Abstract
Background
Child and adolescent overweight and obesity are rising in China, yet how parental words, namely verbal encouragement, and actions, namely parents’ own physical activity, relate to weight status remains unclear. This study estimated the associations of four observed parental role-modelling patterns, defined by parental verbal support and guideline-level parental physical activity, with overweight and obesity in children and adolescents.
Methods
Cross-sectional data were drawn from the Shanghai site of the Physical Activity and Fitness in China, The Youth Study from 2016 to 2020. The study included children and adolescents in Grades 4 to 12, each linked to one responding parent. Parental verbal support and parental physical activity were assessed by questionnaire. Two binary indicators, verbal support and guideline-level parental physical activity, were cross-classified to define four role-modelling patterns: no verbal support and no guideline-level parental physical activity (NS), verbal support only (VO), guideline-level parental physical activity only (AO), and both verbal support and guideline-level parental physical activity (BS). Child height and weight were measured objectively, BMI was calculated, and overweight or obesity was defined using WHO age- and sex-specific references. We used cluster-adjusted logistic regression to estimate adjusted odds ratios (aORs) for overweight or obesity, with adjustment for sociodemographic covariates. Subgroup-specific estimates were examined descriptively, and formal interaction tests were conducted in the full analytic sample.
Results
Data from 18 479 parent–child dyads (aged 9–18 years) were analysed; VO was most common (42.1%), followed by BS (26.8%), NS (22.4%), and AO (8.7%). Overall, 23.2% had overweight or obesity. Compared with NS, AO was associated with lower odds of overweight or obesity (aOR 0.82, 95% CI 0.69–0.96). BS was associated with even lower odds (aOR 0.57, 95% CI 0.50–0.65), suggesting a greater magnitude of association, whereas VO showed no evidence of association (aOR 0.99, 95% CI 0.98–1.01). BS was associated with lower odds across all examined subgroups, whereas the association for AO was more evident in Grades 7–9, non-only children, and higher-income households.
Conclusion
Guideline-level parental physical activity was more consistently associated with healthier weight status among children and adolescents than verbal support alone. The combination of verbal support and guideline-level parental physical activity was associated with the lowest odds of overweight or obesity. Family-based obesity prevention should support parents in developing active routines and aligning verbal encouragement with their own physical activity.
Graphical Abstract
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12889-026-28099-1.
Keywords: Child, Adolescent, Overweight, Obesity; Parental support, Role modeling
Introduction
The global prevalence of overweight and obesity among children and adolescents has reached epidemic proportions, representing a critical public health challenge. Global obesity rates among adolescents have quadrupled since 1990, as reported in The Lancet [1]. This trend is particularly acute in China, where recent analysis of three decades of national survey data (1985–2019) revealed a population-wide shift toward higher and more dispersed BMI, signifying a substantial increase in overweight and obesity of youth [2]. Childhood overweight and obesity carry substantial immediate and long-term health sequelae, including cardiometabolic risks and adverse psychosocial outcomes [3]. This condition is proximally driven by behavioral factors, with insufficient physical activity (PA) recognized as a primary contributor to energy imbalance. The persistence of this epidemic, despite various interventions, underscores the necessity of understanding its determinants within foundational environments, particularly the family [4].
Within the family ecosystem, parents function as the primary socialization agents. Their influence, often conceptualized as parental role modeling, is a critical determinant of children's health. This modeling is not monolithic; it operates through distinct, though potentially related, mechanisms. These include explicit verbal encouragement (i.e., 'words'), such as providing praise or reminders to be active, and behavioral role modeling (i.e., 'actions'), whereby parents demonstrate an active lifestyle themselves [5]. While many studies link this modeling to children's PA, parental influence also has a recognized direct association with weight outcomes [6]. Previous evidence links parental behaviours, including physical activity-related practices and parental weight status, to child body mass index and overweight or obesity [7]. However, the present study focused on parental verbal support and guideline-level parental physical activity, as parental anthropometric data were not available.
Despite this, the literature remains limited in its comprehension of how these distinct modeling mechanisms interact, particularly in relation to child weight outcomes. Many studies have examined these pathways in isolation or aggregated them into a single 'parental support' metric [8]. This approach fails to capture the complexity of parental modeling, especially when 'words' and 'actions' are incongruent. Consequently, it remains unknown whether verbal encouragement, in the absence of behavioral modelling, is associated with lower odds of overweight or obesity. Conversely, it is unclear whether guideline-level parental physical activity shows a similar association when not accompanied by verbal reinforcement. This knowledge gap is particularly salient in the Chinese context, where intense academic pressures may create unique conflicts between parental verbal endorsement of PA and their own modeled behaviors [9].
To address this gap, this study utilized data from a large, population-based survey in Shanghai to estimate the associations between mutually exclusive parental role-modelling patterns and child overweight or obesity. These patterns were defined by the presence or absence of parental verbal encouragement and guideline-level parental physical activity. Furthermore, we explored whether these associations differed across key sociodemographic subgroups. We hypothesized that parental role modelling involving guideline-level parental physical activity would be associated with lower odds of child overweight or obesity, and that the combination of verbal support and guideline-level parental physical activity would show the lowest odds.
Methods
Study design and participants
This cross-sectional study analyzed data from the Shanghai site of the Physical Activity and Fitness in China—The Youth Study (PAFCTYS), a multiyear national surveillance programme conducted annually between October and November from 2016 to 2020. Although PAFCTYS is a national programme, the present analysis was restricted to the Shanghai site and was not designed to generate national prevalence estimates. At the Shanghai site, a multistage, stratified cluster sampling design was used to cover socioeconomic and urban–rural strata [10, 11]. Four urban districts and four rural towns were randomly selected from Shanghai’s 16 administrative divisions. Within each selected location, two primary schools, one junior middle school, and one senior high school were randomly chosen. One to two intact classes per grade (Grades 4–12) were sampled; Grades 1–3 were excluded because younger children may not reliably complete questionnaires.
