapidly evolving society [51, 52]. Experience SA, for older adults. It can boost confidence in older adults
especially if fearful of social situations and negative eval- and keep them emotionally upbeat to cope with stress-
uations, can lead to small network and infrequent social ful threats in social settings, therefore becoming useful in
interaction, making it more difficult to build and sustain preventing the reduction of social network in older adults
intimate social ties [24, 25]. Consequently, they are more suffering from SA [3, 56]. The cognitive-behavioral model
susceptible to social isolation. Second, studies have con- of SA [57], however, posits that individuals with SA will
sistently shown a strongly negative relationship between be preoccupied with self-representations that they regard
social network and loneliness [15, 28, 53]. Social network as faulty or related to not meeting social expectations or
can serve as a protective factor against loneliness when norms. When seeking PSS, older adults may focus on
they satisfy older adults’ demands for a sense of belong- diminished physical functioning, compare themselves
ing and desired social relationships [53]. Due to the influ- to those believed to be highly capable, and acquire low
ence of the traditional culture of filial piety in China [4], self-esteem [58, 59]. On the other hand, when older
most elderly have deep-rooted family values and expect adults suffer from high level SA, they may be more sen-
close, interdependent family relationships. However, as sitive to any perceived differences in social relationships
children move out and the proportion of empty nesters and adopt negative and avoiding attitudes, thereby deny-
gradually increases, the size and density of older adults’ ing appropriately received social support and resulting
family network may gradually decrease [27]. The social in a lack of PSS. Low self-esteem and hypersensitivity to
network will be single and fixed. Many senior citizens social situations weakened the protective role of PSS in
choose to migrate to their adult children’s hometowns, the process of SA-induced social network narrowing. It
accompanied by language and cultural barriers, may con- might also be understood that once stress reaches a cer-
tribute to a decline in existing network, making it chal- tain level, PSS may lose its ability to offset SA. In the case
lenging to make new friends [51, 54]. All these problems of low PSS, older adults’ social networks may have been
can lead to social isolation and loneliness. restricted. Therefore, their social networks changed less
regardless of high or low SA.
Sun et al. BMC Public Health (2024) 24:483 Page 9 of 11

Regarding the pathway of “social network → loneli- validated in previous studies with satisfactory reliabil-
ness”, the study also found that loneliness decreased sig- ity and validity, all variables were based on older adults’
nificantly with the increase of social network in the case self-report, introducing the possibility of self-report bias.
of high PSS, suggesting that PSS plays a facilitating role Finally, due to the dialect problem, only non-random
in the process of social network affecting loneliness, con- sampling, rather than random sampling, could be used
sistent with the conservation of resources theory [34]. in the study, which affected the representativeness of the
When older adults feel high PSS, they gain more value- study sample to some extent.
affirming, self-worth and life satisfaction [3, 30]. They
exhibit more positivity and optimism when interacting Relevance to clinical practice
with others and are able to perceive the benefits of objec- This study provides new perspectives for nursing home
tive social networks, which, in turn considerably improve or community managers to lessen loneliness in older
social networks, such as network size [2, 29], and reduce adults. First, prompt screening and treatment for SA can
loneliness. In addition, the connection between older be an effective measure to identify and reduce loneli-
adults and those in their social networks who provide ness in older adults. The degeneration of psychological
support not only promotes the development of inti- and physical functions brought about by aging makes the
mate relationships, but also increases the likelihood of elderly less adaptable to social and more fearful of nega-
exposure to health information, further boosting mental tive evaluation, so they tend to avoid social situations. It
health [60]. Conversely, the study discovered that in the is necessary to provide cognitive behavioral training to
event of reported poor PSS, the decrease in loneliness did this population to reduce SA. Second, older people with
not change significantly with increasing social network, SA may not initiate social interactions. Social assistance
indicating that the elderly could not effectively access and inviting them to participate in community building
the positive effect of social network when lacking social or in the daily management of the nursing home con-
support. Therefore, they were unable to successfully tribute to the expansion of social network. Third, the
reduce loneliness through expanding their social net- assessment and intervention of PSS also deserve atten-
work. Accordingly, building a strong social support net- tion. Measures to increase the perceived support of older
work is an important way for older adults to maintain a adults, especially emotional support from family and
good psychological state since it may help social network friends, should be taken along with improving social net-
strengthen and mitigate loneliness. Nevertheless, given work to enhance their use of PSS and reduce loneliness.
