A longitudinal, two-cohort study of parenting influences on adolescent physical activity and screen-time during the transition to secondary school in British Columbia

Abstract

Objective

Parents influence children's lifestyle behaviors, though this influence may decline as children gain autonomy. However, the COVID-19 pandemic likely altered these dynamics due to increased time at home. We examined whether associations between parenting practices and adolescents' physical activity and recreational screen-time changed during the transition to secondary school in two cohorts: one assessed pre-pandemic and one during the pandemic.

Methods

Data were collected in British Columbia (2018–2021) at two time points (elementary and secondary school) among 689 parent-adolescent dyads (parent age: 46 ± 5.4 years, 73 % mothers; child age: 13 ± 0.3 years, 52 % female). Forty-two percent were assessed pre-pandemic, and 58 % during the pandemic. Both cohorts self-reported parenting practices, recreational screen-time, and physical activity; non-pandemic children also wore accelerometers. Mixed-effect models with parenting-by-grade interaction terms assessed within-cohort changes over time.

Results

Supportive parenting predicted lower self-reported recreational screen-time and higher moderate-to-vigorous physical activity (MVPA) (non-pandemic: b = −25, 95 %CI:-38,-12 & b = 0.66, 95 %CI:0.38,0.93; pandemic: b = −19, 95 %CI:-30,-8 & b = 0.63, 95 %CI:0.32,0.93). In non-pandemic families, supportive parenting also predicted weekday light and MVPA (b = 12, 95 %CI:4,21 & b = 9, 95 %CI:4,14), though the latter significantly weakened post-transition (b = −11, 95 %CI:-19,-2).

Conclusions

Our findings highlight the importance of positive parenting in supporting children's active lifestyles, particularly during key life transitions.

Highlights

• •Supportive parenting predicts less screen-time and more physical activity in teens.
• •Effects varied by timing of school transition and pandemic-related restrictions.
• •Parenting effects are stronger when teens' behaviors are most at risk of decline.
• •Findings support parent-based strategies during transitions and external stressors.

1. Introduction

Adolescence, particularly the transition from elementary to secondary school, is a critical developmental period marked by psychosocial and environmental changes. During this time, physical activity declines, screen-time increases, and the risk of excess weight gain rises (Chong et al., 2020; Pearson et al., 2017).

The Canadian 24-Hour Movement Guidelines recommend at least 60 min of moderate-to-vigorous physical activity (MVPA) per day and no more than two hours of recreational screen-time for youth aged 5–17 (Tremblay et al., 2016). However, in a nationally representative sample of 22,115 Canadian youth aged 10–17, only 35 % met the activity guideline, and just 8 % met the screen-time recommendation (Janssen et al., 2017). Compliance declined steeply with age, particularly for recreational screen-time. Unlike other forms of screen-time such as educational (school/homework) screen-time, the use of screens for leisure (e.g., watching television, playing video games) is particularly concerning due to its links with negative health outcomes (Sanders et al., 2019). Systematic reviews have shown increased screen-time and sedentary behavior during the transition to secondary school (Pearson et al., 2017) and consistent declines in physical activity across school, after-school, and leisure-time contexts (Chong et al., 2020). Together, these findings point to a broader trend of increasingly inactive lifestyles during adolescence.

The Social-Ecological Model, widely applied both in child development and obesity research (Birch and Ventura, 2009; Bronfenbrenner, 1994), frames health behaviors as shaped by individual and contextual factors, including the family, school, and community. As children grow, they gain more autonomy over their behaviors, and interactions with these ecological layers evolve. For example, the transition to secondary school introduces new peer dynamics, changes in families' routines, and less parental involvement; all of which may shape lifestyle behaviors.

At the interpersonal level, parenting practices play a key role. Supportive parenting refers to a style of parenting that integrates both structure and autonomy support to create an environment where children feel both guided and empowered to be active. Through the use of these practices (e.g., encouragement, involvement, facilitation, modeling, co-participation, limit setting), supportive parenting is associated with higher activity levels (Gustafson and Rhodes, 2006; Trost and Loprinzi, 2011; Xu et al., 2015) and lower screen-time in youth (Xu et al., 2015). Yet, as children gain independence, parental influence may weaken, and findings across age groups are mixed: some studies report similar associations across childhood and adolescence (Trost and Loprinzi, 2011), while others find stronger effects in younger children (Gustafson and Rhodes, 2006).