Across five survey waves, 23,760 students aged 9–18 years and their parents were invited. After excluding incomplete questionnaires, missing anthropometric assessments, and implausible values based on predefined quality checks, 21,506 students and 20,183 parents provided valid data. One completed parental questionnaire was linked to each student using a unique student identifier. Following exclusion of dyads with missing exposure or outcome variables, 18,479 matched pairs were included in the final analytic sample. The study protocol was approved by the Ethics Committee of Shanghai University of Sport (No. 201737). School permission was obtained, and verbal parental consent and student assent were secured, with all data anonymised before analysis.
Procedures
Prior to data collection, parents and students received standardized instructions describing study procedures. At each school visit, trained research staff created an individual study file for every student, recorded demographic characteristics, and assigned a unique numeric identifier. Data collection occurred annually between October and November to align with the municipal fitness assessment schedule.
Anthropometric measurements were obtained in light clothing and without shoes. Height was measured to the nearest 0.1 cm using a calibrated portable stadiometer, and weight to the nearest 0.1 kg using a beam balance. Two examiners independently assessed height and weight, and the mean of the two measurements was used. Students completed a structured, self-administered questionnaire in supervised classroom settings, whereas one parent completed the corresponding questionnaire at home within one week. Both paper and online formats were used depending on school resources; validation studies within PAFCTYS have demonstrated equivalence between formats [12–14]. All questionnaire responses and physical fitness test records underwent double data entry into EpiData software, with automated range checks and manual review for inconsistencies. Records containing missing values, contradictory responses, or duplicated patterns were re-examined and resolved by senior supervisors prior to database lock.
Measures
Parental measures
Parental verbal support for children’s physical activity was assessed using a validated single-item question asking how often parents encouraged their child to participate in physical activity. Responses were recorded on a five-point Likert scale ranging from “never” to “always.” Consistent with previous PAFCTYS studies, responses of “sometimes,” “often,” or “always” were classified as verbal support, whereas “never” or “rarely” were classified as no verbal support [15].
Parental physical activity was assessed using the validated Chinese version of the International Physical Activity Questionnaire-Short Form (IPAQ-SF) [16]. Parents reported the frequency and duration of moderate- and vigorous-intensity activities lasting at least ten minutes during the previous week [11]. Weekly minutes of MVPA were calculated. IPAQ-SF data were cleaned according to international guidelines, and responses exceeding established duration limits or containing implausible values were excluded. Parents meeting the 2020 World Health Organization recommendation of at least 150 min of moderate activity, 75 min of vigorous activity, or an equivalent combination per week were classified as having guideline-level parental physical activity; those below this threshold were classified as not meeting guideline-level parental physical activity [17]. Parental measures were based on the responding parent, as separate maternal and paternal questionnaires were not available. The questionnaire did not reliably identify whether the responding parent was the mother or father; therefore, parental gender could not be used to define exposure groups or included as a covariate.
The four parental role-modelling categories were derived by cross-classifying two binary variables: parental verbal support and guideline-level parental physical activity. Participants were classified as no verbal support and no guideline-level parental physical activity (NS), verbal support only (VO), guideline-level parental physical activity only (AO), or both verbal support and guideline-level parental physical activity (BS).
Child weight status
Children’s anthropometric measurements were collected according to national fitness testing protocols. Body mass index (BMI) was calculated as weight (kg) divided by height (m2) [11, 18]. Weight status was defined using sex- and age-specific BMI cut-offs from the World Health Organization growth reference, classifying participants as overweight/obese versus normal weight [19]. Participants classified as underweight were not included in the binary weight-status outcome, which compared overweight/obesity with normal weight.
Student attitude toward physical activity
Covariates
Covariates were selected a priori based on established associations with youth weight status. Child-level variables included sex, exact age in years, grade level (4th to 12th), ethnicity (Han or minority), residential area (urban or rural), and only-child status. Survey year was included as a study-period variable. Parent and household-level variables included parental age, highest parental education level, and annual per-capita household income, categorised as < 9000, 9000 to 30,000, > 30,000 to 100,000, and > 100,000 RMB [15]. All variables were included as covariates in adjusted models. Dietary intake was not available in the harmonized dataset and therefore could not be included as a covariate. Parental height and weight were also unavailable; therefore, parental BMI or weight status could not be included in the adjusted models.
Statistical analysis
All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). The clustered sampling design was accounted for using cluster-robust standard errors at the school level. Sampling weights were not applied because this analysis was limited to linked parent–child dyads from the Shanghai site and valid expansion factors for the final analytic sample were not available. Descriptive characteristics of parents and children were summarised across the four parental role-modelling categories and presented as means with standard deviations (SD) for continuous variables or proportions with 95% confidence intervals (95% CI) for categorical variables. Group differences were examined using linear regression for continuous variables and Rao-Scott χ2 tests for categorical variables.
The primary analyses used cluster-adjusted logistic regression models to estimate odds ratios (ORs), adjusted odds ratios (aORs), and 95% CIs for child overweight/obesity, with normal weight as the reference outcome category. These models were restricted to participants classified as normal weight or overweight/obesity, with underweight participants excluded from the binary outcome analysis. In the full analytic sample, univariable models were first fitted for each predictor separately, followed by a multivariable model including parental role-modelling category and all covariates described above. Categorical covariates with more than two levels, including grade group, parental education, and household income, were entered as indicator variables. The reference categories were Grades 4 to 6, primary education or below, and household income < 9000 RMB, respectively. Model 1 reports univariable ORs, whereas Model 2 reports multivariable-adjusted odds ratios (aORs). Prespecified subgroup analyses were conducted by child gender, grade group (4th-6th, 7th-9th, 10th-12th), only-child status, residential area, and household income category to estimate subgroup-specific aORs. In these stratified models, the same adjustment set was used as in the primary model, except that the corresponding stratification variable was omitted. Formal effect modification was assessed separately in the full analytic sample by adding product terms between parental role-modelling category and each subgroup variable to the multivariable model. Wald tests were used to evaluate each product term, and the corresponding estimates, standard errors, Wald χ2 statistics, and P values are presented in Table 3. Because these interaction tests provided the formal assessment of effect modification, subgroup-specific estimates were interpreted descriptively. Because these interaction tests provided the formal assessment of effect modification, subgroup-specific estimates were interpreted descriptively.
Table 3.