the relatively limited protective effect of PSS, just trying For example, nursing home or community managers can
to increase their level of PSS may not be appropriate for work with family and friends to establish small social
lonely older adults experiencing SA. How to reduce SA in support groups, organizing regular speech contests,
older adults may be the focus of interventions. group interaction activities, and changing the location of
the event from time to time to help participants with SA
Limitations adjust to different social situations. However, it’s crucial
The current research still has some limitations. First, as not to overstate the role of protective resources such as
the study was designed as a cross-sectional survey, mak- PSS.
ing causal inferences was not possible. However, the
moderated mediation model was based on a theoretical Conclusion
foundation and supported by previous empirical studies, Collectively, although further in-depth research is
so the cross-sectional survey can still provide valuable needed, this research represents an important step in
information about the relationships between variables. clarifying the relationship between SA and loneliness
More longitudinal studies are needed in the future to among elderly. The results addressed the question of
improve the representativeness of the moderated media- how SA affects the loneliness in older adults, highlight-
tion model. Second, the LSNS- 6 scale was used to mea- ing the negative predictive effect of SA on loneliness and
sure social network in the study, which includes relatively identifying conditions under which the mediating effect
objective items characterizing the structure of social of social network and the moderating effect of PSS are
relationships (i.e. size of active network, size of intimate more significant. These findings emphasized the impor-
network and frequency of contact). But the size of social tance of psychosocial factors and revealed the poten-
network and the frequency of contact among different tial mechanisms by which SA affects loneliness in older
older adults might affect the availability and effective- adults. Additionally, it can assist nursing home or com-
ness of resources as well as the study results. It is rec- munity managers in carrying out efficient psychological
ommended that the role of these variables be further interventions to further improve the quality of life and
explored to suggest more detailed interventions. Third, psychological well-being of older adults, meeting the
although the questionnaire used in this study has been emotional needs of society.
Sun et al. BMC Public Health (2024) 24:483 Page 10 of 11

Abbreviations 8. Gale CR, Westbury L, Cooper C. Social isolation and loneliness as risk factors
SA Social Anxiety for the progression of frailty: the English Longitudinal Study of Ageing. Age
PSS Perceived Social Support Ageing. 2018;47(3):392–7. https://doi.org/10.1093/ageing/afx188.
SAD The Social Avoidance and Distress Scale 9. Maes M, Nelemans SA, Danneel S, Fernández-Castilla B, Van den Noortgate W,
DJG-6 The 6-item De Jong Gierveld Loneliness Scale Goossens L, et al. Loneliness and social anxiety across childhood and adoles-
LSNS-6 The Lubben Social Network Scale-6 cence: Multilevel meta-analyses of cross-sectional and longitudinal associa-
PSSS The Perceived Social Support Scale tions. Dev Psychol. 2019;55(7):1548–65. https://doi.org/10.1037/dev0000719.
10. Watson D, Friend R. Measurement of social-evaluative anxiety. J Consult Clin
Acknowledgements Psychol. 1969;33(4):448–57. https://doi.org/10.1037/h0027806.
We would like to thank all participants in this work as well as nursing home 11. Bai JW, Hao XJ, li CZ, Cheng SC. Social avoidance and distress among
and community managers for their support of this study. middle-aged and older adults with chronic diseases and influencing
factors. Chin J Gerontol. 2021;41(17):3810–3. https://doi.org/10.3969/j.
Author contributions issn.1005-9202.2021.17.048.
