Abstract
Backgrounds
Relatively few studies have focused on the associations between family-related factors and adolescents’ mental health in China. This study aimed at investigating the associations between per capita living space, parental socioeconomic status, and depressive symptoms among adolescents.
Methods
We conducted a cross-sectional survey to investigate the mental health of adolescents in Guangzhou, China. Per capita living space was calculated by dividing the total housing size by the number of residents. Parental socioeconomic status included the education and employment status. Depressive symptoms were measured with the Patient Health Questionnaire–9. Mixed logistic regression analyses with school as a random intercept were used to explore the association between per capita living space, socioeconomic status, and depressive symptoms. Restrictive cubic spline (RCS) was used to examine whether the relationship between per capita living space and depression was linear.
Results
Compared with adolescents living in a house with per capita living space less than 10 m2, those living in space of 10–20 m2, 20–30 m2 had significantly lower odds of depressive symptoms, with no significant association observed for space over 30 m2. RCS analysis identified a non-symmetric U-shaped association. With the increase of per capita living space, the risk of depressive symptoms first decreased sharply, then rose and finally arrived at a plateau. Subgroup analyses stratified by sex and housing ownership showed similar nonlinear patterns that were observed in the full sample. Higher parental educational levels and maternal unemployment were associated with lower odds of depressive symptoms, whereas paternal unemployment was associated with higher depression risk only among girls.
Conclusions
Our study highlights the significant and complex influence of family-related social determinants including per capita living space and parental socioeconomic status on adolescents’ depressive symptoms. Extremely small living spaces are associated with depressive symptoms, moderate spaces are linked to less depression, while excessively large spaces offer no additional benefit. Prevention to promote adolescents’ mental health should prioritize adolescents with these socioeconomic disadvantages.
Clinical trial number
not applicable.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12889-025-23779-w.
Keywords: Housing size, Parental socioeconomic status, Depressive symptoms, Adolescents
Introduction
Mental health for children and adolescents is a public health priority both globally and in China. Depression is one of the leading causes of disease and disability in adolescents aged 15–19 years [1] and is associated with elevated risk of suicidal behavioral [2]. A recent meta-analysis reported a pooled depressive symptoms prevalence of 19.85% [3]. According to the National Mental Health Development Report in China (2019–2020), among 15,280 children from 4th to 12th grade, 17.2% exhibited mild depression and 7.4% exhibited severe depression [4]. Furthermore, the COVID-19 pandemic has further exacerbated emotional problems [5]. Considering the increasing disease burden and negative consequences of depressive symptoms, identifying factors associated with depressive symptoms among adolescents is crucial for developing effective prevention and intervention strategies.
Family environments are key social determinants that influence children and adolescents’ academic, behavior, and health outcomes [6–8]. Factors influencing a child’s well-being include their parents’ marital status, education, employment status, income, home environment, and etc [7]. Young people are more exposed to social determinants of mental health, especially in low-an-middle-income countries [9]. Previous research emphasized that the health effects of low socioeconomic status(SES) among children are more significant than in adults [8]. Growing inequalities in family SES in mainland China further generate inequalities in child mental health outcomes. However, relatively few studies have focused on family-related factors. Therefore, exploring the association of social determinants regarding house environment, parental SES and adolescents’ mental health is needed to deepen our understanding of the consequences of the economic development of China on children and adolescents’ mental health, which could be a long-term social issue.
Housing is a fundamental component of the living environment and a significant social determinant of both physical and mental health [10–13]. Several studies have established that poor housing conditions including poor affordability, inferior quality, and inadequate living space, are associated with poor mental health, including depression and anxiety [10, 11, 14]. For example, overcrowding increases psychological stress due to limited privacy and space, contributing to mental health challenges [14, 15].
Housing conditions, such as size, location, living condition, and price, reflect not only a family’s physical living environment but also their economic status, which both significantly influence mental health [16]. While most existing research has been conducted among adults [17], children and adolescents are more vulnerable to housing disadvantages. Early life exposures to housing disadvantage could have more negative influences, as social-emotional development among children and adolescents makes them more sensitive to stressors. Moreover, adolescents spend a substantial amount of time at home, which significantly increased during the quarantines in the COVID-19 pandemic, making housing an important context for mental health. Despite this, research on how housing overcrowding affects adolescents’ mental health remains limited, particularly in rapidly urbanizing contexts like China.
The most common measure of overcrowding, “persons per room” (PPR), ignores the role of housing size in reflecting crowding level and the relative family wealth [16, 18]. Per capita living space may provide a more precise estimate. Larger living spaces could not only provide a better supportive physical environment but also reflect better family economic status, promoting mental health. However, it is unclear whether this benefit consistently continues with living space increasing. Some studies suggested that the positive impact of a larger living space were limited [18]. For instance, research from Britain has shown that an increase in living space has only a weak positive linear effect on the life satisfaction and mental health of men, but moving to a larger home has no positive impact on subjective well-being [18]. This raises the question of whether the relationship of housing size and mental health is linear, or if there is an optimal range of space that maximizes mental health benefits among adolescents?
In addition to housing conditions, family factors such as parental socioeconomic status also influence the health and development of adolescents [8, 16, 19]. Families with a low SES are deprived in multiple ways and suffer from a higher number of stressors related to finances, social relations, employment situations and health complaints than those with a high SES [20]. Existing evidence on the association between parental SES and children’s health was mixed. Peverill et al., and Xiang et al. suggested that low parental educational level was independently associated with poor mental health [21, 22], whereas Zhang et al. found that mother’s education was associated with more depressive symptoms [23]. Evidence reported that only maternal unemployment was associated with a small increase in children’s hospitalization in Sweden [24], but in a US study, only paternal unemployment associated with anxiety and depression among children [25]. In China, Liu et al. suggested a negative association between child health and both parental recent job loss [26], while Pieters et al. found that paternal unemployment reduces children’s nutrient intake, whereas maternal unemployment has beneficial impacts [27]. Most existing studies merely included physical health and nutrition indicators, and relatively less research on mental health.