Parental verbal and action support groups interacted with different variables in the analysis of students' overweight/obesity
| Variable | Category | Estimate | Standard Error | Wald Chi‑Square | P‑value |
|---|---|---|---|---|---|
| Gender | |||||
| NS group × Gender | Reference | Reference | Reference | Reference | |
| VO group × Gender | −0.09 | 0.11 | 0.76 | 0.3818 | |
| AO group × Gender | 0.05 | 0.17 | 0.09 | 0.7697 | |
| BS group × Gender | 0.02 | 0.13 | 0.03 | 0.8531 | |
| Grade | |||||
| NS group × Grade | Reference | Reference | Reference | Reference | |
| VO group × Grade | 0.03 | 0.06 | 0.29 | 0.5896 | |
| AO group × Grade | −0.13 | 0.1 | 1.62 | 0.2028 | |
| BS group × Grade | 0.02 | 0.07 | 0.08 | 0.7785 | |
| Only-child status | |||||
| NS group × Only-child status | Reference | Reference | Reference | Reference | |
| VO group × Only-child status | −0.26 | 0.13 | 3.91 | 0.0479 | |
| AO group × Only-child status | −0.23 | 0.21 | 1.23 | 0.2672 | |
| BS group × Only-child status | 0.12 | 0.15 | 0.64 | 0.4220 | |
| Residence | |||||
| NS group × Residence | Reference | Reference | Reference | Reference | |
| VO group × Residence | 0.22 | 0.1 | 4.27 | 0.0387 | |
| AO group × Residence | −0.06 | 0.17 | 0.14 | 0.712 | |
| BS group × Residence | −0.14 | 0.13 | 1.27 | 0.2602 | |
| Household income | |||||
| NS group × Household income | Reference | Reference | Reference | Reference | |
| VO group × Household income | 0.01 | 0.05 | 0.05 | 0.8269 | |
| AO group × Household income | −0.11 | 0.08 | 1.56 | 0.2111 | |
| BS group × Household income | −0.01 | 0.06 | 0.03 | 0.8667 | |
Parental support groups are contrasted with the reference category “NS group.”
Outcome: normal weight status, defined by age- and sex-specific BMI cutoffs from the Chinese National Student Physical Fitness Standards (2014 revision)
Model definitions: In the evaluation of interactions based on multivariable logistic regression, the following variables were adjusted for: child age, sex, residence, parental age, only-child status, household income per capita, survey year, school grade, ethnicity, and parental education
Complete-case analysis served as the primary analytic approach. Sensitivity analyses included excluding students in examination grades (9 and 12) and additionally adjusting for school level. Statistical significance was defined as a two-sided p-value < 0.05.
Results
Participant characteristics
Participant characteristics are shown in Table 1. The mean age of students was 13.23 (SD 2.64) years; 65.81% lived in urban areas. Girls comprised 52.08% of the sample and boys 47.92%. The proportions in Grades 4th–6th, 7th–9th, and 10th–12th were 35.62%, 30.69%, and 33.69%, respectively. Most students were Han ethnicity (98.80%). Mean parental age was 39.45 (SD 5.59) years; 77.67% of students were only children. Per-capita household income was most commonly > 30,00–100,000 RMB (38.46%), and 90.62% of parents had at least senior high school education. Mean child BMI was 19.89 (SD 3.66) kg/m2. After excluding underweight participants from the binary outcome analysis, 13,629 participants (76.80%) were classified as normal weight and 4,118 (23.20%) as overweight or obese.
Table 1.
Participant characteristics by parental role-modelling category
| Characteristic | All participants | NS group (22.42%) | VO group (42.05%) | AO group (8.74%) | BS group (26.79%) | P value |
|---|---|---|---|---|---|---|
| Survey characteristic | ||||||
| Survey year | 18,479 (100.00) | 4143 (100.00) | 7771 (100.00) | 1615 (100.00) | 4950 (100.00) | |
| 2016 | 3542 (19.17) | 821 (19.82) | 1423 (18.31) | 327 (20.25) | 971 (19.62) | <.0001 |
| 2017 | 3318 (17.96) | 763 (18.42) | 1172 (15.08) | 387 (23.96) | 996 (20.12) | |
| 2018 | 3333 (18.04) | 798 (19.26) | 1258 (16.19) | 360 (22.29) | 917 (18.53) | |
| 2019 | 4368 (23.64) | 843 (20.35) | 2138 (27.51) | 285 (17.65) | 1102 (22.26) | |
| 2020 | 3918 (21.20) | 918 (22.16) | 1780 (22.91) | 256 (15.85) | 964 (19.47) | |
| Child-level characteristics | ||||||
| Age | 13.23 ± 2.64 | 13.50 ± 2.64 | 13.18 ± 2.63 | 13.49 ± 2.67 | 12.99 ± 2.62 | <.0001 |
| Gender | 18,479 (100.00) | 4143 (100.00) | 7771 (100.00) | 1615 (100.00) | 4950 (100.00) | |
| Boys | 8855 (47.92) | 1847 (44.58) | 4032 (51.89) | 664 (41.11) | 2312 (46.71) | <.0001 |
| Girls | 9624 (52.08) | 2296 (55.42) | 3739 (48.11) | 951 (58.89) | 2638 (53.29) | |
| Grade | 18,479 (100.00) | 4143 (100.00) | 7771 (100.00) | 1615 (100.00) | 4950 (100.00) | |