Parent-child dynamics are also shaped by broader external factors. For example, the COVID-19 pandemic introduced widespread disruptions to family routines, parenting practices, school access, and opportunities for physical and social interactions. Studies reported declines in youth health behaviors during the pandemic, with restrictions limiting organized sports and indoor activities and increasing recreational screen-time (Dunton et al., 2020; Ellis et al., 2020; Saunders and Colley, 2024). During the pandemic in Canada, the proportion of older adolescents spending more than three hours per day on social media more than doubled (Ellis et al., 2020), and the percentage of adolescents meeting MVPA recommendations dropped by 10–15 % (Saunders and Colley, 2024).

Furthermore, increased time at home may have altered parent-child interactions in ways that either supported or hindered lifestyle behaviors. For instance, some Canadian studies (Moore et al., 2020, Moore et al., 2021) found that supportive parenting was linked to more physical activity and less screen-time during the pandemic. However, elevated parental stress might have increased coercive practices, which negatively affect children's health behaviors (Lucassen et al., 2021; Park et al., 2021; Seguin et al., 2021).

While previous work has explored behavioral changes during the school transition, pandemic-related behavioral shifts, or the role of supportive parenting, their combined effects remain unclear. This longitudinal study examined how parenting practices relate to adolescents' physical activity and recreational screen-time during the transition to secondary school in two cohorts: one exposed to pandemic conditions and one not. We hypothesized that supportive parenting would predict higher physical activity and lower screen-time, with associations weakening post-transition but strengthening under pandemic conditions due to more time at home.

2. Methods

2.1. Design

This study used data from the HABITs (Health And Behaviors In Teens) cohort study approved by the University of British Columbia Research Ethics Board (H15–018). All participants signed informed consent forms prior to participating in this study. HABITs examined determinants of adolescents' health behaviors during the transition from elementary (grade seven) to secondary (grade eight) school in British Columbia, Canada. A total of 689 parent-child dyads were followed prospectively for one year. Of these, 287 dyads (42 %) completed data collection before pandemic-related school closures began (17/03/2020). The remaining (n = 402, 58 %) participated during the pandemic, when schooling was fully online in grade seven and followed a hybrid model in grade eight. Fig. 1 shows the study timeline.

Fig. 1.

Timeline of recruitment and data collection by grade and cohort in British Columbia, Canada (2017–2021), with relative sample sizes.

2.2. Recruitment and data collection

Participants were recruited from 27 public schools in Vancouver, West Vancouver, Burnaby, Surrey, and Delta. Following district approval, school principals received informational packages, and research staff delivered in-class presentations to students, who then took home consent packages to be signed by themselves and their parents. Eligible students were in grade seven at schools where this was the final grade, could read and write at a grade six level or above, and had a participating parent/guardian. In the non-pandemic cohort, grade seven students completed an online questionnaire at school and wore an accelerometer for eight days, while grade eight data collection occurred at community centers or at the British Columbia Children's Hospital Research Institute, with accelerometers being mailed to their homes. Non-pandemic parents and pandemic parents and children completed all questionnaires online. Families received nominal compensation for participation.

2.3. Measures

Media parenting practices (MPP), measured with five items from the FLASHE study (Nebeling et al., 2017), assessed agreement with practices such as modeling, setting rules, and monitoring screen-time. Responses ranged from 0 to 4, with higher scores indicating greater agreement. Factor analysis supported a 1-factor structure in grade seven and eight: Steiger's Root Mean Square Error of Approximation (RMSEA) 95 %CI = 0.06,-0.12 & 0.05,-0.12, Comparative Fit Index (CFI) = 0.97 & 0.98, Standardized Root Mean Square Residual (SRMR) = 0.04 & 0.03, Cronbach's α = 0.76 & 0.79.

Physical activity parenting practices (PAPP), measured using four FLASHE items (Nebeling et al., 2017) and one from the PAPP Item Bank (Mâsse et al., 2020), assessed facilitation, rules, monitoring, and perceived importance of physical activity. Responses ranged from 0 to 4, with higher scores indicating greater agreement. Factor analysis supported a 1-factor structure in grade seven and eight: RMSEA 95 %CI = 0.04,-0.10 & 0.00,-0.08, CFI = 0.97 & 0.99, SRMR = 0.03 & 0.03, Cronbach's α = 0.72 & 0.70.