Therefore, the aim of this study is to examine the associations between housing size, parental socioeconomic status, and the mental health of adolescents, using a cross-sectional sample from junior and senior high schools in Guangzhou, China. We hypothesize that large housing size and higher parental education have a positive effect on mental health among adolescents, whereas parental unemployment would pose a negative impact on adolescents’ depressive symptoms.
Methods
Study design and participants
We conducted a cross-sectional survey to investigate the mental health of adolescents attending junior high schools and senior high schools in Guangzhou, China between December 2020 and September 2021. A multi-stage, stratified, and cluster random sampling method was used (see Supplementary Figure S1). The sample was stratified by administrative district and school. First, one central district was randomly selected from four central districts in Guangzhou, and one sub-district was randomly selected from seven sub-districts. Then, within each selected district, two junior high schools and two senior high schools were randomly selected. All students in these selected schools were invited to participate in our online survey. Eligible samples were (1) registered junior high school or senior high school students (2), willing to participate in the study. Students who were unable to complete the survey due to severe physical disorders, hearing or visual impairments were excluded. In addition, Grade 9 students from one junior high school and Grade 12 students from two senior high schools were excluded because these classes declined to participate due to ongoing exam preparation. A total of 9179 students were invited. Among them, 493 students (5.4%) declined to participate. There were no significant differences in age and gender between those who participated and those who did not. The research was approved by the institutional committee in Affiliated Brain Hospital, Guangzhou Medical University. Written informed consent was obtained from all participating students and their parents before the investigation.
Measurement and data collection
We used REDCap online questionnaire platform to collect data. Students who agree to participate in the study completed the self-administrated questionnaire at the computer center of their school, and one research assistant would be there in case that students have any question or need more information. Meanwhile, we also sent a questionnaire link with QR code to parents to collect information on parents and families. We deleted questionnaires with over 20% of missing data, as well as those finished in two minutes.
Depressive symptoms
Depressive symptoms measured with the Patient Health Questionnaire–9 (PHQ-9) was used as the indicator of mental health of adolescents [28, 29]. The PHQ-9 is a self-report instrument and screens for depressive symptoms. PHQ-9 consists of 9 items with a total summation score of 27. Higher score means more and severer symptoms of depression, with a score of 10 or above indicating depressive symptoms. Cronbach’s alpha among the participating sample is 0.884, indicating PHQ-9 has good reliability.
Per capita living space
Per capita living space was used as a measure of average housing areas. We collected data on the total housing area (m2), the number of residents living in this house, and the residential address. Per capita living space was calculated by dividing the total housing area by the number of residents. Extreme values were identified and verified using residential address information. Data with clearly erroneous information were recoded as missing data. All analyses were employed in adolescents with non-missing data of per capita living space.
For the regression model, per capita living space was categorized into four groups: 0–10 m2, 11–20 m2, 21–30 m2, and over 30 m2.
Parental socioeconomic status
Parental educational level was categorized in three different groups as primary and below, secondary, college and above; employment status of parents was divided into two groups: being employed and unemployed.
Covariates
We collected data on adolescents’ sociodemographic characteristics through a structured questionnaire. The covariates included age (in years), sex (male or female), grade (academic year level), school (school attended), residential student status (defined as residing in a student boarding house for three or more days per week, categorized as yes or no), and the only child status (whether the adolescent is an only child, categorized as yes or no). Family-related characteristics were also assessed, including family structure, classified into one of the following categories: nuclear family (adolescent lives with two parents), extended family (lives with parents and other relatives), single-parent family (lives with one parent), skip-generation family (lives with grandparents or great-grandparents without parents), blended family (lives with one biological parent and their partner and/or step-siblings), or other structure (any other arrangement not fitting the above categories). Additional variables included house ownership (whether the primary residence is owned or rented) and living location, classified as urban, suburban, or rural area based on the administrative designation of the adolescent’s residence. There were no missing data for any of the covariates.
Statistical methods
Participants’ characteristics were presented as the mean (standard deviation, SD) or median (interquartile range, IQR) for continuous variables and as the frequency (percentage) for categorical variables. Descriptive analyses were employed to describe and compare the sociodemographic characteristics of students, housing characteristics, depressive symptoms, parental educational levels and employment status by per capita living space category.
Univariable and multivariable mixed logistic regression analyses, with school as a random effect, were used to investigate the associations between per capita living space, parental SES (education and employment), and depressive symptoms. For the mixed logistic regression model, per capita living space was categorized into four groups: 0–10 m2, 11–20 m2, 21–30 m2, 31 m2 and above. Adolescents’ sociodemographic characteristics, family structure, house ownership, and the living location were adjusted. A restrictive cubic spline (RCS) model, using continuous per capita living space, was performed to examine whether the relationship between per capita living space and mental health is linear or not; the RCS model also included school as a random effect. All analyses were performed using STATA 17.0 (Stata Corporation, College Station, TX, USA) with a significance level of 0.05.