| 4th to 6th graders | 6582 (35.62) | 1342 (32.39) | 2803 (36.07) | 512 (31.70) | 1925 (38.89) | <.0001 |
| 7th to 9th graders | 5671 (30.69) | 1216 (29.35) | 2413 (31.05) | 502 (31.08) | 1540 (31.11) | |
| 10th to 12th graders | 6226 (33.69) | 1585 (38.26) | 2555 (32.88) | 601 (37.21) | 1485 (30.00) | |
| Residential area | 18,371 (100.00) | 4119 (100.00) | 7732 (100.00) | 1605 (100.00) | 4915 (100.00) | |
| Urban | 12,090 (65.81) | 2582 (62.69) | 5180 (66.99) | 990 (61.68) | 3338 (67.91) | 0.0001 |
| Rural | 6281 (34.19) | 1537 (37.31) | 2552 (33.01) | 615 (38.32) | 1577 (32.09) | |
| Ethnicity | 18,479 (100.00) | 4143 (100.00) | 7771 (100.00) | 1615 (100.00) | 4950 (100.00) | |
| Han | 18,258 (98.80) | 4109 (99.18) | 7674 (98.75) | 1594 (98.70) | 4881 (98.61) | 0.1931 |
| Minority | 221 (1.20) | 34 (0.82) | 97 (1.25) | 21 (1.30) | 69 (1.39) | |
| Only-child status | 18,467 (100.00) | 4141 (100.00) | 7766 (100.00) | 1615 (100.00) | 4945 (100.00) | |
| Only child | 14,343 (77.67) | 3329 (80.39) | 6120 (78.81) | 1218 (75.42) | 3676 (74.34) | <.0001 |
| Non-only child | 4124 (22.33) | 812 (19.61) | 1646 (21.19) | 397 (24.58) | 1269 (25.66) | |
| BMI | 19.89 ± 3.66 | 20.21 ± 3.90 | 20.03 ± 3.79 | 19.98 ± 3.48 | 19.37 ± 3.21 | <.0001 |
| BMI for age z score | 0.21 ± 1.08 | 0.25 ± 1.15 | 0.29 ± 1.11 | 0.20 ± 1.02 | 0.10 ± 0.94 | <.0001 |
| Weight status | 17,747 (100.00) | 3955 (100.00) | 7508 (100.00) | 1559 (100.00) | 4725 (100.00) | |
| Normal | 13,629 (76.80) | 2934 (74.18) | 5542 (73.81) | 1222 (78.38) | 3931 (83.20) | <.0001 |
| Overweight or obesity | 4118 (23.20) | 1021 (25.82) | 1966 (26.19) | 337 (21.62) | 794 (16.80) | |
| Parent and household level characteristics | ||||||
| Parental age | 39.45 ± 5.59 | 39.69 ± 5.69 | 39.50 ± 5.55 | 39.22 ± 5.58 | 39.27 ± 5.55 | 0.7134 |
| Parental education | 14,806 (100.00) | 3294 (100.00) | 6295 (100.00) | 1275 (100.00) | 3942 (100.00) | |
| Primary or below | 1389 (9.38) | 314 (9.53) | 529 (8.40) | 189 (14.82) | 357 (9.06) | |
| Secondary to bachelor’s | 12,172 (82.21) | 2768 (84.03) | 5197 (82.56) | 991 (77.73) | 3216 (81.58) | |
| Postgraduate | 1245 (8.41) | 212 (6.44) | 569 (9.04) | 95 (7.45) | 369 (9.36) | |
| Household income | 18,330 (100.00) | 4106 (100.00) | 7709 (100.00) | 1600 (100.00) | 4915 (100.00) | |
| < 9,000 | 2275 (12.41) | 612 (14.91) | 900 (11.67) | 237 (14.81) | 526 (10.70) | <.0001 |
| 9,000 to 30,000 | 3792 (20.69) | 902 (21.97) | 1476 (19.15) | 407 (25.44) | 1007 (20.49) | |
| > 30,000 to 100,000 | 7049 (38.46) | 1648 (40.14) | 2952 (38.29) | 612 (38.25) | 1837 (37.38) | |
| > 100,000 | 5214 (28.45) | 944 (22.99) | 2381 (30.89) | 344 (21.50) | 1545 (31.43) | |
NS no verbal support and no guideline-level parental physical activity, VO verbal support only, AO guideline-level parental physical activity only, BS both verbal support and guideline-level parental physical activity, PA Physical Activity, BMI Body Mass Index, MVPA Moderate-to-Vigorous Physical Activity
Parental role-modelling categories, defined by verbal support and guideline-level parental physical activity, were distributed as follows: NS 22.42% (n = 4143), VO 42.05% (n = 7771), AO 8.74% (n = 1615), and BS 26.79% (n = 4950). Boys were more common than girls in the VO group, whereas girls predominated in NS, AO, and BS. The BS group decreased in proportion with higher grade level and showed the highest levels of household income > 100,000 RMB (31.43%) and postgraduate education (9.36%). Mean BMI was lowest in BS (19.37 [SD 3.21] kg/m2) and highest in NS (20.21 [SD 3.90] kg/m2). Overweight or obesity prevalence was highest in VO (26.19%) and lowest in BS (16.80%).
Parental role modeling and child overweight/obesity
Table 2 presents the full-sample logistic regression results for child overweight or obesity. Model 1 shows the univariable associations for each predictor, whereas Model 2 shows the multivariable-adjusted associations after all listed covariates were entered simultaneously. Compared with NS, lower odds were observed for AO (Model 1 OR 0.79, 95% CI 0.69–0.91; Model 2 aOR 0.82, 95% CI 0.69–0.96) and for BS (Model 1 OR 0.58, 95% CI 0.52–0.64; Model 2 aOR 0.57, 95% CI 0.50–0.65). VO showed no significant association after adjustment. Age, sex, residential area, survey year, and only-child status were also related to overweight or obesity, with lower odds observed at older ages and in later survey years, and among girls, rural students, and only children (Fig. 1).
Table 2.