Recreational screen-time on weekdays was assessed with one item from the Take Action Survey (French et al., 2011). Students self-reported time spent using a computer, tablet or cellphone outside of school/paid work during the most recent weekday. Response options ranged from 0 to 240+ minutes.

Physical activity was measured as self-reported days/week with ≥60 min of MVPA (based on the Physical Activity Questionnaire for Children (Kowalski et al., 1997)), and accelerometer-based daily minutes of light and MVPA. Only the non-pandemic cohort wore Axivity AX3 accelerometers (Axivity Ltd., Newcastle University, UK) for eight days (100 Hz, ±8 g). Three reviewers used OmGui software (Open Movement & Newcastle University, UK) and activity diaries to identify non-wear (transportation) periods at the start/end of the data collection. Day one (afternoon) and day eight (morning) were merged as they occurred on the same weekday. Finally, data were processed in R using the GGIR package (Migueles et al., 2019), with intensity classified via validated cut-points for 8–14-year-olds (light: 7-19 g; MVPA: ≥20 g (Phillips et al., 2013)).

2.4. Statistical analyses

Baseline cohort differences were assessed using two-sample Student t-tests and Chi-square tests (X²). Paired t-tests examined within-cohort changes over time. Linear mixed-effect models tested first the main effects of parenting practices on adolescents' health behaviors, followed by parenting-by-grade interaction terms to assess change over time. Interactions were then decomposed into lower-order components, and simple effects were graphed for interpretation. Based on theoretical expectations, simple effects were explored in all models regardless of interaction significance (Aiken et al., 1991). While statistical significance was set at p < 0.05, we also assessed the public health significance of behavioral changes using predefined thresholds representing weak, moderate, and strong effects. These thresholds correspond to estimated annual weight changes of approximately one, one and a half, or two pounds, respectively. Supplemental Table 1 details the cut-offs for recreational screen-time, light activity, and MVPA by weekday type. These were calculated using standard energy expenditure formulas and metabolic equivalent values (Butte et al., 2018), with annual energy increases converted to weight change estimates. All models were stratified by cohort and adjusted for adolescent and parent sex, parent age, education, household income, and ethnicity/race. Missing data ranged from 2 to 16 %. Analyses were conducted in Stata (StataCorp, College Station, TX: StataCorp LLC).

3. Results

Table 1 shows demographic characteristics at baseline, with no differences between cohorts.

Table 1.

Demographic characteristics of grade seven students in British Columbia, Canada, comparing non-pandemic (2018–2019) and pandemic (2020) cohorts.

Demographics		Total N = 689		Non-Pandemic N = 287		Pandemic N = 402
		n (%) / mean ± SD		n (%) / mean ± SD		n (%) / mean ± SD
Child sex	Male	317	(47)	138	(48)	179	(46)
	Female	358	(53)	148	(52)	210	(54)
Child age		12.9	± 0.4	12.8	± 0.5	12.9	± 0.3
Parent sex	Male	156	(24)	76	(27)	80	(22)
	Female	490	(76)	201	(73)	289	(78)
Parent age		46.1	± 5.3	46.3	± 5.1	46.0	± 5.4
Education	Non-university degree	281	(43)	118	(43)	163	(44)
	University degree	368	(57)	159	(57)	209	(56)
Marital status	Married / Common-law	83	(13)	39	(14)	44	(12)
	Single / Divorced / Widowed	552	(87)	233	(86)	319	(88)
Household income ($1CAD = ∼$.77USD)	<$70,000 CAD	178	(31)	80	(32)	98	(30)
	$70,000–$125,000 CAD	198	(34)	96	(38)	102	(31)
	>$125,000 CAD	202	(35)	75	(30)	127	(39)
Ethnicity / Race	Caucasian / White	235	(36)	105	(38)	130	(35)
	East & Southeast Asian	219	(34)	93	(34)	126	(34)
	South Asian	100	(16)	43	(16)	57	(15)
	Other	91	(14)	36	(13)	55	(15)

Notes: % may not add up to 100 % due to rounding. No significant differences emerged between samples at baseline except for child age due to temporality, as the pandemic sample had data collected slightly later during the year. CAD = Canadian dollars. USD = United States of America's dollars. CAD to USD conversion rate based on the exchange rate in October 2017, the month recruitment began.