Results
Participants characteristics
A total of 8686 students completed the survey, 121 (1.4%) of them were excluded due to missing data on per capita living space. 8565 adolescents were included in the analysis. Table 1 presents the sociodemographic, housing characteristics, and depressive symptoms of participants according to per capita living space categories. The mean age of participating adolescents was 15.6 (SD = 1.6) years, and 50.4% were girls. The proportion of students living in households with per capita living space of 0–10 m2, 10–20 m2, 20–30, and 30 above m2 was 2.4%, 32.3%, 37.4%, and 17.9%, respectively. The median per capita living space was 25 m² (IQR: 20, 33 m²) (see Supplementary Figure S2). The number of household residents decreased with larger living space (mean 5.5 persons in ≤ 10 m² group vs. 3.5 persons in > 30 m² group). The median total housing size increased from 40 m² (IQR: 30–60 m²) in the 0–10 m² group to 130 m² (IQR: 100–180 m²) in the > 30 m² group. Home ownership also increased across per capita living space categories, from 53.0% in the 0–10 m² group to 87.5% in the > 30 m² group. The proportion of only children increased with larger living space, from 24.8% in the 0–10 m² group to 52.7% in the > 30 m² group. Regarding family structure, 4793 (56.0%) adolescents were from nuclear families, 2142 (25.0%) from extended families, 637 (7.4%) from single-parent families, 224 (2.6%) from skip-generation families, and 148 (1.7%) from blended families. The proportion of students from nuclear families increased with larger living space, while extended and skip-generation family structures were more common among those with smaller living spaces. The mean PHQ-9 score was 5.1 (4.9), with 14.9% of adolescents experiencing depressive symptoms in the last two weeks. The prevalence of depressive symptoms (PHQ-9 ≥ 10) was highest in the 0–10 m² group (20.3%), lowest in the 10–20 m² group (13.9%), and slightly higher in the > 30 m² group (16.5%).
Table 1.
Characteristics of adolescents by per capita living space category
| Characteristics | 0–10 m2 (N = 202) |
10–20 m2 (N = 2765) |
20–30 m2 (N = 3207) |
Over 30 m2 (N = 2391) |
Total (N = 8565) |
|---|---|---|---|---|---|
| Age, mean (SD) | 15.1 (1.8) | 15.4 (1.7) | 15.7 (1.6) | 15.8 (1.6) | 15.6 (1.6) |
| Gender | |||||
| Boys | 116 (57.4) | 1252 (45.3) | 1633 (50.9) | 1244 (52.0) | 4245 (49.6) |
| Girls | 86 (42.6) | 1513 (54.7) | 1574 (49.1) | 1147 (48.0) | 4320 (50.4) |
| Grade | |||||
| Grade 7 | 46 (22.8) | 473 (17.1) | 412 (12.8) | 290 (12.1) | 1221 (14.3) |
| Grade 8 | 40 (19.8) | 410 (14.8) | 357 (11.1) | 240 (10.0) | 1047 (12.2) |
| Grade 9 | 15 (7.4) | 270 (9.8) | 294 (9.2) | 193 (8.1) | 772 (9.0) |
| Grade 10 | 41 (20.3) | 647 (23.4) | 901 (28.1) | 745 (31.2) | 2334 (27.3) |
| Grade 11 | 36 (17.8) | 690 (25.0) | 892 (27.8) | 671 (28.1) | 2289 (26.7) |
| Grade 12 | 24 (11.9) | 275 (9.9) | 351 (10.9) | 252 (10.5) | 902 (10.5) |
| Boarding status | |||||
| Yes | 107 (53.0) | 1336 (48.3) | 1544 (48.1) | 1288 (53.9) | 4275 (49.9) |
| No | 95 (47.0) | 1429 (51.7) | 1663 (51.9) | 1103 (46.1) | 4290 (50.1) |
| The only one child | |||||
| Yes | 50 (24.8) | 578 (20.9) | 1298 (40.5) | 1260 (52.7) | 3186 (37.2) |
| No | 152 (75.2) | 2187 (79.1) | 1909 (59.5) | 1131 (47.3) | 5379 (62.8) |
| Family structure | |||||
| Nuclear family | 89 (44.1) | 1405 (50.8) | 1945 (60.6) | 1354 (56.6) | 4793 (56.0) |
| Others | 31 (15.3) | 261 (9.4) | 198 (6.2) | 131 (5.5) | 621 (7.3) |
| Extended family | 62 (30.7) | 885 (32.0) | 721 (22.5) | 474 (19.8) | 2142 (25.0) |
| Single-parent family | 12 (5.9) | 102 (3.7) | 195 (6.1) | 328 (13.7) | 637 (7.4) |
| Skip-generation family | 6 (3.0) | 66 (2.4) | 91 (2.8) | 61 (2.6) | 224 (2.6) |
| Blended family | 2 (1.0) | 46 (1.7) | 57 (1.8) | 43 (1.8) | 148 (1.7) |
| Education level-father | |||||
| Primary or below | 19 (9.4) | 186 (6.7) | 148 (4.6) | 120 (5.0) | 473 (5.5) |
| Junior high | 77 (38.1) | 815 (29.5) | 737 (23.0) | 477 (19.9) | 2106 (24.6) |
| Senior high | 78 (38.6) | 896 (32.4) | 954 (29.7) | 630 (26.3) | 2558 (29.9) |
| Junior college | 15 (7.4) | 382 (13.8) | 508 (15.8) | 326 (13.6) | 1231 (14.4) |
| College and above | 13 (6.4) | 486 (17.6) | 860 (26.8) | 838 (35.0) | 2197 (25.7) |
| Education level-mother | |||||
| Primary or below | 34 (16.8) | 346 (12.5) | 257 (8.0) | 151 (6.3) | 788 (9.2) |
