Associated factors of loneliness among primary school students

Abstract

Loneliness is associated with various negative mental and physical health outcomes. Studies on factors associated with loneliness can inform its early screening and prevention. However, little is known about what factors are associated with loneliness among Chinese young children. The present study aimed to identify the various loneliness-associated factors (demographic, personal, school-related, and family-related) among primary school students in Hong Kong. A total of 258 students and their parents from six primary schools in Hong Kong completed an online anonymous survey from June to October 2020. Loneliness (i.e., UCLA Loneliness Scale 3 total score ≥ 3) was reported by 14% of the students. Multivariable mixed effects logistic regression suggested loneliness was positively associated with a lower happiness level at school, poorer independence skills, a lower level of satisfaction with parents, and lower child-rearing expenditure. There was no clustering effect of school on the associations. The current study found important demographic, personal, school-related, and family-related factors of loneliness among school-age children, with caution suggested in their interpretation considering the cross-sectional nature of this study. Future studies with a larger sample, preferably longitudinal ones, are needed to substantiate these associations and uncover their underlying mechanisms.

Keywords

Child health childhood mental health school closure school environment family factor

Loneliness is the negative emotional response to the discrepancy between one’s desired and actual social connections (Cacioppo & Hawkley, 2009; Perlman & Peplau, 1981). Evidence show that children aged five or six already have a basic understanding of loneliness (Cassidy & Asher, 1992). In fact, loneliness is commonly reported among school-age children, with prevalence estimated to be 14%–20% (Bartels et al., 2008; Lempinen et al., 2018). Childhood loneliness is associated with many negative outcomes, such as conduct and emotional problems (Lempinen et al., 2018), social incompetence (Zhang et al., 2014), and poor academic performance (Asher & Paquette, 2003). It has also been linked to low self-esteem, poor peer acceptance, and unsatisfying relationships (Krause-Parello, 2008; Rokach & Sha’ked, 2013). Childhood loneliness also predicted future loneliness (Bartels et al., 2008), depression and anxiety (Qualter et al., 2010, 2013; Xerxa et al., 2021), suicidal behaviors (Schinka et al., 2012), and poor physical health (Harris et al., 2013) later in adolescence and adulthood.

The mechanisms of childhood loneliness are complicated (Heinrich & Gullone, 2006). Different life stages could be associated with different social needs and sources of loneliness (Heinrich & Gullone, 2006; Qualter et al., 2015). As in adulthood, childhood loneliness is associated with multiple factors in different aspects of children’s life but is particularly susceptible to the influence of developmental or transitional life changes (Lim et al., 2020). At the personal level, Suanet and van Tilburg (2019) highlighted the role of psychological resources (e.g., self-efficacy) in coping with loneliness-associated distress. School and family environments are also particularly relevant to childhood loneliness (e.g., peer friendship and rejection, Qualter et al., 2015; attachment difficulties, Heinrich & Gullone, 2006). Research on childhood loneliness will benefit from examination of a spectrum of personal and environmental factors (Heinrich & Gullone, 2006; Lim et al., 2020).

While empirical evidence is abundant for predictors of loneliness among adolescents and adults (see review by Solmi et al., 2020), less attention has been paid to those among younger children. Findings are mixed concerning the impact of gender and grade level (i.e., Primary 1–6;¹ Schinka et al., 2013; Qualter et al., 2010; Yan et al., 2018). There is emerging evidence for the roles of school and family environments in childhood loneliness (e.g., Arslan, 2021; Palikara et al., 2021; Yan et al., 2018). Yet very few studies have examined the demographic, personal, family-related, and school-related factors of childhood loneliness in the same sample. Additionally, questions remain as to whether and to what extent these factors predict levels of loneliness among Chinese children in Hong Kong.

Experiences of loneliness are shaped by culture (van Staden & Coetzee, 2010). It has been suggested that compared to Western cultures, Chinese culture places a greater emphasis on interdependence, group harmony, and family relationships (Markus & Kitayama, 1991). These cultural differences may in turn affect people’s experiences and understandings of loneliness. For example, in collectivist cultures, which is featured by tighter social networks and more traditional kinship bonds, loneliness seems to be less prevalent (Barreto et al., 2021) and more strongly predicted by absent interactions with family (vs. friends; Lykes & Kemmelmeier, 2014), compared to individualistic cultures. Little evidence is available for the factors of childhood loneliness in Hong Kong, with the only local investigation being conducted more than 20 years ago and including both primary and secondary school students (Lau et al., 1999), which found more severe loneliness among younger boys. In mainland China, childhood loneliness is associated with various demographic (e.g., high economic pressure, Ying et al., 2019), personal (e.g., low self-support, Yao et al., 2022; high perceived stress, Kuo et al., 2021), and interpersonal (e.g., poor relationships with parents or peers, Ying et al., 2019; Zhang et al., 2014) factors. Again, these is a lack of research examining these factors in one group of children in the cultural background of Hong Kong. There is a pressing need to update this field of research with a wider spectrum of factors of childhood loneliness examined.

Upon the outbreak of the coronavirus disease 2019 (COVID-19), the Hong Kong government implemented strict pandemic control measures including city-wide school closures. In primary schools, it was not until 2023, February 15 that whole-day in-person classes were fully resumed (Muthanna, 2022). Loneliness due to social isolation can have significant negative consequences among children that could last for up to nine years (Loades et al., 2020). There is initial evidence for the residual and long-lasting effect of pandemic-related lockdown on mental health (Ambrosetti et al., 2021; Ma et al., 2022). It is thus crucial to identify the predictors of loneliness among younger children during a period of pandemic-related social isolation, such as school closure.

Considering the substantial and potentially long-term impact of pandemic-related loneliness on younger children, a study that examines a range of personal and environmental predictors of childhood loneliness would be fitting, as it will not only help identify children at higher risk for loneliness but also enlighten the development of psychosocial interventions for loneliness in this population in the post-COVID era. The current study aimed to address this research gap by exploring the contribution of a range of demographic, personal, school-related, and family-related factors to loneliness among primary school students in Hong Kong.