Overall logistic regression models for child overweight/obesity in the full analytic sample
| Variable | Category | Model 1, OR (95% CI) | P value | Model 2, aOR (95% CI) | P value |
|---|---|---|---|---|---|
| Age (years) | per 1-year increase | 0.96(0.95–0.98) | <.0001 | 0.94(0.90–0.98) | 0.005 |
| Gender | Girls vs boys | 0.45(0.42–0.48) | <.0001 | 0.44(0.40–0.48) | <.0001 |
| Residential area | Urban vs rural | 1.54(1.43–1.65) | <.0001 | 1.81(1.66–1.97) | <.0001 |
| Ethnicity | Minority vs Han | 1.39(0.98–1.98) | 0.0684 | 1.57(0.98–2.53) | 0.0614 |
| Grade | per higher grade | 0.90(0.86–0.94) | <.0001 | 1.09(0.95–1.25) | 0.2225 |
| Survey year | per later year | 0.94(0.92–0.97) | <.0001 | 0.94(0.90–0.97) | 0.0008 |
| Parental age (years) | per 1-year increase | 1.00(0.99–1.00) | 0.5949 | 1.01(1.00–1.02) | 0.1155 |
| Parental education | per higher category | 0.96(0.87–1.05) | 0.3829 | 1.01(0.91–1.11) | 0.9066 |
| Only-child status | Only child vs non-only child | 0.93(0.85–1.01) | 0.0842 | 0.90(0.81–1.00) | 0.0473 |
| Household income | per higher category | 0.95(0.92–0.98) | 0.0052 | 0.98(0.93–1.02) | 0.2543 |
| Parental role-modelling category | NS group | 1 [Reference] | — | 1 [Reference] | — |
| VO group | 1.02(0.93–1.11) | 0.6679 | 0.98(0.88–1.09) | 0.7555 | |
| AO group | 0.79(0.69–0.91) | 0.0011 | 0.82(0.69–0.96) | 0.0158 | |
| BS group | 0.58(0.52–0.64) | <.0001 | 0.57(0.50–0.65) | <.0001 |
For continuous and ordinal predictors, ORs and aORs represent the change per 1-unit or one-category increase. Parental role-modelling categories were contrasted with NS as the reference category
Outcome: child overweight/obesity versus normal weight, defined using WHO age-specific and sex-specific BMI references. Higher ORs or aORs indicate higher odds of overweight/obesity
Model definitions: Model 1, cluster-adjusted univariable logistic regression fitted separately for each predictor. Model 2, cluster-adjusted multivariable logistic regression including child age, gender, grade, ethnicity, residential area, survey year, parental age, parental education, only-child status, household income, and parental role-modelling category
Abbreviations: OR odds ratio, aOR adjusted odds ratio, CI confidence interval, BMI body mass index, NS no verbal support and no guideline-level parental physical activity, VO verbal support only, AO guideline-level parental physical activity only, BS both verbal support and guideline-level parental physical activity
Fig. 1.
Flow diagram of participant selection and data inclusion
Figure 2 presents subgroup-specific aORs for parental role-modelling categories from stratified multivariable models. These estimates focus on the associations of VO, AO, and BS with overweight or obesity within each subgroup, using NS as the reference group. Formal interaction tests were conducted in the full analytic sample. The corresponding product-term models are presented in Table 3. No significant interaction was observed for BS with any subgroup variable, supporting the overall consistency of the BS association across strata. Two nominally significant interaction terms were observed for VO with only-child status and residential area, whereas the remaining interaction terms were not statistically significant. Because VO showed no consistent association in the primary or stratified models, these isolated interaction findings were interpreted cautiously, and subgroup estimates were considered descriptive.
Fig. 2.
Subgroup-specific adjusted associations between parental role-modelling patterns and child overweight or obesity
Subgroup analyses were largely consistent with the main findings. By gender, BS was associated with lower odds of overweight or obesity in boys (aOR 0.58, 95% CI 0.49 to 0.68) and girls (aOR 0.56, 95% CI 0.46 to 0.68), whereas VO and AO showed no significant association in either group. Across grade groups, BS remained significant in Grades 4 to 6 (aOR 0.62, 95% CI 0.50 to 0.76), Grades 7 to 9 (aOR 0.40, 95% CI 0.31 to 0.52), and Grades 10 to 12 (aOR 0.69, 95% CI 0.55 to 0.85). VO was non-significant in all grade groups. AO was significant only among students in Grades 7 to 9 (aOR 0.66, 95% CI 0.48 to 0.91), but not in Grades 4 to 6 or Grades 10 to 12.
By only-child status, BS was significant for only children (aOR 0.55, 95% CI 0.47–0.64) and non-only children (aOR 0.61, 95% CI 0.47–0.81). VO remained non-significant (only child aOR 1.03, 95% CI 0.92–1.16; non-only child aOR 0.82, 95% CI 0.64–1.04). AO was significant among non-only children (aOR 0.67, 95% CI 0.47–0.96) but not only children (aOR 0.87, 95% CI 0.72–1.04).
By residential area, BS was associated with lower odds of overweight or obesity in both urban students (aOR 0.60, 95% CI 0.51 to 0.71) and rural students (aOR 0.49, 95% CI 0.40 to 0.60), while VO and AO were not significant in either area. Across income strata, BS was significant for < 9000 RMB (aOR 0.49, 95% CI 0.33 to 0.71), 9000 to 30,000 RMB (aOR 0.58, 95% CI 0.43 to 0.77), > 30,000 to 100,000 RMB (aOR 0.64, 95% CI 0.52 to 0.78), and > 100,000 RMB (aOR 0.50, 95% CI 0.39 to 0.63). VO was non-significant in all income groups. AO was significant only in the > 100,000 RMB group (aOR 0.59, 95% CI 0.42 to 0.85), but not in lower-income groups.
Multiple-imputation sensitivity analyses were consistent with the main findings: AO (aOR 0.79, 95% CI 0.78–0.80) and BS (aOR 0.56, 95% CI 0.54–0.58) remained significant relative to NS, whereas VO did not (aOR 0.99, 95% CI 0.98–1.01).
Discussion
This cross-sectional study estimated the adjusted associations of four observed parental role-modelling patterns with overweight and obesity among children and adolescents in Shanghai. In addition to parental role-modelling patterns, several child-level and study-period variables were associated with overweight or obesity in the adjusted model. The findings can be summarized in three main points. First, in the full-sample adjusted model, guideline-level parental physical activity was associated with lower odds of overweight or obesity, particularly when combined with verbal support. Compared with the NS group, the AO group had lower odds of overweight or obesity, the BS group had the lowest odds, and the VO group showed no evidence of association. Second, parental role-modelling patterns displayed a graded distribution aligned with child weight outcomes: verbal-only support was the most common pattern, while the dual-support pattern was less prevalent; children in the BS group had the lowest mean BMI and the lowest prevalence of overweight or obesity, whereas those in the VO group had the highest prevalence. Third, subgroup-specific estimates showed that the inverse association for BS was broadly consistent across sex, grade group, residential area, only-child status, and household income. Although the estimates for AO appeared to vary across some subgroups, formal interaction tests were not statistically significant; these patterns should therefore be viewed as descriptive rather than evidence of confirmed effect modification. Collectively, these results suggest that the parental role-modelling pattern combining verbal encouragement with guideline-level parental physical activity was more strongly associated with lower odds of overweight or obesity than verbal encouragement alone.