| Junior high | 75 (37.1) | 895 (32.4) | 843 (26.3) | 557 (23.3) | 2370 (27.7) |
| Senior high | 54 (26.7) | 722 (26.1) | 863 (26.9) | 589 (24.6) | 2228 (26.0) |
| Junior college | 26 (12.9) | 390 (14.1) | 499 (15.6) | 384 (16.1) | 1299 (15.2) |
| College and above | 13 (6.4) | 412 (14.9) | 745 (23.2) | 710 (29.7) | 1880 (21.9) |
| Paternal employment | |||||
| Employed | 177 (87.6) | 2580 (93.3) | 2996 (93.4) | 2214 (92.6) | 7967 (93.0) |
| Unemployed | 25 (12.4) | 185 (6.7) | 211 (6.6) | 177 (7.4) | 598 (7.0) |
| Maternal employment | |||||
| Employed | 167 (82.7) | 2231 (80.7) | 2516 (78.5) | 1873 (78.3) | 6787 (79.2) |
| Unemployed | 35 (17.3) | 534 (19.3) | 691 (21.5) | 518 (21.7) | 1778 (20.8) |
| House ownership | |||||
| Owned | 107 (53.0) | 1898 (68.6) | 2548 (79.5) | 2091 (87.5) | 6644 (77.6) |
| Rent | 95 (47.0) | 867 (31.4) | 659 (20.5) | 300 (12.5) | 1921 (22.4) |
| Location | |||||
| City | 148 (73.3) | 2326 (84.1) | 2683 (83.7) | 1930 (80.7) | 7087 (82.7) |
| Suburban | 39 (19.3) | 334 (12.1) | 420 (13.1) | 348 (14.6) | 1141 (13.3) |
| Rural area | 15 (7.4) | 105 (3.8) | 104 (3.2) | 113 (4.7) | 337 (3.9) |
|
Total housing size, median (IQR) |
40.0 (30.0, 60.0) | 80.0 (70.0, 100.0) | 100.0 (80.0, 110.0) | 130.0 (100.0, 180.0) | 100.0 (75.0, 120.0) |
| Number of residents, mean (SD) | 5.5 (2.5) | 4.8 (1.3) | 3.9 (1.0) | 3.5 (1.1) | 4.1 (1.3) |
| Score of PHQ-9 | 5.7 (5.7) | 5.0 (4.8) | 5.1 (4.9) | 5.3 (5.0) | 5.1 (4.9) |
| Depressive symptom | 41 (20.3) | 384 (13.9) | 459 (14.3) | 394 (16.5) | 1278 (14.9) |
The associations between per capita living area, parental SES, and depressive symptoms
We used univariable and multivariable mixed logistic regression to examine the associations between per capita living area, parental SES, and depressive symptoms. Both unadjusted and adjusted ORs are presented in Table 2. Adjustment for potential confounders did not substantially change the direction of associations, though the magnitude of some associations was slightly attenuated. After adjusting for covariates, adolescents living in houses with 10–20 m2 of per capita living space (OR [95% CI]: 0.57 [0.40, 0.83]) or 20–30 m² (OR [95% CI]: 0.66 [0.46, 0.96]) had significantly lower odds of depressive symptoms compared to those living in houses with a per capita living space less than 10 m2. No significant association was observed for those living in houses with a per capita living space larger than 30 m2 (OR [95%CI]: 0.80 [0.55, 1.16]). No significance between the “over 30m2” and “0-10m2” categories did not indicate equal risks, but suggested a potential non-linear relationship.
Table 2.
Associations between per capita living space, parental SES and depressive symptoms
| Depressive symptoms | Unadjusted OR | 95% CI | P | Adjusted OR a | 95% CI | P |
|---|---|---|---|---|---|---|
| Per Capita Living Space (Ref = 0–10 m2) | ||||||
| 10–20 m2 | 0.62 | 0.43, 0.90 | 0.010 | 0.57 | 0.40, 0.83 | 0.003 |
| 20–30 m2 | 0.66 | 0.46, 0.94 | 0.021 | 0.66 | 0.46, 0.96 | 0.028 |
| Over 30 m2 | 0.79 | 0.55, 1.13 | 0.192 | 0.80 | 0.55, 1.16 | 0.243 |
| Education level-father (Ref = primary or below) | ||||||
| Secondary | 0.70 | 0.55, 0.89 | 0.004 | 0.75 | 0.58, 0.97 | 0.025 |
| College and above | 0.72 | 0.56, 0.93 | 0.012 | 0.78 | 0.59, 1.04 | 0.086 |
| Education level-mother (Ref = primary or below) | ||||||
| Secondary | 0.73 | 0.60, 0.90 | 0.002 | 0.73 | 0.59, 0.90 | 0.003 |
| College and above | 0.77 | 0.62, 0.96 | 0.021 | 0.75 | 0.58, 0.96 | 0.023 |
| Paternal employment (Ref = employed) | ||||||
| Unemployed | 1.27 | 1.02, 1.58 | 0.032 | 1.27 | 1.01, 1.60 | 0.038 |
| Maternal employment (Ref = employed) | ||||||
| Unemployed | 0.89 | 0.76, 1.03 | 0.125 | 0.84 | 0.72, 0.99 | 0.032 |
aAdjusted for sex, grade, being an only child, boarding status, family structure, house ownership, and living location, with school included as a random effect
To examine whether the associations between per capita living space and depressive symptoms were consistent across different sex and house ownership statuses, we separately add a per capita living space × sex interaction term, and a per capita living space × house ownership interaction term into the mixed model. Results of subgroup analysis are showed in Table 3. No significant interactions (all P for interaction > 0.05) were found between per capita living space and sex, nor between per capita living space and house ownership.
Table 3.