Method

Participants

This is a secondary analysis on the dataset of a school-based cross-sectional study conducted from June to October 2020, a period of mandatory school closure (Zheng et al., 2022). Following institutional ethics committee approval, an invitation email was sent to 165 randomly selected primary schools (including government, aided, Direct Subsidy Scheme [DSS], and private schools) introducing the aim, content, timeframe, funding source, confidentiality, and contact information of the research project and asking if they were willing to participate, followed by at least three email reminders and phone calls. Of all invited schools, six (three aided, two DSS, and one private) agreed to participate in the study. The schools then sent out invitation emails to all the students and their parents that briefly introduced the research project and provided the QR codes and weblinks to access the online questionnaires (including one parent questionnaire and one child questionnaire) and the instructions for completing the questionnaires. All participants filled in the survey anonymously and voluntarily. No compensation was provided for their participation.

Table 1 summarizes the characteristics of all participating child-parent pairs. A total of 258 primary school students were included in the study (205 boys: 79.5%) with an average age of 9.00 (SD = 1.94, median = 9.02, range = 5.98–12.09), among whom about half (n = 133) came from Primary 1–3 (age: M = 7.35, SD = 0.94) with the rest coming from Primary 4–6 (age: M = 10.80, SD = 0.82). The proportion of students showing loneliness (i.e., UCLA-3 score ≥3) was 14.0% in the current sample, which is consistent with that found in previous studies (14%–20%; Bartels et al., 2008; Lempinen et al., 2018).

Table 1.

Participant characteristics (n = 258).

	n	%
Gender^a
Boy	205	79.5
Girl	53	20.5
Age^b (M, SD)	9.00	1.94
Grade level^a
Low (Primary 1–3)	133	51.6
High (Primary 4–6)	125	48.5
Participating parent’s relationship to the child^b
Biological mother	214	83.0
Biological father	41	15.9
Others	3	1.2
Participating parent’s marital status^b
Married	251	97.3
Divorced/separated/never married	7	2.7
Mother employment status^b
Full-time employee	99	38.4
Part-time employee/home-maker/others	159	61.6
Father employment status^b
Full-time employee	189	73.3
Part-time employee/home-maker/others	69	26.7
Mother education level^b
Master’s degree or above	80	31.1
Associate/diploma/bachelor’s degree	98	38.1
Secondary school or below	79	30.7
Father education level^b
Master’s degree or above	87	34.0
Associate/diploma/bachelor’s degree	93	36.3
Secondary school or below	76	29.7
Number of persons residing in the household^b (M, SD)	4.28	1.14
Number of caregivers for the child^b (M, SD)	1.79	0.95
Have siblings^b	161	62.4
Monthly household income^b
Below $10,000	6	2.3
$10,000–39,999	69	26.7
$40,000–69,999	41	15.9
$70,000–99,999	32	12.4
$100,000 or above	110	42.6
Expenses incurred for the child over the past year^b
Below $20,000	38	15.3
$20,000–99,999	97	39.1
$100,000–199,999	70	28.2
$200,000–299,999	23	9.3
$300,000 or above	20	8.1

^aReported by students.

^bReported by parents.

Instrument

An expert panel comprising researchers and practitioners in public health, education, and journalism designed the online survey, which consisted of a child questionnaire and a parent questionnaire. A Chinese preliminary version was first developed and then independently reviewed and commented on by each panel member. After several rounds of discussion and revision, the panel agreed upon a shorter final version that could be completed within 5–10 minutes. Considering younger children might have limited literacy skills and need cognitive scaffolding when responding to the questionnaire, audios of survey questions were available to Primary 1–3 students; colorful pictures/emojis were also embedded in the questionnaire. To cater to the bilingual educational needs in Hong Kong, the Chinese questionnaire was translated into English by two bilingual researchers and checked by other two with experiences in public health and education. Several rounds of revision and proofreading were conducted on the English version to ensure the equivalency of all questions and response options to those in the Chinese version. A pilot study involving 45 primary school students (at least five came from each of the six grade levels) and their parents suggested good survey acceptability. The survey was put into use in the formal study after minor revisions made with feedback collected in the pilot.

The survey was built in a professional online survey platform (QuestionPro) and could be completed using electronic devices including computers, tablets, and smartphones. Separate weblinks and QR codes were provided for the child and parent questionnaires. The students could fill in the child questionnaire either by themselves or with the assistance of their parents if needed. Before starting the survey, participants were asked to give informed consent. They could begin the survey by clicking “Yes” and terminate it by clicking “No”.

Measurement

Study outcome

Childhood loneliness was measured with the UCLA Loneliness Scale 3 (UCLA-3; Hughes et al., 2004). Students rated their feelings of loneliness on a three-point Likert scale (0 = almost never, 1 = sometimes, 2 = always; Joshi et al., 2015). The total score was computed by summing up all three items and ranged from zero to six. Since the total scores were positively skewed, we followed previous practices (Liu et al., 2020; Steptoe et al., 2013) and dichotomized the variable based on the total score of three (not lonely: < 3; lonely: ≥ 3). For students from Primary 1–3, the Chinese question “How often do you feel isolated from others?” was rephrased into “Do you believe that you have no friends and your classmates do not want to play with you?” for better comprehension by younger students. The UCLA-3 in this study showed acceptable internal reliability in the full sample and Primary 1–3 and 4–6 students (Cronbach’s α = .78, .70, and .81, respectively).