Several covariate associations also warrant consideration. Older age was associated with lower odds of overweight or obesity after adjustment. This finding may partly reflect age-specific BMI references, pubertal growth, and age-related differences in weight-related behaviours; however, because age was examined cross-sectionally, it should not be interpreted as evidence that individual obesity risk decreases with age. Girls had substantially lower odds of overweight or obesity than boys, which is consistent with sex differences in adolescent growth patterns, body composition, and weight-related social or behavioural norms. Residential area was also important: urban students had higher odds of overweight or obesity than rural students, possibly reflecting greater exposure to energy-dense food environments, sedentary study and screen behaviours, and urban lifestyle patterns. Finally, later survey years were associated with lower odds of overweight or obesity. This period association should be interpreted cautiously because the analysis used repeated cross-sectional, unweighted Shanghai-site data; it may reflect changes in sample composition, local school health initiatives, or other secular factors rather than a causal time trend.
A key finding was that verbal support alone was not associated with child overweight or obesity, whereas guideline-level parental physical activity, particularly when combined with verbal encouragement, was associated with lower odds. This pattern introduces tension with a common interpretation in prior work that parental encouragement is itself a key driver of healthier weight. Many studies from Western settings and China report beneficial associations of “parental support” with youth physical activity or adiposity, yet these studies typically operationalize support as a composite that conflates encouragement, co-participation, logistic help, and parents’ own activity [8, 20]. Associations attributed to “parental support” in prior composite measures may partly reflect parents’ own activity patterns rather than verbal encouragement alone. By cross-classifying verbal support and guideline-level parental physical activity into mutually exclusive categories, this study compared “verbal support only” with patterns involving parents’ own guideline-level physical activity. The findings suggest that verbal support alone was not associated with lower odds of overweight or obesity in this cross-sectional sample. One plausible explanation is that children interpret exhortations from inactive parents as low-credibility or inconsistent guidance, which can weaken internalization of the health message through cognitive dissonance and reduced motivational salience [21]. In the Shanghai context, where academic pressure and time scarcity are normative, verbal prompts to be active may be further reframed by children as performance demands or symbolic advice detached from lived family priorities, potentially eliciting disengagement rather than sustained behavior change [22]. In contrast, parents who meet guideline-level physical activity may provide a more visible and consistent behavioural example, in line with social cognitive theory [23]. Such behaviour may also signal family commitment to health more clearly than verbal encouragement alone [24]. These considerations suggest that the association observed for guideline-level parental physical activity, especially when aligned with encouragement, is both methodologically clarified and behaviourally plausible in contemporary Chinese families.
Extending these findings, the contrast between AO and BS suggests that parental encouragement may be most relevant when supported by parents’ own guideline-level physical activity. Although AO was associated with lower odds of overweight or obesity than NS, the lower adjusted odds in the BS group suggest a stronger association when verbal support and guideline-level parental physical activity co-occurred. This pattern accords with previous evidence that children fare best when parental support is broader in scope, combining modelling with shared activity, practical facilitation, and positive emotional input [25]. A simple explanation is that parents’ own activity shows children what an active lifestyle looks like, while encouragement helps them understand why it matters and can be enjoyable, making the behavior easier to accept and sustain [26]. In practice, dual support may also reflect richer parent–child interaction around activity, including exercising together, arranging time and opportunities, providing access or equipment, and offering praise [27]. Through these combined pathways, aligned words and guideline-level parental physical activity may create a more consistent family health climate in which physical activity is a shared norm.
Seen in this light, the descriptive gradient across modelling patterns is especially informative. Verbal support only was the most prevalent pattern, whereas the pattern combining both verbal support and guideline-level parental physical activity was less common. Children and adolescents in the BS group had the lowest mean BMI and the lowest prevalence of overweight or obesity, while those in the VO group had the highest prevalence. This mismatch suggests that the most common support pattern was not the one most strongly associated with healthier weight status. The gradient also aligns with the concept of healthy lifestyle clustering, whereby favorable health behaviors tend to co-occur with more advantaged social conditions [28]. Prior studies indicate that families with higher socioeconomic status are more likely to show convergent health practices, including regular physical activity, healthier diets, and lower exposure to other lifestyle risks, rather than isolated supportive behaviors [29]. In this study, BS families had higher income and education, which may signal greater health literacy, more predictable routines, and more discretionary resources to support activity. By contrast, VO may capture parents who value physical activity but face constraints such as time scarcity, limited resources, or poorer health that restrict behavioral enactment. These graded differences underscore that parental role-modelling patterns are embedded in broader socioeconomic and psychological contexts, and they suggest that prevention efforts should move beyond encouraging parents to say more, toward helping families combine verbal support with achievable parental activity routines.
The subgroup analyses provided descriptive support for the main finding that BS showed the most consistent association across population groups. The interaction models further supported this interpretation, as no significant interaction was observed between BS and any examined subgroup variable. Although VO showed nominal interactions with only-child status and residential area, these findings should be interpreted cautiously because VO was not associated with overweight or obesity in the main model and multiple exploratory interaction terms were assessed. The BS pattern was associated with lower odds of overweight or obesity across the examined subgroups, including sex, grade group, only-child status, residential area, and household income, suggesting a relatively consistent association across population groups. In contrast, the association for AO appeared less consistent in stratified analyses, reaching significance only among students in Grades 7 to 9, among non-only children, and among those from higher-income households. Because formal interaction tests were not statistically significant, these subgroup differences should be interpreted cautiously. A developmental explanation is plausible: early adolescence is a period of increasing autonomy in which young people can act on observed parental habits while still being strongly shaped by parental norms, making behavioral modelling alone more influential than in younger children who depend on direct support, or older adolescents whose routines are more constrained by peers and academic demands [30]. Family structure may further modify these pathways, as siblings can amplify exposure to parents’ activity through shared play and mutual reinforcement, whereas only children may experience stronger academic supervision and fewer home-based activity partners, limiting the impact of modelling without encouragement [31, 32]. Socioeconomic resources offer a complementary lens. In higher-income settings, access to safe spaces, organized programmes, and equipment may allow children to convert parental example into their own activity, whereas in lower-income contexts, structural barriers may blunt the effect of modelling alone [33]. These patterns suggest that BS may provide a more consistent family-level signal than guideline-level parental physical activity alone, while the absence of statistically significant interactions limits any inference about true subgroup modification.