Associations between per capita living space and depressive symptoms stratified by sex and housing property, and between paternal employment and depressive symptoms by sex
| Depressive symptoms | Depressive cases/total | Adjusted OR | 95% CI | P |
|---|---|---|---|---|
| Per capita living space | ||||
| Sex | ||||
| Male | ||||
| 0–10 m2 | 21/116 | 1.00 | Ref | |
| 10–20 m2 | 127/1252 | 0.52 | 0.31, 0.88 | 0.014 |
| 20–30 m2 | 187/1633 | 0.63 | 0.38, 1.06 | 0.081 |
| Over 30 m2 | 153/1244 | 0.70 | 0.41, 1.18 | 0.178 |
| Female | ||||
| 0–10 m2 | 20/86 | 1.00 | Ref | |
| 10–20 m2 | 257/1513 | 0.64 | 0.38, 1.09 | 0.099 |
| 20–30 m2 | 272/1574 | 0.71 | 0.42, 1.21 | 0.206 |
| Over 30 m2 | 241/1147 | 0.93 | 0.54, 1.61 | 0.805 |
| House ownership | ||||
| Owned house | ||||
| 0–10 m2 | 19/107 | 1.00 | Ref | |
| 10–20 m2 | 272/1898 | 0.69 | 0.41, 1.17 | 0.165 |
| 20–30 m2 | 365/2548 | 0.78 | 0.46, 1.31 | 0.345 |
| Over 30 m2 | 338/2091 | 0.95 | 0.56, 1.61 | 0.844 |
| Rented house | ||||
| 0–10 m2 | 22/95 | 1.00 | Ref | |
| 10–20 m2 | 112/867 | 0.43 | 0.25, 0.73 | 0.002 |
| 20–30 m2 | 94/659 | 0.52 | 0.30, 0.91 | 0.020 |
| Over 30 m2 | 56/300 | 0.65 | 0.36, 1.17 | 0.152 |
| Paternal employment | ||||
| Male | ||||
| Father employed | 456/3944 | 1.00 | Ref | |
| Father unemployed | 32/301 | 0.83 | 0.56, 1.23 | 0.356 |
| Female | ||||
| Father employed | 714/4023 | 1.00 | Ref | |
| Father unemployed | 76/297 | 1.64 | 1.23, 2.18 | 0.001 |
Given the multivariable logistic regression models suggested a non-linear association between per capita living space and depressive symptoms, a restrictive cubic spline analysis (RCS) was conducted to visualize the relationship. We used the multivariable mixed logistic model for the RCS analysis. The RCS analysis identified a significant non-symmetric U-shaped relationship (P for linear test < 0.01) between per capita living space and depressive symptoms (Fig. 1). With the increase of per capita living space, the odds of depressive symptoms sharply declined until the per capita living space reaching 20 m2; beyond this point, the risk of depressive symptoms began to rise and arrived at a plateau with a gradual upward trend after 30 m2. Subgroup analyses stratified by sex and housing ownership showed similar nonlinear patterns that observed in the full sample (Fig. 1). In a sensitivity analysis excluding the bottom 1% of values within the 0–10 m² group and the top 1% of values within the > 30 m² group, the overall U-shaped pattern of the RCS curve remained consistent, supporting the robustness of the observed non-linear association (see Supplementary Figure S3).
Fig. 1.
Restricted cubic spline model of the odds ratios of depression symptoms with the per capita living space. Adjusted for sex, grade, being an only child, boarding status, family structure, house ownership, and living location, with school included as a random effect. a represents the whole research sample. The dark grey line represents the ORs for depressive symptoms and the light grey areas represent the 95% confidence intervals. b stratified by sex. c stratified by house ownership
Regarding parental education, higher parental educational levels were associated with lower odds of depressive symptoms after adjusted for covariates. In terms of parental employment, paternal unemployment (OR [95%CI]: 1.27 [1.01,1.60]) was positively associated with depressive symptoms. Maternal unemployment was not significantly associated with depressive symptoms in the univariable analysis but became significantly associated with lower odds of depressive symptoms (adjusted OR = 0.84, 95% CI: 0.72,0.99) after adjustment. To examine whether the associations between parental SES and depressive symptoms differed by sex, we tested interactions between each parental SES indicator and sex. Results showed a significant paternal employment × sex interaction (P for interaction = 0.011). Figure 2 confirmed this interaction. The prevalence of depressive symptoms was higher among females (17.8%) compared to males (11.6%) when fathers were employed; among those whose fathers were unemployed, the prevalence significantly increases for females to 25.6%, while it remains relatively stable for males. In addition, the predicted probability of depressive symptoms increased significantly with paternal unemployment among females, while it remains low and relatively unchanged for males. Paternal unemployment was negatively associated with depressive symptoms among girls, while no significant association was observed among boys.
Fig. 2.
Associations between paternal unemployment and depressive symptoms stratified by sex. a Prevalence of depressive symptoms among adolescents by paternal employment status, stratified by sex. b predicted probability of depressive symptoms associated with paternal unemployment, stratified by sex
Discussions
We found 14.9% of adolescents were experiencing depressive symptoms. Results of the study identified a non-symmetric U-shaped relationship between per capita living areas and depressive symptoms: increasing living space from very low levels to moderate levels is associated with lower odds of depressive symptoms, but further increases do not appear to confer additional mental health benefits. In addition, we found that maternal unemployment was associated with less depressive symptoms, while paternal unemployment was associated with higher risk of depressive symptoms among female adolescents. These findings expand existing knowledge on how social determinants influence adolescent mental health.
This is a novel finding that has identified a non-symmetric U-shaped relationship between per capita living areas and depressive symptoms, challenging the assumption that larger living areas was associated with fewer depressive symptoms. Compared to adolescents living in extremely crowded conditions (per capita living space less than 10 m²), increasing per capita living space to 10–20 m² and 20–30 m² was associated with lower odds of depressive symptoms, indicating that alleviating extreme crowding could improve mental health for adolescents. However, this improvement does not persist as per capita living space continue to increase, indicating that the mental health benefits plateau after a certain threshold. Results of restricted cubic spline analysis confirmed this non-symmetric U-shaped relationship, showing that the risk of depressive symptoms sharply declined until the per capita living areas approximately 20 m², beyond which the risk began to rise and plateaus after 30 m². These findings were consistent with previous research demonstrating negative association between household overcrowding and mental health [12, 14]. Overcrowding usually coexists with other housing and economic disadvantages [18], increase stress and pressure [12] and inviolate personal spaces [15]. Interventions should prioritize improving living conditions for adolescents living in overcrowded environments.