Table 2 describes health behaviors and parenting practices in elementary school, as well as unadjusted changes after the transition to secondary school. Compared to their non-pandemic peers, the pandemic cohort reported lower supportive MPP and higher screen-time at baseline. During the transition, screen-time increased in the non-pandemic group but decreased in the pandemic group. Supportive PAPP and self-reported physical activity declined in both cohorts, as did accelerometer-measured physical activity in the non-pandemic cohort.

Table 2.

Descriptive statistics on parenting practices and adolescents' health behaviors before and after the transition from elementary to secondary school in British Columbia, Canada, comparing non-pandemic (2018–2020) and pandemic (2020−2021) cohorts.

	Non-Pandemic N = 287		Pandemic N = 402
	Elementary school	Change after the transition	Elementary school	Change after the transition
	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD
Parenting practices (0–4 range)
Supportive media parenting practices	2.9 ± 0.7 a	−0.2 ± 0.5 ↓	2.7 ± 0.6 a	−0.1 ± 0.6 ↓
Supportive physical activity parenting practices	2.8 ± 0.6	−0.1 ± 0.5 ↓	2.8 ± 0.6	−0.1 ± 0.6 ↓
Adolescents' behaviors
Recreational screen-time, min/weekday	88 ± 68 a	31 ± 78 ↑	170 ± 77 a	−20 ± 95 ↓
Self-reported days/week with 60+ MVPA minutes	3.8 ± 1.8	−0.3 ± 1.8 ↓	3.9 ± 2.3	−0.3 ± 2.4 ↓
AX3-MVPA, minutes/weekday	80 ± 32	−14 ± 30 ↓	–	–
AX3-MVPA, minutes/weekend day	52 ± 32	−7 ± 30 ↓	–	–
AX3-Light physical activity, minutes/weekday	205 ± 50	−29 ± 50 ↓	–	–
AX3-Light physical activity, minutes/weekend day	178 ± 67	−34 ± 65 ↓	–	–

Notes: ^a Significant difference (p < 0.05) between samples. ↓ or ↑ Significant (p < 0.05) decrease or increase, respectively, within a sample. – Data unavailable. MVPA = Moderate-to-vigorous physical activity. AX3 refers to the type of accelerometer used (i.e., Axivity AX3 accelerometers).

Table 3 summarizes associations between supportive parenting and adolescents' health behaviors. Prior to the pandemic, transitioning to secondary school was linked to declines in health behaviors: weekday screen-time rose by 30 min/day, weekly MVPA days dropped by ∼1/3 day, and accelerometry-based MVPA declined by 14 and seven minutes on weekdays and weekends, respectively, and by 29 and 31 min for light activity. In contrast, screen-time declined by 20 min/day during the pandemic, and no changes were observed in self-reported MVPA.

Table 3.

Changes in adolescents' health behaviors during the transition from elementary to secondary school in British Columbia, Canada (2018–2021), associated with the transition itself and with a one-unit improvement in parenting practices.

	Main effects
	School transition		Supportive parenting
	b	[95 %CI]	b	[95 %CI]
Non-Pandemic, n = 287
Recreational screen-time, minutes/weekday	30	[16,44]***	−25	[−38,-12]***
Self-reported days/week with 60+ MVPA min	−0.36	[−0.61,-0.12]**	0.66	[0.38,0.93]***
AX3-MVPA, minutes/weekday	−14	[−19,-9]***	9	[4,14]**
AX3-MVPA, minutes/weekend day	−7	[−13,−1]*	6	[−0.5,12]
AX3-Light physical activity, minutes/weekday	−29	[−37,-20]***	12	[4,21]**
AX3-Light physical activity, minutes/weekend day	−31	[−43,-19]***	8	[−5,21]
Pandemic, n = 402
Recreational screen-time minutes/weekday	−20	[−31,-8]**	−19	[−30,-8]**
Self-reported days/week with 60+ MVPA minutes	−0.22	[−0.52,0.08]	0.63	[0.32,0.93]***

Notes: b = unstandardized regression coefficients. 95 %CI = 95 % confidence intervals. Stars indicate statistical significance, where * p < 0.05, ** p ≤ 0.01, *** p ≤ 0.00. MVPA = moderate-to-vigorous physical activity. All models included the following covariates: adolescents' and parents' sex, parents' age, education, household income and race/ethnicity. AX3 refers to the type of accelerometer used (i.e., Axivity AX3 accelerometers).