Several strengths and limitations warrant consideration. Strengths include the large PAFCTYS sample covering Grades 4 to 12, which provided substantial statistical power. The mutually exclusive NS, VO, AO, and BS classification is a key methodological contribution, because it distinguishes observed parental role-modelling patterns defined by verbal encouragement and guideline-level parental physical activity, allowing clearer presentation of associations than prior composite support measures. Child height and weight were measured objectively, and overweight or obesity was defined using WHO age- and sex-specific references, ensuring a standardized and internationally comparable outcome. Limitations should also be acknowledged. The cross-sectional design precludes causal inference and leaves potential for reverse causation and residual confounding, particularly from unmeasured dietary intake, which is closely related to child weight status. Parental activity and verbal encouragement were self-reported, and IPAQ-SF based estimates are susceptible to recall and social desirability biases. Importantly, the IPAQ-SF classification reflects whether parents met recommended physical activity levels, rather than the complete presence or absence of physical activity or directly observed parent–child role-modelling behaviour. In addition, parental height and weight were not collected, so we could not assess parental BMI or determine whether parental weight status confounded the observed associations. Because parental information was provided by one responding parent, and the respondent’s gender was not reliably recorded, we could not distinguish maternal from paternal role-modelling effects or adjust for parental gender. This may have introduced residual confounding and may have obscured parent-specific associations, particularly if maternal and paternal support or activity patterns differ in their influence on child weight status. Verbal support was measured with a single item that could not capture the quality or context of encouragement, and the action measure did not assess parent–child co-participation or situational modelling. Finally, the study was conducted in a highly urbanized setting, and the specific social and educational context may limit generalizability to less urbanized or lower-resource regions. In addition, because this analysis was based on unweighted Shanghai-site data, the findings should be interpreted as associations within this study population rather than national population estimates.
These findings have implications for family-based obesity prevention. At the family level, they suggest that consistent verbal support and parents’ own guideline-level physical activity may be a practical target for obesity prevention. Interventions should therefore help parents build stable activity habits themselves and promote parent–child physical activity as a shared routine, rather than focusing only on encouraging children. From a health equity perspective, the weaker and less consistent association for AO in lower-resource groups suggests that guideline-level parental physical activity alone may be insufficiently reflected in child weight status where opportunities for activity are limited. Policies should expand affordable and accessible spaces and programmes for family physical activity in rural and low-income communities so that parental example can translate into child behavior. In practice, pediatric and school health staff should routinely ask about parents’ own activity and support patterns, and schools can strengthen home-school links through simple parent–child activity initiatives to reinforce consistent messages across settings.
Conclusion
This large cross-sectional study suggests that parental role-modelling patterns involving guideline-level parental physical activity were associated with healthier weight status among children and adolescents, whereas verbal encouragement alone showed no evidence of association. Compared with NS, AO was associated with lower odds of overweight or obesity, and BS showed the lowest and most consistent odds across demographic groups. VO showed no evidence of association with lower odds of overweight or obesity. Obesity prevention efforts may therefore support parents in building sustainable active routines and sharing activity with their children, while recognizing that these findings reflect adjusted cross-sectional associations rather than causal or independent effects.
Supplementary Information
Acknowledgements
We thank the students, parents/guardians, teachers, and school staff for their participation and assistance with data collection and data entry.
Abbreviations
- WHO
World Health Organization
- PA
Physical Activity
- MVPA
Moderate-to-Vigorous Physical Activity
- PAFCTYS
Physical Activity and Fitness in China-The Youth Study
- IPAQ-SF
International Physical Activity Questionnaire-Short Form
- NS
No verbal support and no guideline-level parental physical activity
- VO
Verbal support only
- AO
Guideline-level parental physical activity only
- BS
Both verbal support and guideline-level parental physical activity
- OR
Odds Ratio
- aOR
Adjusted Odds Ratio
- CI
Confidence Interval
- SD
Standard Deviation
- IQR
Interquartile Range
- RMB
Renminbi (Chinese Yuan)
Authors’ contributions
JYW, SYW, and ZZ conceived the study design, collected and analyzed data, and drafted the manuscript; YY, YXL and SHG participated in the study design and statistical analysis; YT and YL contributed to data interpretation, provided critical feedback, and edited the manuscript; ZBC assisted with data management and supported the analytical work; PJC and JZ supervised the study and provided critical revisions to the manuscript. All authors have read and approved the final version of the manuscript, and agree with the order of presentation of the authors.
Funding
This study was supported by the Key Project of the National Social Science Foundation of China (No. 16ZDA227), the Shanghai Oriental Talent Program (JS2025098), and the Research and Innovation Grant for Graduate Students, Shanghai University of Sport (YJSCX-2025–025).
Data availability
No datasets were generated or analysed during the current study.
Declarations
Ethics approval and consent to participate
This study was approved by the Ethics Committee of Shanghai University of Sport. In accordance with national regulations, written informed consent was not required. Verbal informed consent was obtained from parents or legal guardians, and verbal assent was obtained from all participating students prior to data collection. All procedures were conducted in accordance with the Declaration of Helsinki.
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.
Contributor Information
Peijie Chen, Email: chenpeijie@sus.edu.cn.
Jie Zhuang, Email: zhuangjiesh@163.com.