The non-symmetric U-shaped relationship finding also revealed that simply living a larger house does not necessarily equate to better mental health for adolescents. While a larger per capita living space often associated with a better physical living environment and relatively higher family socioeconomic status, larger objective living space alone might be insufficient to promote better mental health among adolescents. One possible explanation is that a too large house might reduce the opportunities for parent-child communication, potentially weakening family bonds. Additionally, although housing size may reflect a family’s relative economic status, it provides very limited insight. It is also important to note that unmeasured confounders, such as parental mental health, family stress, and household dynamics may influence both housing conditions and adolescent mental health. Furthermore, factors beyond physical living—such as housing quality, neighborhood safety, access to amenities, and perceived privacy—likely play significant roles in shaping mental health outcomes [30, 31]. Given the observational nature of this study, causal relationships cannot be established. Future research should explore those additional psychosocial and housing-related factors to better understand the comprehensive influence of housing on adolescents’ mental health.
Consistent with previous studies [8, 19, 32–34], we found that higher parental educational levels were associated with lower odds of depressive symptoms, emphasizing the protective role of parental education in adolescent mental well-being. For instance, the BELLA study found that children of mothers with low education had significantly more mental health problems during childhood and adolescence than children of mothers with higher educational levels [35]. Higher parental education is often associated with better socioeconomic status, greater access to resources for children’s mental health, and the ability to create a stable and supportive environment for the psychological well-being of children. In addition, well-educated parents are more likely to provide effective coping strategies to deal with stressful life situations, and mitigate the impact of stressful life situations on their children’s mental health [20].
One interesting phenomenon we found was that paternal unemployment was associated with adolescents’ depressive symptoms, while maternal unemployment buffered adolescents’ depressive symptoms. This divergence reflects the traditional gender roles [27], following the East Asian cultural norm of ‘husband handling external affairs and wife managing domestic responsibilities’. As a primary breadwinner, father’s unemployment could lead to financial stress and perceived masculine compromise, potentially posing a negative influence on parent-child relations and adolescents’ mental well-being [36, 37]. In contrast, maternal unemployment might allow mothers to spend more time caring for their children, fostering a more supportive home environment. Furthermore, we found that the association between paternal unemployment and depressive symptoms was observed only among girls, not boys. This could be explained by differences in gender socialization. Girls might be more sensitive to family and emotional stressors and are more likely to disclose and internalize stresses and negative emotions [38, 39].
Our study is the first to explore the associations between per capita living space, parental socioeconomic status, and depressive symptoms among Chinese adolescents. However, several limitations should be noted when interpreting the results. Firstly, our study is limited by its cross-sectional design, which prevent us from determine the temporal sequence between exposure (e.g., per capita living space, parental SES) and depressive symptoms. Thus, causal inferences cannot be made, and longitudinal research is needed to test our results. However, per capita living space and parental SES are relatively stable factors that are unlikely to change frequently over time, which might reduce the potential bias. Secondly, the data of per capita living space were self-reported and could be susceptible to reporting or recall bias. We took several steps to address this limitation by rechecking the data with detailed addresses and removing extreme outliers. Future studies could benefit from using objective measurements, such as those obtained from property records or spatial mapping. Thirdly, our study did not collect data on household income or other housing-related factors, such as housing quality, neighborhood environment, or perceived privacy in the house, which might be potential mediators or confounders in the relationship between housing size and depressive symptoms. Although we adjusted for a range of covariates, the possibility of residual confounding cannot be entirely ruled out, and future research should include more housing-related information. Finally, since the study included students from only eight schools in two districts of Guangzhou, the full range of economic variability across all districts in the city may not be fully represented. In addition, most of the data were collected during the COVID-19 pandemic, which may limit the generalizability of our findings. Future studies should aim to include a broader and more diverse sample to enhance the representativeness and external validity of the results.
Conclusion
Our study highlights the significant and complex influence of family-related social determinants — including per capita living space and parental socioeconomic status— on adolescents’ depressive symptoms. Among these factors, our findings reveal a non-symmetric U-shaped relationship between per capita living space and depressive symptoms, where very small living spaces are linked to poorer mental health, moderate spaces are linked to lower odds of depressive symptoms, and larger spaces offer no additional benefit. These findings underscore the necessity of incorporating these social determinants into research, prevention, and intervention targeting adolescent’s mental health. Addressing socioeconomic health disparities in mental health is not only the responsibility of individuals or families, but of society as a whole. Programs to promote adolescents’ mental health should adopt a multi-level approach and prioritize adolescents experiencing socioeconomic disadvantages.
Supplementary Information
Acknowledgements
We thank all adolescents and their parents who participated in this study. We acknowledge all research personnel in the Affiliated Brain Hospital of Guangzhou Medical University.
Authors’ contributions
Conception and design of the study: Dr. Wang and Prof. Zhou. Data curation: Wang. Formal analysis, Methodology, Software, Visualization: Wang. Writing- original draft: Wang. Writing-review & editing: Wang, Sun, and Zhou. Funding acquisition: Zhou. Resource: Sun and Zhou.