Independent variables

Potential demographic, personal, school-related, and family-related factors of loneliness were assessed with a range of rigorously designed single items (see “Instrument” above). For categorical variables on five- or three-point Likert-scales, the points were compressed to ensure a sufficient number in each cross-tabulation group. Continuous variables entered into analyses were dichotomized based on the mean total score for the ease of interpretation. Our preliminary univariable logistic regression analyses yielded 22 significant predictors of loneliness (see Supplemental Table 1 for complete results). Among them, thirteen had a less clear theoretical relationship with loneliness, were conceptually similar to other variables, had problems of multicollinearity (i.e., variance inflation factors [VIFs] ≥ 10), yielded a poorer model fit, or were reported by a smaller sample. These variables were removed from the subsequent analyses. The following nine variables served as the independent variables (IVs) of the present study:

A. Demographic factors:

(1) Student’s self-reported gender,

(2) Student’s self-reported grade level (Primary 1–6),

(3) Parent-reported child-rearing expenditure (“What is the approximate amount of the expenses incurred for your child in the past year in your household, including tuition fees, living expenses, medical expenses, insurance expenses and all other relevant expenses?”; in Hong Kong dollars),

B. Personal factors:

(4) Parent-reported student’s levels of independence skills (“In your opinion, your child’s independence skills, such as the ability to independently arrange and deal with his/her daily life, learning and playing activities, are: 1 = very good, 2 = quite good, 3 = average, 4 = quite poor, 5 = very poor, 6 = hard to say”),

(5) Parent-reported levels of pressure the student was facing in various aspects (“Do you think the pressure faced by your child in various aspects is high? 1 = very low, 2 = low, 3 = average, 4 = high, 5 = very high”),

C. School-related factors:

(6) Student’s self-reported happiness level at school (“If you are to give a mark, the mark representing the happiness you feel at school is ___ mark(s)”; ranged from 0 to 10, a higher score indicating a higher level of happiness),

(7) Student’s self-reported relationship with classmates (“Do you have a good relationship with your classmates? 1 = very bad, 2 = quite bad, 3 = average, 4 = quite good, 5 = very good”),

D. Family-related factors:

(8) Student’s self-reported happiness level at home (“If you are to give a mark, the mark representing the happiness you feel at home is ___ mark(s)”; ranged from 0 to 10, a higher score indicating a higher level of happiness),

(9) Student’s self-reported level of satisfaction with parents (“Overall speaking, what is your degree of satisfaction with your father and mother? 1 = very dissatisfied, 2 = not quite satisfied, 3 = half-and-half, 4 = satisfied, 5 = very satisfied”).

Statistical analysis

The Stata 16 software (StataCorp, 2019) was used for statistical analyses. Frequency and percentage were used to describe categorical variables, and M and SD were used for continuous variables. Given that our participating students mainly came from six primary schools in Hong Kong and that important school-related IVs were examined, our data had a hierarchical structure, which violated the independence assumption of traditional logistic regression models. Therefore, mixed effects logistic regression models were fitted to test the association between IVs and loneliness using school² as a random variable. To test the significance of clustering effect by school, the intra-class Correlation Coefficient (ICC), Likelihood Ratio (LR) test, Median Odds Ratio (MOR), and cluster variance were estimated. The fit of the mixed effects versus traditional logistic regression model was assessed using deviance (-2LLR), Akakie Information Criteria (AIC), and Bayesian Information Criteria (BIC).

A series of univariable mixed effects logistic regression analyses were performed to assess the association between each IV and levels of loneliness, and results were presented as crude odds ratios (CORs; odds ratios with only one variable entered in the model at a time) with 95% confidence intervals (CIs). Only IVs that significantly predicted loneliness in the univariable mixed effects regression analyses (p < .05) entered a multivariable mixed effects logistic regression model to examine their associations with loneliness, with adjusted odds ratios (AORs; odds ratios that control for other variables in the model) and 95% CIs presented. No multicollinearity was observed among the IVs (VIFs < 10).

Additional analyses on a larger sample

Since the study explored a wide variety of (including parent-reported) correlates of childhood loneliness, the sample was limited to a small, matched child-parent pairs (i.e., paired sample, n = 258). However, there was a larger dataset with student but not pairable parent responses (i.e., student-only sample, n = 781). To confirm our findings with a larger sample size, we conducted additional analyses to test the associations between loneliness and all student-reported (11 of all 22) IVs selected for the current study³ in the student-only sample. Only two demographic variables (gender, grade level) were reported by students and available for this larger sample (Supplemental Table 2). The student-only sample had a higher percentage of girls (32.8% vs. 20.5%) but a similar percentage of Primary 1–3 students (50.1% vs. 48.4%), compared to the paired sample.

Results

Predictors of childhood loneliness in the paired sample

Results of univariable mixed effects logistic regression analyses were summarized in Table 2. Loneliness was significantly predicted by being a girl, coming from Primary 4–6, lower child-rearing expenditure over the past year, poorer independence skills, a higher level of pressure, a lower happiness level at school, poorer relationships with classmates, a lower level of happiness at home, and a lower level of satisfaction with parents. These variables were entered into the multivariable mixed effects logistic regression model.

Table 2.

Factors associated with loneliness, prevalence of loneliness, and results of univariable and multivariable mixed effects logistic regression analyses (n = 258).