In this study, HX, LZ and YW conceived and designed the study. SS, and LW 12. Gum AM, King-Kallimanis B, Kohn R. Prevalence of mood, anxiety, and
designed the questionnaires. LW, JL, JZ and SQ collected the questionnaires. substance-abuse disorders for older americans in the national comorbidity
SS, LW, JL and HL performed the data analyses. SS and YW, LZ and HX wrote survey-replication. Am J Geriatr Psychiatry. 2009;17(9):769–81. https://doi.
the manuscript. All authors read and approved the manuscript. org/10.1097/JGP.0b013e3181ad4f5a.
13. Lim MH, Rodebaugh TL, Zyphur MJ, Gleeson JF. Loneliness over time: the
Funding crucial role of social anxiety. J Abnorm Psychol. 2016;125(5):620–30. https://
The research was supported by the Natural Science Foundation of Zhejiang doi.org/10.1037/abn0000162.
Province [grant number LGF20H250003]; Wenzhou Basic Scientific Research 14. Hoffman YSG, Grossman ES, Bergman YS, Bodner E. The link between social
Project [grant number Y20220105] and the National Social Science Foundation anxiety and intimate loneliness is stronger for older adults than for younger
of China [grant number 22BRK015]. adults. Aging Ment Health. 2021;25(7):1246–53. https://doi.org/10.1080/1360
7863.2020.1774741.
Data availability 15. Cacioppo JT, Cacioppo S. Chapter three - loneliness in the modern age:
The data that support the findings of this study are available from the an evolutionary theory of loneliness (ETL). In: Olson JM, editor. Advances
corresponding author, upon reasonable request. in experimental social psychology. Volume 58. Academic Press; 2018. pp.
127–97.
16. Lieberz J, Shamay-Tsoory SG, Saporta N, Kanterman A, Gorni J, Esser T, et al.
Declarations Behavioral and neural dissociation of social anxiety and loneliness. J Neurosci.
2022;42(12):2570–83. https://doi.org/10.1523/jneurosci.2029-21.2022.
Ethics approval and consent to participate 17. Chau AKC, So SH, Sun X, Zhu C, Chiu CD, Chan RCK, et al. The co-occurrence
This study was approved by the Institutional Review Board of Wenzhou of multidimensional loneliness with depression, social anxiety and paranoia
Medical University (approval number: 2021-011). All methods were performed in non-clinical young adults: a latent profile analysis. Front Psychiatry.
in accordance with the relevant rules and regulations of the Declaration of 2022;13:931558. https://doi.org/10.3389/fpsyt.2022.931558.
Helsinki. Informed consent was obtained from all participants, and they were 18. Danneel S, Nelemans S, Spithoven A, Bastin M, Bijttebier P, Colpin H, et al.
kept anonymous. Internalizing problems in adolescence: linking loneliness, social anxiety
symptoms, and depressive symptoms over Time. J Abnorm Child Psychol.
Consent for publication 2019;47(10):1691–705. https://doi.org/10.1007/s10802-019-00539-0.
Not applicable. 19. Bruce LD, Wu JS, Lustig SL, Russell DW, Nemecek DA. Loneliness in the United
States: a 2018 National Panel Survey of demographic, structural, cognitive,
Competing interests and behavioral characteristics. Am J Health Promot. 2019;33(8):1123–33.
The authors declare no competing interests. https://doi.org/10.1177/0890117119856551.
20. Alden LE, Taylor CT. Interpersonal processes in social phobia. Clin Psychol Rev.
Received: 13 August 2023 / Accepted: 17 January 2024 2004;24(7):857–82. https://doi.org/10.1016/j.cpr.2004.07.006.
21. Kim SSY, Liu M, Qiao A, Miller LC. I want to be alone, but I don’t want to be
lonely: uncertainty management regarding social situations among College
students with social anxiety disorder. Health Commun. 2022;37(13):1650–60.
https://doi.org/10.1080/10410236.2021.1912890.