Associations between supportive parenting and health behaviors were similar in both cohorts. A one-unit improvement in MPP was associated with 25- and 19-min reductions in weekday screen-time in the non-pandemic and pandemic cohorts, respectively. A one-unit increase in supportive PAPP was linked to an increase in MVPA of ∼2/3 day in both cohorts. In the non-pandemic cohort only, more supportive PAPP also predicted increases of 12 and nine weekday minutes in light and MVPA, respectively.

Table 4 and Fig. 2 show the parenting-by-grade interactions. For MPP (Fig. 2a), effects on screen-time were stronger in secondary school in the non-pandemic cohort (screen-time reductions of 36 vs. 18 min/day), and in elementary school for the pandemic group (screen-time reductions of 27 vs.12 min/day). Though the parenting-by-grade interaction was not significant in either cohort, it was borderline during the pandemic (p = 0.09).

Table 4.

Interaction between school transition and parenting practices on adolescents' health behaviors in British Columbia, Canada, comparing non-pandemic (2018–2020) and pandemic (2020–2021) cohorts.

	Simple Effects					Interaction
	School transition at parenting = 0 a		Supportive parenting by grade b			Supportive parenting * school transition c
	b	[95 %CI]	Grade	b	[95 %CI]	b	[95 %CI]
Non-Pandemic N = 278
Recreational screen-time, minutes/weekday	77	[19,136]**	7	−18	[−34,-3]*	−17	[−38,4]
			8	−36	[−54,-17]***
Self-reported days/week with 60+ MVPA minutes	−1.39	[−2.55,-0.23]*	7	0.50	[0.18,0.82]**	0.38	[−0.04,0.79]
			8	0.88	[0.51,1.24]***
AX3-MVPA, minutes/weekday	14	[−9,39]	7	13	[6,19]***	−11	[−19,-2]*
			8	2	[−6,10]
AX3-MVPA, minutes/weekend day	−4	[−33,24]	7	6	[−1,13]	−1	[−11,9]
			8	5	[−4,14]
AX3-Light physical activity, minutes/weekday	−10	[−51,31]	7	15	[5,25]**	−7	[−22,8]
			8	8	[−5,21]
AX3-Light physical activity, minutes/weekend day	15	[−45,76]	7	14	[−0.7,29]	−17	[−39,5]
			8	−3	[−22,16]
Pandemic N = 402
Recreational screen-time, minutes/weekday	−61	[−111,−11]*	7	−27	[−41,-13]***	16	[−3,34]
			8	−12	[−26,3]
Self-reported days/week with 60+ MVPA minutes	0.43	[−0.96,1.83]	7	0.74	[0.35,1.13]***	−0.24	[−0.74,0.26]
			8	0.50	[0.09,0.90]*

Notes: b = unstandardized regression coefficients. 95 %CI = 95 % confidence intervals. Stars indicate statistical significance, where * p < 0.05, ** p ≤ 0.01, *** p ≤ 0.00. MVPA = moderate-to-vigorous physical activity. ^a = Change in behaviors in grade 8 (secondary school) relative to grade 7 (elementary school) when parenting practice has the lowest value (=0). ^b = Change in behaviors, decomposed by grade, for a one unit increase in supportive parenting. ^c = Difference in slope in supportive parenting predicting behaviors between grades. All models included the following covariates: adolescents' and parents' sex, parents' age, education, household income and race/ethnicity. AX3 refers to the type of accelerometer used (i.e., Axivity AX3 accelerometers).

Fig. 2.

Decomposed interactions between supportive parenting and school transition on adolescents' health behaviors, comparing non-pandemic (2018–2020) and pandemic (2020–2021) cohorts in British Columbia, Canada.

Supportive PAPP was consistently associated with higher self-reported MVPA (Fig. 2b), with weekly increases ranging from 1/2–9/10 of a day. Although parenting-by-grade interactions were not significant in either cohort, the effect approached significance among non-pandemic families (p = 0.08).

PAPP also predicted greater accelerometry-measured light and MVPA on weekdays during elementary school only, with increases of 15 and 13 min/day, respectively (Fig. 2c-d). These associations were not observed after the transition. Yet, the parenting-by-grade interaction was significant only for weekday MVPA (p = 0.02), reflecting an 11-min difference in slopes by grade.