	Prevalence of loneliness, n (%)		COR		AOR^a
	Not lonely	Lonely	M (95%CI)	p-value	M (95%CI)	p-value
Demographic and personal factors
Gender^b
Boy	183 (89.3)	22 (10.7)	Ref		Ref
Girl	39 (73.6)	14 (26.4)	2.99 (1.40, 6.35)	.004**	1.72 (0.61, 4.80)	.303
Grade level^b
High (Primary 1–3)	123 (92.5)	10 (7.5)	Ref		Ref
Low (Primary 4–6)	99 (79.2)	26 (20.8)	3.09 (1.40, 6.84)	.005**	2.43 (0.91, 6.49)	.075
Expenses incurred for the child over the past year^c
HK$100,000 or above	106 (93.8)	7 (6.2)	Ref		Ref
Below HK$100,000	108 (80.0)	27 (20.0)	3.57 (1.41, 9.06)	.007**	3.15 (1.08, 9.09)	.036*
Independence skills^c
Very good/Quite good	99 (93.4)	7 (6.6)	Ref		Ref
Average	95 (84.8)	17 (15.2)	2.49 (0.98, 6.34)	.055	1.86 (0.63, 5.54)	.259
Quite poor/Very poor/Hard to say	28 (70.0)	12 (30.0)	5.59 (1.97, 15.87)	.001**	4.08 (1.16, 14.35)	.029*
Levels of pressure^c
Average/Low/Very low	184 (89.3)	22 (10.7)	Ref		Ref
Very high/High	38 (73.1)	14 (26.9)	2.85 (1.30, 6.22)	.009**	2.59 (0.98, 6.79)	.054
School-related factors
Happiness at school^b
9–10	161 (93.6)	11 (6.4)	Ref		Ref
0–8	61 (70.9)	25 (29.1)	6.18 (2.82, 13.55)	<.001***	3.13 (1.20, 8.18)	.020*
Relationship with classmates^b
Very good/Good	212 (88.3)	28 (11.7)	Ref		Ref
Average/Bad	10 (55.6)	8 (44.4)	5.90 (2.08, 16.70)	.001**	2.07 (0.48, 8.87)	.328
Family-related factors
Happiness at home^b
9–10	161 (91.0)	16 (9.0)	Ref		Ref
0–8	61 (75.3)	20 (24.7)	3.22 (1.55, 6.68)	.002**	1.76 (0.67, 4.62)	.253
Satisfaction with parents^a
Very satisfied/Satisfied	199 (89.6)	23 (10.4)	Ref		Ref
Half-and-half/Not quite satisfied/Very dissatisfied	23 (63.9)	13 (36.1)	4.71 (2.06, 10.78)	<.001***	3.13 (1.11, 8.84)	.032*

Note. AOR = adjusted odds ratio. CI = confidence interval. COR = crude odds ratio.

^aSample size for the multivariable logistic regression model n = 248.

^bReported by students.

^cReported by parents.

*p < .05, **p < .01, ***p < .001.

In the multivariable model that included only univariably significant IVs, loneliness remained significantly associated with a lower happiness level at school (AOR = 3.13, 95% CI 1.20 to 8.18, p = .020), poorer independence skills (quite poor/very poor/hard to say: AOR = 4.08, 95% CI 1.16 to 14.35, p = .029), a higher level of satisfaction with parents (AOR = 3.13, 95% CI 1.11 to 8.84, p = .032), and lower child-rearing expenditure (AOR = 3.15, 95% CI 1.08 to 9.09, p = .036; Table 2). The ICC value and cluster variance of 0.00, the MOR value of 1.00, and the non-significant LR-test (p = .000) revealed little variation between schools. The mixed effects logistic regression using school as the clustering variable showed a similar model fit (deviance, AIC, BIC) to that yielded by traditional logistic regression (Table 3).⁴ Therefore, there was a limited impact of school on the associations between IVs and loneliness.

Table 3.

Model comparison between traditional and mixed effects logistic regression analyses (n = 248).

	Traditional regression analysis	Mixed effects regression analysis
LLR	−69.29	−69.29
AIC	164.28	162.58
BIC	209.95	204.74
Deviance	138.58	138.58
ICC		0.00
LR-test	LR test versus logistic model: $χ^{2}$ (1) = 0.00, p = .000
MOR		1.00
Cluster variance		0.00

Note. AIC = Akaike Information Criteria. BIC = Bayesian Information Criteria. ICC = Intra-class Correlation Coefficient. LLR = log-likelihood ratio. LR-test = Likelihood Ratio test. MOR = Median Odds Ratio.

Predictors of childhood loneliness in the student-only sample

Results of the additional analyses on the larger, student-only sample were summarized in Supplemental Table 3. All IVs significantly predicted loneliness in the univariable mixed effects regression model. In the multivariable mixed effects model with all IVs entered, loneliness was significantly associated with a lower happiness level at school and a higher level of satisfaction with parents, confirming our findings in the smaller, paired sample. Loneliness was also significantly associated with being a girl, coming from Primary 4–6, a lower level of perceived safety at school (not quite/absolutely not), stronger intention of school transfer (strongly agree/agree/average), poorer relationships with classmates, and a lower happiness level at home. There was little clustering effect by school (Supplemental Table 4).

Discussion

Loneliness among younger children is associated with a series of severe negative psychological and physical health consequences (Schinka et al., 2012; Xerxa et al., 2021). Particularly, loneliness due to COVID-19-related social isolation can have a substantial, long-lasting impact on children (Ambrosetti et al., 2021; Loades et al., 2020; Ma et al., 2022). This study examined a wide range of demographic, personal, school-related, and family-related factors of loneliness in a group of primary school students in Hong Kong. We found loneliness was significantly predicted by a lower happiness level at school and a lower level of satisfaction with parents as reported by the students. Loneliness was also significantly associated with poorer independence skills and lower child-rearing expenditure as reported by the parents. The associations of loneliness and the foregoing factors did not differ across participating schools. Some of our results were confirmed in a larger, student-only sample. These findings were discussed in turn below.

A lower level of happiness at school was associated with loneliness. Our finding was in line with the literature that suggested school dislike (Rönkä et al., 2017), low school belongingness (Arslan, 2021; Palikara et al., 2021), and victimization at school (Bayat et al., 2021) significantly predicted loneliness among the youth. We were not able to determine the direction of this association, given the cross-sectional nature of the present study. Children who were lonely might find their school life less happy. Conversely, a happier school life could serve as a protector of loneliness. Notably, happiness level at school remained a strong predictor of loneliness in the multivariable model, whereas relationship with classmates did not. Possibly childhood loneliness was influenced by the school environment as a whole, with the contribution of not only peer relationship but also other factors (e.g., student-teacher relationship, academic adjustment; Bayat et al., 2021; Rokach & Sha’ked, 2013). A concern is that each participating school in the current study might be associated with a different level of students-perceived happiness, which could bias its relationship found with loneliness. However, the mixed effects logistic regression analyses revealed little school effect in the associations between loneliness and the IVs (including happiness level at school), suggesting consistent findings across schools.