Funding
This work was supported by the Key Project of Guangzhou Municipal Health Commission [grant numbers: 20191A031003]; Guangzhou Research-oriented Hospital. Science and Technology Program of Guangzhou (grant numbers: 2024A04J10000, 2024A03J0223). Guangzhou Municipal Key Discipline in Medicine (2025-2027).
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. The research was approved by the institutional committee in Affiliated Brain Hospital, Guangzhou Medical University. Written informed consent was obtained from all participating students and their parents before the investigation.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Gore FM, Bloem PJ, Patton GC, Ferguson J, Joseph V, Coffey C, et al. Global burden of disease in young people aged 10–24 years: a systematic analysis. Lancet (London England). 2011;377(9783):2093–102. 10.1016/s0140-6736(11)60512-6. [DOI] [PubMed] [Google Scholar]
- 2.Pozuelo JR, Desborough L, Stein A, Cipriani A. Systematic review and Meta-analysis: depressive symptoms and risky behaviors among adolescents in Low- and Middle-Income countries. J Am Acad Child Adolesc Psychiatry. 2022;61(2):255–76. 10.1016/j.jaac.2021.05.005. [DOI] [PubMed] [Google Scholar]
- 3.Rao WW, Xu DD, Cao XL, Wen SY, Che WI, Ng CH, et al. Prevalence of depressive symptoms in children and adolescents in china: A meta-analysis of observational studies. Psychiatry Res. 2019;272:790–6. 10.1016/j.psychres.2018.12.133. [DOI] [PubMed] [Google Scholar]
- 4.Fu X, Zhang K, Chen X, Chen Z. Report on National mental health development in China (2019–2020). Bluebook on mental health series. Beijing: Social Sciences Academic; 2021. [Google Scholar]
- 5.Qi H, Liu R, Feng Y, Luo J, Lei SM, Cheung T, et al. Prevalence of depression and its associated factors among adolescents in China during the early stage of the COVID-19 outbreak. Peerj. 2021;9. 10.7717/peerj.12223. [DOI] [PMC free article] [PubMed]
- 6.Tamura K, Morrison J, Pikhart H. Children’s behavioural problems and its associations with socioeconomic position and early parenting environment: findings from the UK millennium cohort study. Epidemiol Psychiatric Sci. 2020;29:e155. 10.1017/s2045796020000700. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bush NR, Wakschlag LS, LeWinn KZ, Hertz-Picciotto I, Nozadi SS, Pieper S, et al. Family environment, neurodevelopmental risk, and the environmental influences on child health outcomes (ECHO) initiative: looking back and moving forward. Front Psychiatry. 2020;11. 10.3389/fpsyt.2020.00547. [DOI] [PMC free article] [PubMed]
- 8.Reiss F. Socioeconomic inequalities and mental health problems in children and adolescents: A systematic review. Soc Sci Med. 2013;90:24–31. 10.1016/j.socscimed.2013.04.026. [DOI] [PubMed] [Google Scholar]
- 9.Rose-Clarke K, Bitta M, Evans-Lacko S, Jokinen T, Jordans M, Nyongesa MK, et al. Centring youth mental health discourse on low-income and middle-income countries. Lancet Psychiatry. 2024;11(9):671–2. 10.1016/S2215-0366(24)00211-6. [DOI] [PubMed] [Google Scholar]
- 10.Ruiz-Tagle J, Urria I. Household overcrowding trajectories and mental well-being. Soc Sci Med. 2022;296. 10.1016/j.socscimed.2021.114051. [DOI] [PubMed]
- 11.Liu PC, Zhou LF, Chen SJ, Wang XJ. Association between household overcrowding and depressive mood among Chinese residents. J Affect Disord. 2021;290:74–80. 10.1016/j.jad.2021.04.066. [DOI] [PubMed] [Google Scholar]
- 12.Marsh R, Salika T, Crozier S, Robinson S, Cooper C, Godfrey K, et al. The association between crowding within households and behavioural problems in children: longitudinal data from the Southampton women’s survey. Paediatr Perinat Epidemiol. 2019;33(3):195–203. 10.1111/ppe.12550. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hansen CB, Larsen CVL, Bjerregaard P, Riva M. The effect of household crowding and composition on health in an Inuit cohort in Greenland. Scand J Public Health. 2020;49(8):921–30. 10.1177/1403494820929496. [DOI] [PubMed] [Google Scholar]
- 14.Singh A, Daniel L, Baker E, Bentley R. Housing disadvantage and poor mental health: A systematic review. Am J Prev Med. 2019;57(2):262–72. 10.1016/j.amepre.2019.03.018. [DOI] [PubMed] [Google Scholar]
- 15.Weber C, Gatersleben B, Jagannath S, Fuchslin B, Delabrida ZNC. Crowding and aggression during the COVID-19 lockdown in the united kingdom: the relationship between residential density, subjective crowding, privacy, and aggression. J Environ Psychol. 2024;96. 10.1016/j.jenvp.2024.102335.