References 22. Berkman LF. The assessment of social networks and social sup-
1. Barakat MM, Elattar NF, Zaki HN, Depression. Anxiety and loneliness among port in the elderly. J Am Geriatr Soc. 1983;31(12):743–9. https://doi.
Elderly Living in Geriatric homes. Am J Nurs Res. 2019;7(4):400–11. https://doi. org/10.1111/j.1532-5415.1983.tb03393.x.
org/10.12691/ajnr-7-4-1. 23. Donev D, Pavlekovic G, Zaletel-Kragelj L. Social Networks and Social Sup-
2. Kemperman A, van den Berg P, Weijs-Perrée M, Uijtdewillegen K. Loneliness port in Health Promotion Programmes. 2008. https://doi.org/10.2390/
of older adults: Social Network and the living environment. Int J Environ Res biecoll-mhcp4-1.7.
Public Health. 2019;16(3). https://doi.org/10.3390/ijerph16030406. 24. Falk Dahl CA, Dahl AA. Lifestyle and social network in individuals with high
3. Jia Y, Yue Y. Perceived social support mediates loneliness and social isolation: level of social phobia/anxiety symptoms: a community-based study. Soc
a cross-sectional study of Chinese older adults relocated for poverty relief. Int Psychiatry Psychiatr Epidemiol. 2010;45(3):309–17. https://doi.org/10.1007/
J Geriatr Psychiatry. 2023;38(5):e5940. https://doi.org/10.1002/gps.5940. s00127-009-0069-6.
4. Wei K, Liu Y, Yang J, Gu N, Cao X, Zhao X, et al. Living arrangement modi- 25. Teo AR, Lerrigo R, Rogers MA. The role of social isolation in social anxi-
fies the associations of loneliness with adverse health outcomes in older ety disorder: a systematic review and meta-analysis. J Anxiety Disord.
adults: evidence from the CLHLS. BMC Geriatr. 2022;22(1):59. https://doi. 2013;27(4):353–64. https://doi.org/10.1016/j.janxdis.2013.03.010.
org/10.1186/s12877-021-02742-5. 26. Berkman LF, Glass T, Brissette I, Seeman TE. From social integration to health:
5. Surkalim DL, Luo M, Eres R, Gebel K, van Buskirk J, Bauman A, et al. The preva- Durkheim in the new millennium. Soc Sci Med. 2000;51(6):843–57. https://
lence of loneliness across 113 countries: systematic review and meta-analysis. doi.org/10.1016/s0277-9536(00)00065-4.
BMJ. 2022;376:e067068. https://doi.org/10.1136/bmj-2021-067068. 27. Chang H, Zhou J, Wang Z. The impact of social capital on successful ageing
6. Cacioppo S, Grippo AJ, London S, Goossens L, Cacioppo JT. Loneliness: clini- of empty nesters: a cross-sectional study. J Adv Nurs. 2022. https://doi.
cal import and interventions. Perspect Psychol Sci. 2015;10(2):238–49. https:// org/10.1111/jan.15509.
doi.org/10.1177/1745691615570616. 28. Santini ZI, Jose PE, York Cornwell E, Koyanagi A, Nielsen L, Hinrichsen C, et al.
7. Cacioppo JT, Hawkley LC. Perceived social isolation and cognition. Trends Social disconnectedness, perceived isolation, and symptoms of depression
Cogn Sci. 2009;13(10):447–54. https://doi.org/10.1016/j.tics.2009.06.005. and anxiety among older americans (NSHAP): a longitudinal mediation
Sun et al. BMC Public Health (2024) 24:483 Page 11 of 11

analysis. Lancet Public Health. 2020;5(1):e62–e70. https://doi.org/10.1016/ 46. Zheng X, Li H. How Chinese children’s filial piety beliefs affect their parents’
s2468-2667(19)30230-0. life satisfaction and loneliness. Psych J. 2022;11(6):863–72. https://doi.