Weaker independence skills significantly predicted loneliness. Again, we were unable to conclude on the direction of this relationship due to the cross-sectional design of the study. Poor independence skills might result from high loneliness, considering the negative impact of loneliness on social functioning and health (Harris et al., 2013; Zhang et al., 2014). Alternatively, stronger independence skills might alleviate children’s feelings of loneliness. Our results confirmed the previous finding that among a group of Chinese primary schoolers, lower loneliness was significantly associated with a higher level of self-support (i.e., the process of children becoming independent of what they have depended on; Yao et al., 2022). A possibility is that children with stronger independence skills were more active and resourceful in (i) building social bonds and (ii) developing a more positive attitude towards solitary activities and thus experienced less loneliness (Henning et al., 2021; Yao et al., 2022). Our finding, if replicated in further research, may suggest independence skill training as a beneficial component of loneliness prevention programs for children.

Another significant predictor of loneliness was a lower level of satisfaction with parents. The critical role of parents in children’s experience of loneliness has been emphasized (Heinrich & Gullone, 2006), particularly in Chinese cultures where family relationships are highly valued (Markus & Kitayama, 1991). Our finding added to the literature highlighting a satisfying relationship with parents in protecting Chinese children against loneliness (Yan et al., 2018; Ying et al., 2019). It is intriguing that loneliness was not significantly predicted by students’ perceived level of happiness at home. It is possible that a happy family life had a positive effect on children’s mental health in general, whereas more satisfaction with parents was specifically related to the fulfilling of social needs at home (i.e., not lonely). The relative contribution of these two constructs to children’s experience of loneliness should be clarified in future research.

There was a significant association between lower child-rearing expenditure and loneliness. Interestingly, our preliminary analyses found no association between loneliness and household income, which contrasted previous findings that higher family income predicted lower loneliness (Schinka et al., 2012; Sorhagen & Wurster, 2017). The child-rearing expenditure depends on not only the amount of income but also how it is spent (Bianchi et al., 2004), which could be a better indicator of how much attention is paid to the child and therefore a more proximal factor of child well-being. More research is needed to clarify the relative importance of these two factors to childhood loneliness.

We did not find an association of loneliness with gender or grade level. It has been suggested that a gender difference in loneliness does not appear in childhood but emerges later during adolescence, with adolescent boys being lonelier (Koenig & Abrams, 1999). Our results supported this notion by showing similar levels of loneliness among primary school boys and girls. The absence of gender difference in the current sample might also stem from the imbalanced gender ratio in the current sample, where girls were underrepresented (20.5% vs. 48.0% for primary school students in Hong Kong; Hong Kong Education Bureau, 2021). Previous findings regarding the relationship between grade level and loneliness were mixed (e.g., Qualter et al., 2010; Yan et al., 2018), which warrants further research.

Additional analyses in a larger, student-only sample indicated the significance of lower levels of happiness at school and satisfaction with parents in predicting loneliness, which confirmed our findings in the paired sample. Whereas in this larger sample, loneliness among children was also associated with other factors (e.g., being a girl, coming from Primary 4–6). The aforementioned factors might become less predictive of loneliness when other (probably more significant) parent-reported variables (e.g., child rearing expenditure) were controlled for. It is also possible that the small sample size of paired sample limited the statistical power of the multivariable model. Finally, the student-only sample had a relatively more balanced gender ratio (percentage of girls: 32.8%). That said, implications drawn from findings in the student-only sample were limited, with few demographic information available (e.g., age, household income) and lack of control for parent-reported factors in the statistical analyses.

The current study has several strengths. First, this study is one of the first investigations on loneliness-associated factors among younger children in Hong Kong. Second, we examined primary school students from all grade levels and factors that might be specific to childhood loneliness. By collecting both children and parent responses, we were able to explore a wide range of aspects of children’s lives (personal, school, family). Additional analyses in a larger sample partly confirmed our main findings, with the caveat of no parent-reported outcomes in the model. Third, mixed effects logistic analyses were conducted to account for school effect on the hypothesized associations, and results suggested that our findings were consistent for students from different schools. Finally, the present study offered valuable evidence for the correlates of children’s mental health during the COVID-19 pandemic, which may continue to contribute to the long-term psychological impact of the pandemic.

This study has several limitations. First, due to difficulties in collecting matched parent-child responses during the pandemic, we had a small sample. Compared to the primary school student population in Hong Kong, our sample was characterized by fewer girls (20.5% vs. 48.0%) and an older age (9 vs. 8.36 years), despite a similar percentage of Primary 1–3 students (51.6% vs. 50.2%; Hong Kong Education Bureau, 2021). Our additional analyses on student-only sample provided a larger sample with a more balanced gender ratio but lacked important parent-reported information. The generalizability of our results was thus limited. This exploratory study nevertheless rendered an important reference for future studies with a larger, paired parent-child sample with a more balanced gender ratio. Future studies may also collect more information about children’s identity (e.g., gender identity, disability, sexual orientation).

A second limitation is that our respondents, recruited through convenience sampling, probably had a better mental health status (i.e., self-selection; Farrokhi & Mahmoudi-Hamidabad, 2012), which might produce an underestimated proportion of students exhibiting loneliness. Third, except for loneliness, all included variables were measured with single questions rather than well-structured scales. That said, brief and direct single questions could be more comprehensible and easily answered by primary schoolers. Single-item measures of the included variables have also been used previously and showed good reliability (e.g., Fülöp et al., 2020; Herman et al., 2018). Fourth, the scores of loneliness and IVs were dichotomized to address the skewed score distribution or for the ease of interpretation, which might have led to information loss or lower statistical power (Erees, 2022). Finally, the cross-sectional nature of this study cautioned against conclusions of causal relationship, which should be validated in further qualitative and longitudinal studies.