References
- 1.Song, X., et al., Trends and predictions to 2030 in demographic structures and metabolic health for children and adolescents in China: analysis of national school health surveys from 2000 to 2019. Lancet Child Adolesc Health. 2025;9(8):530-543. [DOI] [PMC free article] [PubMed]
- 2.Luo D, et al. Long-term trends and urban-rural disparities in the physical growth of children and adolescents in China: an analysis of five national school surveys over three decades. Lancet Child Adolesc Health. 2023;7(11):762–72. [DOI] [PubMed] [Google Scholar]
- 3.Weihrauch-Blüher S, Schwarz P, Klusmann JH. Childhood obesity: increased risk for cardiometabolic disease and cancer in adulthood. Metabolism. 2019;92:147–52. [DOI] [PubMed] [Google Scholar]
- 4.Gråstén A, et al. Physical activity strategies for preventing school-aged children’s overweight and obesity: a systematic review. J Public Health Policy. 2025;46(3):484–502. [DOI] [PubMed] [Google Scholar]
- 5.Yaffe Y, Levental. Paternal and maternal parenting style and adolescents’ physical activity motivation: indirect effects through parental physical activity related practices. Curr Psychol. 2024;43(8):7297–307. [Google Scholar]
- 6.Petersen TL, et al. Association between parent and child physical activity: a systematic review. Int J Behav Nutr Phys Act. 2020;17(1):67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lee JS, Jin MH, Lee HJ. Global relationship between parent and child obesity: a systematic review and meta-analysis. Clin Exp Pediatr. 2022;65(1):35–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Carter JS, et al. Examining Direct and Indirect Mechanisms of Parental Influences on Youth Physical Activity and Body Mass Index. J Child Fam Stud. 2022;31(4):991–1006. [Google Scholar]
- 9.Wang J, et al. Association between moderate-to-vigorous physical activity trajectories and academic achievement in Chinese primary school children: a 3-year longitudinal study. BMC Public Health. 2025;25(1):341. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Chen P. Physical activity, physical fitness, and body mass index in the Chinese child and adolescent populations: An update from the 2016 Physical Activity and Fitness in China-The Youth Study. J Sport Health Sci. 2017;6(4):381–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Fan X, Cao ZB. Physical activity among Chinese school-aged children: National prevalence estimates from the 2016 Physical Activity and Fitness in China-The Youth Study. J Sport Health Sci. 2017;6(4):388–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Liu Y, et al. Associations between parental support for physical activity and moderate-to-vigorous physical activity among Chinese school children: A cross-sectional study. J Sport Health Sci. 2017;6(4):410–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chen ST, et al. Meeting 24-h movement guidelines: Prevalence, correlates, and the relationships with overweight and obesity among Chinese children and adolescents. J Sport Health Sci. 2021;10(3):349–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wang L, Tang Y, Luo J. School and community physical activity characteristics and moderate-to-vigorous physical activity among Chinese school-aged children: A multilevel path model analysis. J Sport Health Sci. 2017;6(4):416–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Xin F, et al. Prevalence and correlates of meeting the muscle-strengthening exercise recommendations among Chinese children and adolescents: Results from 2019 Physical Activity and Fitness in China-The Youth Study. J Sport Health Sci. 2022;11(3):358–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Craig CL, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003;35(8):1381–95. [DOI] [PubMed] [Google Scholar]
- 17.Bull FC, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020;54(24):1451–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ministry of Education of the People’s Republic of China; General Administration of Sport of China. National Student Physical Fitness and Health Standard (Revised 2014). Beijing, China. 2014.
- 19.de Onis M, et al. Development of a WHO growth reference for school-aged children and adolescents. Bull World Health Organ. 2007;85(9):660–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Epstein LH, et al. Family-Based Behavioral Treatment for Childhood Obesity Implemented in Pediatric Primary Care: A Randomized Clinical Trial. JAMA. 2023;329(22):1947–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Corriveau KH, et al. Do as I do, not as I say: Actions speak louder than words in preschoolers learning from others. J Exp Child Psychol. 2016;143:179–87. [DOI] [PubMed] [Google Scholar]
- 22.Zhang D, et al. Associations of physical activity with academic achievement and academic burden in Chinese children and adolescents: do gender and school grade matter. BMC Public Health. 2022;22(1):1496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Saunders RP, et al. A longitudinal examination of direct and indirect influences of parental behaviors on child physical activity. J Child Fam Stud. 2024;33(7):2262–70. [Google Scholar]
- 24.Xu W, et al. The effect of parental expectancy value beliefs on children’s physical activity: the chain mediating role of parental exercise commitment and children’s exercise self-efficacy. Front Psychol. 2025;16:1599121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Aleid AM, et al. The impact of parental involvement in the prevention and management of obesity in children: a systematic review and meta-analysis of randomized controlled trials. Children (Basel). 2024. 10.3390/children11060739. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Liu Y, et al. Physical activity maintenance and increase in Chinese children and adolescents: the role of intrinsic motivation and parental support. Front Public Health. 2023;11:1175439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Rhodes RE, Perdew M, Malli S. Correlates of parental support of child and youth physical activity: a systematic review. Int J Behav Med. 2020;27(6):636–46. [DOI] [PubMed] [Google Scholar]
- 28.Neshteruk CD, et al. Identifying patterns of physical activity and screen time parenting practices and associations with preschool children’s physical activity and adiposity. Prev Med Rep. 2020;18:101068. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Peral-Suárez Á, et al. Change in clusters of lifestyle behaviours from childhood to adolescence: a longitudinal analysis. Eur J Pediatr. 2024;183(10):4507–18. [DOI] [PubMed] [Google Scholar]
- 30.Teuber Z, et al. Autonomy-related parenting profiles and their effects on adolescents’ academic and psychological development: a longitudinal person-oriented analysis. J Youth Adolesc. 2022;51(7):1333–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kracht CL, Sisson SB. Sibling influence on children’s objectively measured physical activity: a meta-analysis and systematic review. BMJ Open Sport Exerc Med. 2018;4(1):e000405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Gao D, et al. The Association Between Single-Child Status and Risk of Abdominal Obesity: Result From a Cross-Sectional Study of China. Front Pediatr. 2021;9:697047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Andersen OK, et al. Socioeconomic position, built environment and physical activity among children and adolescents: a systematic review of mediating and moderating effects. Int J Behav Nutr Phys Act. 2022;19(1):149. [DOI] [PMC free article] [PubMed] [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
No datasets were generated or analysed during the current study.