29. Domènech-Abella J, Mundó J, Switsers L, van Tilburg T, Fernández D, Aznar- org/10.1002/pchj.558.
Lou I. Social network size, loneliness, physical functioning and depressive 47. Sol K, Brauer S, Antonucci TC. Longitudinal associations between loneliness
symptoms among older adults: examining reciprocal associations in four and self-rated Health among Black and White older adults. J Gerontol B Psy-
waves of the Longitudinal Aging Study Amsterdam (LASA). Int J Geriatr chol Sci Soc Sci. 2023;78(4):639–48. https://doi.org/10.1093/geronb/gbac200.
Psychiatry. 2021;36(10):1541–9. https://doi.org/10.1002/gps.5560. 48. Hayes AF. Introduction to mediation, moderation, and conditional process
30. Moreno-Tamayo K, Manrique-Espinoza B, Ramírez-García E, Sánchez-García analysis: A regression-based approach. New York, NY, US: Guilford Press; 2013.
S. Social isolation undermines quality of life in older adults. Int Psychogeriatr. xvii, 507-xvii, p.
2020;32(11):1283–92. https://doi.org/10.1017/s1041610219000310. 49. Podsakoff PM, MacKenzie SB, Lee JY, Podsakoff NP. Common method
31. Zhao D, Zhao M, Wang N, Fu M, Wang A. Relationships among social biases in behavioral research: a critical review of the literature and
isolation, depression, loneliness and quality of life in the community- recommended remedies. J Appl Psychol. 2003;88(5):879–903. https://doi.
dewelling elderly. J Nurs Sci. 2020;35(15):92–5. https://doi.org/10.3870/j. org/10.1037/0021-9010.88.5.879.
issn.101-4152.2020.15.092. 50. Tang DD, Wen ZL. Statistical approaches for testing common method bias:
32. Xiao S, Yang D. The effects of social support on physical and mental health. problems and suggestions. J Psychol Sci. 2020;43(1):215–23. https://doi.
Chin Mental Health J. 1987;04:183–7. org/10.16719/j.cnki.1671-6981.20200130.
33. Liu Y, Aungsuroch Y. Work stress, perceived social support, self-efficacy and 51. Singh A, Misra N. Loneliness, depression and sociability in old age. Ind Psy-
burnout among Chinese registered nurses. J Nurs Manag. 2019;27(7):1445– chiatry J. 2009;18(1):51–5. https://doi.org/10.4103/0972-6748.57861.
53. https://doi.org/10.1111/jonm.12828. 52. Fu TT, Zhang QZ, Shen CZ. Meta-synthesis of qualitative studies on the
34. Hobfoll SE. The influence of culture, community, and the nested-self in the elderly’s experience towards self-perception of aging. Military Nurs.
stress process: advancing conservation of resources theory. Appl Psychology: 2019;36(08):28–31. https://doi.org/10.3969/j.issn.1008-9993.2019.08.007.
Int Rev. 2001;50:337–70. https://doi.org/10.1111/1464-0597.00062. 53. de Jong Gierveld J, van Tilburg T, Dykstra PA. Loneliness and social isolation.
35. Ren Y, Li M. Influence of physical exercise on social anxiety of left-behind In: Vangelisti AL, Perlman D, editors. The Cambridge Handbook of Personal
children in rural areas in China: the mediator and moderator role of perceived Relationships. Cambridge Handbooks in psychology. Cambridge: Cambridge
social support. J Affect Disord. 2020;266:223–9. https://doi.org/10.1016/j. University Press; 2006. pp. 485–500.
jad.2020.01.152. 54. Luo EL, Mei SW, Wu KR, Wu YF, Jin T. A study on the construction model of the
36. Xiao SJ, Shi L, Dong F, Zhang JC, Xue BL, Ouyang P, et al. The impact of new social network of elderly urban immigrants. Urban Insight. 2023;01119–
chronic diseases on loneliness among the older adults: the mediating effect 31. https://doi.org/10.3969/j.issn.1674-7178.2023.01.011.
of cognitive function and moderating role of perceived social support. Mod 55. Rueger SY, Malecki CK, Pyun Y, Aycock C, Coyle S. A meta-analytic review
Prev Med. 2021;48(15):2777–80. of the association between perceived social support and depression in
37. Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: a flexible statistical power childhood and adolescence. Psychol Bull. 2016;142(10):1017–67. https://doi.
analysis program for the social, behavioral, and biomedical sciences. Behav org/10.1037/bul0000058.