Our findings hold significant practical implications for loneliness prevention among younger children, particularly for Chinese and other Asian populations. Our data, although collected during a period of pandemic-associated school closure, could be informative to treatments that target the residual psychological impact of the pandemic or loneliness due to social isolation in general. First, our study highlighted the importance of a happy, caring, and warm school environment. School-based universal interventions have been successful in promoting mental well-being among students in and out of Hong Kong (Durlak et al., 2011; Lai et al., 2016; Shek & Zhu, 2020), which may show promise in preventing loneliness in primary schools. Second, continued attention should be paid to improving children’s satisfaction with their parents, which seems to be a particularly strong protector of childhood loneliness inside a family. Third, if supported by empirical evidence, independence skill training might be an effective component of loneliness prevention programs for primary school students.

The present study provided several important research directions. First, the contributions of different aspects of school and family life to childhood loneliness remain to be clarified. Second, child-rearing expenditure (relative to household income) might better capture the effect of family support on loneliness, which should be considered in future investigations. Third, efforts should be made to investigate the loneliness-associated factors found in the current study in the post-COVID context. Overall, future studies may benefit from a longitudinal design with a larger sample size to further examine the interplay of various loneliness-associated factors in this study to extend our findings.

Conclusion

The present study provided valuable evidence for the correlates of loneliness among primary school students in Hong Kong. Several important predictors of loneliness were identified, including a lower happiness level at school, poorer independence skills, a lower level of satisfaction with parents, and lower child-rearing expenditure. Our findings hold significant implications for childhood loneliness research and prevention. Further research with a larger sample, particularly longitudinal studies, is needed to clarify the contribution of these factors to loneliness for high-risk child screening and targeted intervention designs.

Supplemental Material

Supplemental Material - Associated factors of loneliness among primary school students

Supplemental Material for Associated factors of loneliness among primary school students by Wenyue Wang, Jiawen Liang, Dexing Zhang, Zijun Xu, Dicken C. C. Chan, Grace Yaojie Xie, Yang Gao, Lu Niu, Elsa Lau and Samuel Y. S. Wong in Journal of Social and Personal Relationships.

Footnotes

Authors’ contribution

WW contributed to the first draft, finalized the statistical analysis, and wrote the manuscript. JL contributed to an early draft and preliminary data analysis. DZ conceived the study, participated in study design and coordination, and revised the manuscript. ZX and DC contributed to the data cleaning and analysis. SW contributed to revision and overall supervision. All authors have read, revised, and approved the final manuscript.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Social Innovation Research Collaboration Platform (SIRCP) Fund under the British Council.

Ethical statement

Open research statement

As part of IARR’s encouragement of open research practices, the authors have provided the following information: This research was not pre-registered. The data and materials used in the research are available. Access to data can be obtained by contacting the corresponding author at zhangdxdaisy@cuhk.edu.hk.

ORCID iD

Dexing Zhang

Supplemental Material

Supplemental material for this article is available online.

Notes

References

Ambrosetti

Macheret

Folliet

Wullschleger

Amerio

Aguglia

Serafini

Prada

Kaiser

Bondolfi

Sarasin

Costanza

(2021). Psychiatric emergency admissions during and after COVID-19 lockdown: Short-term impact and long-term implications on mental health. BMC Psychiatry, 21(1), 465. https://doi.org/10.1186/s12888-021-03469-8

Arslan

(2021). School belongingness, well-being, and mental health among adolescents: Exploring the role of loneliness. Australian Journal of Psychology, 73(1), 70–80. https://doi.org/10.1080/00049530.2021.1904499

Asher

S. R.

Paquette

J. A.

(2003). Loneliness and peer relations in childhood. Current Directions in Psychological Science, 12(3), 75–78. https://doi.org/10.1111/1467-8721.01233

Barreto

Victor

Hammond

Eccles

Richins

M. T.

Qualter

(2021). Loneliness around the world: Age, gender, and cultural differences in loneliness. Personality and Individual Differences, 169(3), 110066. https://doi.org/10.1016/j.paid.2020.110066

Bartels

Cacioppo

J. T.

Hudziak

J. J.

Boomsma

D. I.

(2008). Genetic and environmental contributions to stability in loneliness throughout childhood. American Journal of Medical Genetics, 147(3), 385–391. https://doi.org/10.1002/ajmg.b.30608

Bayat

Fokkema

Mujakovic

Ruiter

R. A.

(2021). Contextual correlates of loneliness in adolescents. Children and Youth Services Review, 127(3), 106083. https://doi.org/10.1016/j.childyouth.2021.106083

Bianchi

S. M.

Cohen

Raley

Nomaguchi

(2004). Inequality in parental investment in child-rearing: Expenditures, time, and health. In Neckerman

(Ed.), Social inequality (pp. 189–220). Russell Sage Foundation Press.

Cacioppo

J. T.

Hawkley

L. C.

(2009). Perceived social isolation and cognition. Trends in Cognitive Sciences, 13(10), 447–454. https://doi.org/10.1016/j.tics.2009.06.005

Cassidy

Asher

S. R.

(1992). Loneliness and peer relations in young children. Child Development, 63(2), 350–365. https://doi.org/10.1111/j.1467-8624.1992.tb01632.x

10.

Durlak

J. A.

Weissberg

R. P.

Dymnicki

A. B.

Taylor

R. D.

Schellinger

K. B.

(2011). The impact of enhancing students’ social and emotional learning: A meta-analysis of school-based universal interventions. Child Development, 82(1), 405–432. https://doi.org/10.1111/j.1467-8624.2010.01564.x

11.

Erees

(2022). Effects of dichotomizing continuous outcome on efficiencies of measures of explained variation in logistic regression: Simulation study and application. Communications in Statistics: Case Studies, Data Analysis and Applications, 8(4), 663–681. https://doi.org/10.1080/23737484.2022.2139019

12.