- 16.Moulton V, Goodman A, Nasim B, Ploubidis GB, Gambaro L. Parental wealth and children’s cognitive ability, mental, and physical health: evidence from the UK millennium cohort study. Child Dev. 2021;92(1):115–23. 10.1111/cdev.13413. [DOI] [PubMed] [Google Scholar]
- 17.Rautio N, Filatova S, Lehtiniemi H, Miettunen J. Living environment and its relationship to depressive mood: A systematic review. Int J Soc Psychiatry. 2018;64(1):92–103. 10.1177/0020764017744582. [DOI] [PubMed] [Google Scholar]
- 18.Foye C. The relationship between size of living space and subjective Well-Being. J Happiness Stud. 2017;18(2):427–61. 10.1007/s10902-016-9732-2. [Google Scholar]
- 19.Sun XY, Yuan YQ. Associations between parental socioeconomic status and mental health in Chinese children: the mediating roles of parenting practices. Child Adolesc Mental Health. 2024;29(3):292–8. 10.1111/camh.12715. [DOI] [PubMed] [Google Scholar]
- 20.Reiss F, Meyrose AK, Otto C, Lampert T, Klasen F, Ravens-Sieberer U. Socioeconomic status, stressful life situations and mental health problems in children and adolescents: results of the German BELLA cohort-study. PLoS ONE. 2019;14(3). 10.1371/journal.pone.0213700. [DOI] [PMC free article] [PubMed]
- 21.Xiang YH, Cao R, Li XJ. Parental education level and adolescent depression: A multi-country meta-analysis. J Affect Disord. 2024;347:645–55. 10.1016/j.jad.2023.11.081. [DOI] [PubMed] [Google Scholar]
- 22.Peverill M, Dirks MA, Narvaja T, Herts KL, Comer JS, McLaughlin KA. Socioeconomic status and child psychopathology in the united states: A meta-analysis of population-based studies. Clin Psychol Rev. 2021;83. 10.1016/j.cpr.2020.101933. [DOI] [PMC free article] [PubMed]
- 23.Zhang P, Huang N, Yang F, Yan W, Zhang B, Liu X, et al. Determinants of depressive symptoms at individual, school and Province levels: a National survey of 398,520 Chinese children and adolescents. Public Health. 2024;229:33–41. 10.1016/j.puhe.2024.01.015. [DOI] [PubMed] [Google Scholar]
- 24.Mörk E, Sjögren A, Svaleryd H. Parental unemployment and child health. Cesifo Economic Stud. 2014;60(2):366–401. 10.1093/cesifo/ifu016. [Google Scholar]
- 25.Schaller J, Zerpa M. Short-Run effects of parental job loss on child health. Am J Health Econ. 2019;5(1):8–41. 10.1162/ajhe_a_00106. [Google Scholar]
- 26.Liu H, Zhao Z. Parental job loss and children’s health: ten years after the massive layoff of the soes’ workers in China. China Econ Rev. 2014;31:303–19. 10.1016/j.chieco.2014.10.007. [Google Scholar]
- 27.Pieters J, Rawlings S. Parental unemployment and child health in China. Rev Econ Househ. 2020;18(1):207–37. 10.1007/s11150-019-09457-y. [Google Scholar]
- 28.Wang W, Bian Q, Zhao Y, Li X, Wang W, Du J, et al. Reliability and validity of the Chinese version of the patient health questionnaire (PHQ-9) in the general population. Gen Hosp Psychiatry. 2014;36(5):539–44. 10.1016/j.genhosppsych.2014.05.021. [DOI] [PubMed] [Google Scholar]
- 29.Zhang YL, Liang W, Chen ZM, Zhang HM, Zhang JH, Weng XQ, et al. Validity and reliability of patient health Questionnaire-9 and patient health Questionnaire-2 to screen for depression among college students in China. Asia Pac Psychiatry. 2013;5(4):268–75. 10.1111/appy.12103. [DOI] [PubMed] [Google Scholar]
- 30.Xie SH. Quality matters: housing and the mental health of rural migrants in urban China. Hous Stud. 2019;34(9):1422–44. 10.1080/02673037.2019.1577956. [Google Scholar]
- 31.Rollings KA, Wells NM, Evans GW, Bednarz A, Yang YZ. Housing and neighborhood physical quality: children’s mental health and motivation. J Environ Psychol. 2017;50:17–23. 10.1016/j.jenvp.2017.01.004. [Google Scholar]
- 32.Yang DL, Hu SX, Li MN. The Influence of Family Socioeconomic Status on Adolescents’ Mental Health in China. Int J Environ Res Public Health. 2022;19(13). 10.3390/ijerph19137824. [DOI] [PMC free article] [PubMed]
- 33.Weinberg D, Stevens G, Duinhof EL, Finkenauer C. Adolescent socioeconomic status and mental health inequalities in the netherlands, 2001–2017. Int J Environ Res Public Health. 2019;16(19). 10.3390/ijerph16193605. [DOI] [PMC free article] [PubMed]
- 34.Ochi M, Fujiwara T, Mizuki R, Kawakami N, World Mental Hlth Japan Survey G. Association of socioeconomic status in childhood with major depression and generalized anxiety disorder: results from the world mental health Japan survey 2002–2006. BMC Public Health. 2014;14. 10.1186/1471-2458-14-359. [DOI] [PMC free article] [PubMed]
- 35.Meyrose AK, Klasen F, Otto C, Gniewosz G, Lampert T, Ravens-Sieberer U. Benefits of maternal education for mental health trajectories across childhood and adolescence. Soc Sci Med. 2018;202:170–8. 10.1016/j.socscimed.2018.02.026. [DOI] [PubMed] [Google Scholar]
- 36.Ortiz LPA, Farrell MP. FATHERS UNEMPLOYMENT AND ADOLESCENTS SELF-CONCEPT. Adolescence. 1993;28(112):937–49. [PubMed] [Google Scholar]
- 37.Lui L, Kee-lee C. Stay-at-home fathers in Hong kong: trends and characteristics. Asian Popul Stud. 2019;15(3):282–301. 10.1080/17441730.2019.1660095. [Google Scholar]
- 38.Graber JA, Sontag LM. Internalizing problems during adolescence. In: Lerner RM, Steinberg L, editors. Handbook of adolescent psychology. 3rd ed. Wiley; 2009. pp. 642–82. 10.1002/9780470479193.adlpsy001020.
- 39.Garside RB, Klimes-Dougan B. Socialization of discrete negative emotions: genderdifferences and links with psychological distress. Sex Roles. 2002;47:115–28. 10.1023/a:1021090904785. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.