Res Methods. 2007;39(2):175–91. https://doi.org/10.3758/bf03193146. 56. Liu XY, Jin YC, An JX. Social support and loneliness in young adults: a
38. Tinsley HEA, Tinsley DJ. Uses of factor analysis in counseling psy- meoderated mediation model. J Psychol Sci. 2020;43(03):586–92. https://doi.
chology research. J Couns Psychol. 1987;34:414–24. https://doi. org/10.16719/j.cnki.1671-6981.20200311.
org/10.1037/0022-0167.34.4.414. 57. Rapee RM, Heimberg RG. A cognitive-behavioral model of anxiety in
39. Wang XD, Wang XL, Ma H. Rating scales for mental health (supplement). social phobia. Behav Res Ther. 1997;35(8):741–56. https://doi.org/10.1016/
Beijing: Chinese Mental Health Journal; 1999. s0005-7967(97)00022-3.
40. Liu YR. Research on the involvement of group work in the community inte- 58. Goodman FR, Kelso KC, Wiernik BM, Kashdan TB. Social comparisons and
gration of old drifters [master]. Dalian Maritime University; 2020. social anxiety in daily life: an experience-sampling approach. J Abnorm
41. Gierveld JDJ, Tilburg TV. A 6-Item scale for overall, emotional, and Social Psychol. 2021;130(5):468–89. https://doi.org/10.1037/abn0000671.
Loneliness:confirmatory tests on Survey Data. Res Aging. 2006;28(5):582–98. 59. Huang WJ, Fu GS, Tan LN, Li CZ, Wei QL, Gao Q et al. Self-esteem and
https://doi.org/10.1177/0164027506289723. life satisfaction of urban empty nesters and influencing factors. Chi-
42. Leung GT, de Jong Gierveld J, Lam LC. Validation of the Chinese translation of nese Journal of Gerontology. 2021;41(06):1326-9. https://doi.org/3969/j.
the 6-item De Jong Gierveld loneliness scale in elderly Chinese. Int Psychoge- issn.1005-9202.2021.06.056
riatr. 2008;20(6):1262–72. https://doi.org/10.1017/s1041610208007552. 60. Henry JD, Coundouris SP, Mead J, Thompson B, Hubbard RE, Grainger SA.
43. Lubben J, Blozik E, Gillmann G, Iliffe S, von Renteln Kruse W, Beck JC, et al. Social Frailty in Late Adulthood: Social Cognitive and Psychological Well-
Performance of an abbreviated version of the Lubben Social Network Scale Being correlates. J Gerontol B Psychol Sci Soc Sci. 2023;78(1):87–96. https://
among three European community-dwelling older adult populations. Geron- doi.org/10.1093/geronb/gbac157.
tologist. 2006;46(4):503–13. https://doi.org/10.1093/geront/46.4.503.
44. Chang Q, Sha F, Chan CH, Yip PSF. Validation of an abbreviated version of the
Lubben Social Network Scale (LSNS-6) and its associations with suicidality Publisher’s Note
among older adults in China. PLoS ONE. 2018;13(8):e0201612. https://doi. Springer Nature remains neutral with regard to jurisdictional claims in
org/10.1371/journal.pone.0201612. published maps and institutional affiliations.
45. Zimet GD, Powell SS, Farley GK, Werkman S, Berkoff KA. Psychometric char-
acteristics of the Multidimensional Scale of Perceived Social Support. J Pers
Assess. 1990;55(3–4):610–7. https://doi.org/10.1080/00223891.1990.9674095.