Farrokhi

Mahmoudi-Hamidabad

(2012). Rethinking convenience sampling: Defining quality criteria. Theory and Practice in Language Studies, 2(4), 784. https://doi.org/10.4304/tpls.2.4.784-792

13.

Fülöp

Bőthe

Gál

É.

Cachia

J. Y. A.

Demetrovics

Orosz

(2020). A two-study validation of a single-item measure of relationship satisfaction: RAS-1. Current Psychology, 41(4), 2109–2121. https://doi.org/10.1007/s12144-020-00727-y

14.

Harris

R. A.

Qualter

Robinson

S. J.

(2013). Loneliness trajectories from middle childhood to pre-adolescence: Impact on perceived health and sleep disturbance. Journal of Adolescence, 36(6), 1295–1304. https://doi.org/10.1016/j.adolescence.2012.12.009

15.

Heinrich

L. M.

Gullone

(2006). The clinical significance of loneliness: A literature review. Clinical Psychology Review, 26(6), 695–718. https://doi.org/10.1016/j.cpr.2006.04.002

16.

Henning

Segel-Karpas

Bjälkebring

Berg

A. I.

(2021). Autonomy and loneliness - longitudinal within- and between-person associations among Swedish older adults. Aging & Mental Health, 26(12), 2416–2423. https://doi.org/10.1080/13607863.2021.2000937

17.

Herman

K. C.

Hickmon-Rosa

J. E.

Reinke

W. M.

(2018). Empirically derived profiles of teacher stress, burnout, self-efficacy, and coping and associated student outcomes. Journal of Positive Behavior Interventions, 20(2), 90–100. https://doi.org/10.1177/1098300717732066

18.

Hughes

M. E.

Waite

L. J.

Hawkley

L. C.

Cacioppo

J. T.

(2004). A short scale for measuring loneliness in large surveys: Results from two population-based studies. Research on Aging, 26(6), 655–672. https://doi.org/10.1177/0164027504268574

19.

Joshi

Kale

Chandel

Pal

D. K.

(2015). Likert scale: Explored and explained. British Journal of Applied Science & Technology, 7(4), 396–403. https://doi.org/10.9734/bjast/2015/14975

20.

Koenig

L. J.

Abrams

R. F.

(1999). Adolescent loneliness and adjustment: A focus on gender differences. In Rotenberg

K. J.

Hymel

(Eds.), Loneliness in childhood and adolescence (pp. 296–322). Cambridge University Press.

21.

Hong Kong Education Bureau . (2021). Student enrolment statistics, 2020/21 (Kindergartens, primary and secondary schools). Hong Kong Education Bureau

22.

Krause-Parello

C. A.

(2008). Loneliness in the school setting. The Journal of School Nursing, 24(2), 66–70. https://doi.org/10.1177/1059840508.0240020301

23.

Kuo

B. C. H.

Huang

Lin

(2021). Self-esteem, resilience, social support, and acculturative stress as predictors of loneliness in Chinese internal migrant children: A model-testing longitudinal study. Journal of Psychology, 155(4), 387–405. https://doi.org/10.1080/00223980.2021.1891854

24.

Lai

E. S.

Kwok

C. L.

Wong

P. W.

K. W.

Law

Y. W.

Yip

P. S.

(2016). The effectiveness and sustainability of a universal school-based programme for preventing depression in Chinese adolescents: A follow-up study using quasi-experimental design. PLoS One, 11(2), e0149854. https://doi.org/10.1371/journal.pone.0149854

25.

Lau

Chan

D. W.

Lau

P. S.

(1999). Facets of loneliness and depression among Chinese children and adolescents. The Journal of Social Psychology, 139(6), 713–729. https://doi.org/10.1080/00224549909598251

26.

Lempinen

Junttila

Sourander

(2018). Loneliness and friendships among eight-year-old children: Time-trends over a 24-year period. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 59(2), 171–179. https://doi.org/10.1111/jcpp.12807

27.

Lim

M. H.

Eres

Vasan

(2020). Understanding loneliness in the twenty-first century: An update on correlates, risk factors, and potential solutions. Social Psychiatry and Psychiatric Epidemiology, 55(7), 793–810. https://doi.org/10.1007/s00127-020-01889-7

28.

Liu

C. H.

Zhang

Wong

G. T. F.

Hyun

Hahm

H. C.

(2020). Factors associated with depression, anxiety, and PTSD symptomatology during the COVID-19 pandemic: Clinical implications for U.S. young adult mental health. Psychiatry Research, 290, 113172. https://doi.org/10.1016/j.psychres.2020.113172

29.

Loades

M. E.

Chatburn

Higson-Sweeney

Reynolds

Shafran

Brigden

Linney

McManus

M. N.

Borwick

Crawley

(2020). Rapid systematic review: The impact of social isolation and loneliness on the mental health of children and adolescents in the context of COVID-19. Journal of the American Academy of Child & Adolescent Psychiatry, 59(11), 1218–1239.e3. https://doi.org/10.1016/j.jaac.2020.05.009

30.

Lykes

V. A.

Kemmelmeier

(2014). What predicts loneliness? Cultural difference between individualistic and collectivistic societies in Europe. Journal of Cross-Cultural Psychology, 45(3), 468–490. https://doi.org/10.1177/0022022113509881

31.

Yao

Hao

(2022). The lasting effects of COVID-19 lockdowns on mental health. Applied Economics Letters, 31(3), 247–258. https://doi.org/10.1080/13504851.2.022.2130864

32.

Markus

H. R.

Kitayama

(1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98(2), 224–253. https://doi.org/10.1037/0033-295X.98.2.224

33.

Muthanna

(2022, December 29). Hong Kong schools to resume full-day in-person classes, Starting February 1, 2023. The HK Hub. https://thehkhub.com/hong-kong-schools-to-resume-full-day-in-person-classes-starting-february-1-2023/

34.

Palikara

Castro-Kemp

Gaona

Eirinaki

(2021). The mediating role of school belonging in the relationship between socioemotional well-being and loneliness in primary school age children. Australian Journal of Psychology, 73(1), 24–34. https://doi.org/10.1080/00049530.2021.1882270

35.

Perlman

Peplau

L. A.

(1981). Toward a social psychology of loneliness. Personal Relationships, 3, 31–56.

36.

Qualter

Brown

S. L.

Munn

Rotenberg

K. J.

(2010). Childhood loneliness as a predictor of adolescent depressive symptoms: An 8-year longitudinal study. European Child & Adolescent Psychiatry, 19(6), 493–501. https://doi.org/10.1007/s00787-009-0059-y

37.

Qualter

Brown

S. L.

Rotenberg

K. J.

Vanhalst

Harris

R. A.

Goossens

Bangee

Munn

(2013). Trajectories of loneliness during childhood and adolescence: Predictors and health outcomes. Journal of Adolescence, 36(6), 1283–1293. https://doi.org/10.1016/j.adolescence.2013.01.005

38.

Qualter

Vanhalst

Harris

Van Roekel

Lodder

Bangee

Maes

Verhagen

(2015). Loneliness across the life span. Perspectives on Psychological Science, 10(2), 250–264. https://doi.org/10.1177/1745691615568999

39.

Rokach

Sha’ked

(2013). Loneliness in schools: Teachers, students and their peers. Psychology and Education, 50(3-4), 89–97.

40.

Rönkä

A. R.

Sunnari

Rautio

Koiranen

Taanila

(2017). Associations between school liking, loneliness and social relations among adolescents: Northern Finland Birth Cohort 1986 study. International Journal of Adolescence and Youth, 22(1), 93–106. https://doi.org/10.1080/02673843.2015.1136659

41.

Schinka

K. C.

van Dulmen

M. H.

Bossarte

Swahn

(2012). Association between loneliness and suicidality during middle childhood and adolescence: Longitudinal effects and the role of demographic characteristics. Journal of Psychology, 146(1–2), 105–118. https://doi.org/10.1080/00223980.2011.584084

42.

Schinka

K. C.

van Dulmen

M. H.

Mata

A. D.

Bossarte

Swahn

(2013). Psychosocial predictors and outcomes of loneliness trajectories from childhood to early adolescence. Journal of Adolescence, 36(6), 1251–1260. https://doi.org/10.1016/j.adolescence.2013.08.002

43.

Shek

D. T.

Zhu

(2020). Promotion of thriving among Hong Kong Chinese adolescents: Evidence from eight-wave data. Research on Social Work Practice, 30(8), 870–883. https://doi.org/10.1177/1049731520947156

44.

Solmi

Veronese

Galvano

Favaro

Ostinelli

E. G.

Noventa

Favaretto

Tudor

Finessi

Shin

J. I.

Smith

Koyanagi

Cester

Bolzetta

Cotroneo

Maggi

Demurtas

De Leo

Trabucchi

(2020). Factors associated with loneliness: An umbrella review of observational studies. Journal of Affective Disorders, 271, 131–138. https://doi.org/10.1016/j.jad.2020.03.075

45.

Sorhagen

N. S.

Wurster

T. J.

(2017). Income within context: Relative income matters for adolescent social satisfaction and mental health. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 58(6), 736–743. https://doi.org/10.1111/jcpp.12695

46.

StataCorp . (2019). Stata statistical software: Release 16. StataCorp LLC.

47.

Steptoe

Shankar

Demakakos

Wardle

(2013). Social isolation, loneliness, and all-cause mortality in older men and women. Proceedings of the National Academy of Sciences of the United States of America, 110(15), 5797–5801. https://doi.org/10.1073/pnas.1219686110

48.

Suanet

van Tilburg

T. G.

(2019). Loneliness declines across birth cohorts: The impact of mastery and self-efficacy. Psychology and Aging, 34(8), 1134–1143. https://doi.org/10.1037/pag0000357

49.

van Staden

W. C.

Coetzee

(2010). Conceptual relations between loneliness and culture. Current Opinion in Psychiatry, 23(6), 524–529. https://doi.org/10.1097/YCO.0b013e32833f2ff9

50.

Xerxa

Rescorla

L. A.

Shanahan

Tiemeier

Copeland

W. E.

(2021). Childhood loneliness as a specific risk factor for adult psychiatric disorders. Psychological Medicine, 53(1), 227–235. https://doi.org/10.1017/S0033291721001422

51.

Yan

Feng

Schoppe-Sullivan

S. J.

(2018). Longitudinal associations between parent-child relationships in middle childhood and child-perceived loneliness. Journal of Family Psychology, 32(6), 841–847. https://doi.org/10.1037/fam0000446

52.

Yao

Pang

Xiao

Peng

(2022). Self-support and loneliness among Chinese primary school students: A moderated mediation model. Frontiers in Psychology, 12, 773421. https://doi.org/10.3389/fpsyg.2021.773421

53.

Ying

Yan

Shen

Jia

Lin

(2019). Economic pressure and loneliness in migrant children in China: The mediating roles of parent-child communication and parental warmth. Child Psychiatry and Human Development, 50(1), 142–149. https://doi.org/10.1007/s10578-018-0827-3

54.

Zhang

You

Fan

Gao

Cohen

Hsueh

Zhou

(2014). Friendship quality, social preference, proximity prestige, and self-perceived social competence: Interactive influences on children's loneliness. Journal of School Psychology, 52(5), 511–526. https://doi.org/10.1016/j.jsp.2014.06.001

55.

Zheng

Zhang

Lau

E. N. S.

Zhang

P. K. H.

Yang

Mak

E. C. W.

Wong

S. Y. S.

(2022). Primary school students’ online learning during Coronavirus Disease 2019: Factors associated with satisfaction, perceived effectiveness, and preference. Frontiers in Psychology, 13, 784826. https://doi.org/10.3389/fpsyg.2022.784826

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.66 MB