4. Discussion

This study examined the longitudinal effects of supportive parenting on adolescents' recreational screen-time and physical activity during the transition to secondary school, before and during the pandemic. Grounded in the Social-Ecological Model (Birch and Ventura, 2009; Bronfenbrenner, 1994), our findings illustrate how behavioral trajectories are shaped by the family system, developmental transitions, and contextual disruptions. Across cohorts, declines were observed in parenting support and youth health behaviors during the school transition. However, supportive parenting generally buffered against these declines, with several associations reaching moderate-to-strong effects.

In the non-pandemic cohort, adolescents' screen-time increased during the transition, consistent with previous studies showing unhealthy shifts at this time (Chong et al., 2020; Emke et al., 2023; Pearson et al., 2017). This increase coincided with a drop in supportive MPP, predicting higher screen-time both pre- and post-transition. As children gain autonomy, parents may relax screen-time rules and expectations to avoid conflicts (Zhang et al., 2018). The interaction between MPP and grade was not significant, suggesting a consistent effect of parenting over time. Interestingly, the association was slightly stronger in secondary school where screen-time was higher, possibly reflecting increased parental engagement to mitigate behavioral shifts triggered by the transition. These findings reflect the Social-Ecological Model's emphasis on the continued importance of the family-level influences (Birch and Ventura, 2009; Bronfenbrenner, 1994), even as adolescents become more autonomous.

In contrast, the pandemic cohort showed a screen-time decline during the transition, though levels remained elevated compared to pre-pandemic norms (Colley and Saunders, 2023; Colley and Watt, 2022; Ellis et al., 2020; Moore et al., 2021; Stockwell et al., 2021). Supportive MPP were lower than in the non-pandemic cohort at both time points. Despite a borderline-significant interaction, MPP no longer predicted screen-time after the transition. These patterns align with research grounded in the Family Stress Model (Masarik and Conger, 2017), which suggests that stress (e.g., from job loss or disrupted routines) undermines parenting quality and increases coercive practices (Hartshorne et al., 2021; Lucassen et al., 2021; Park et al., 2021; Seguin et al., 2021). Parental efforts may also have declined post-transition due to expectations that in-person schooling would naturally reduce screen-time. Our findings suggest that once high screen-time patterns are established, they are difficult to reverse. A recent review warned of a slow or limited recovery in family dynamics post-pandemic, with ‘residual’ stressors continuing to disrupt family functioning (Campione-Barr et al., 2024). As suggested, many studies assume that stress dissipated once restrictions were lifted, but long-term recovery remains uncertain. The lack of MPP effects in secondary school within the pandemic cohort underscores this challenge, and highlights the need for family-focused strategies to address entrenched behaviors. These findings also reinforce how broader ecological disruptions can interact with family processes to shape adolescent health trajectories.

In the non-pandemic cohort, physical activity declined after the transition, echoing prior findings attributing these drops to academic demands, shifting peer norms, increased autonomy, and environmental changes (Chong et al., 2020; Emke et al., 2023; Pearson et al., 2017). In the pandemic cohort, self-reported physical activity remained stable across the transition, and was similar to the non-pandemic cohort at baseline. Most pandemic-related declines in physical activity have been linked to reliance on school- and sport-based activities and their suspension (Do et al., 2022). However, in British Columbia, outdoor activities were encouraged during the pandemic, which might have buffered these effects and underscores the role of the community-level environment within the Socio-Ecological Model (Birch and Ventura, 2009; Bronfenbrenner, 1994) in shaping health behaviors.

Supportive PAPP were consistently associated with higher self-reported physical activity in both cohorts, with no significant interactions by grade. This suggests that the benefits of supportive PAPP persist across the school transition. Similar to the screen-time results, the associations were slightly stronger during timepoints when adolescents' behaviors were more adversely impacted. In the non-pandemic sample, this occurred after the transition; in the pandemic sample, it occurred earlier, when school closures and restrictions began. These findings align with the Socio-Ecological perspective (Birch and Ventura, 2009; Bronfenbrenner, 1994) and reinforce the protective role of parents in promoting healthy behaviors amid both developmental and external disruptions.

For accelerometry-measured physical activity (available only for the non-pandemic cohort), supportive PAPP was associated with greater light and MVPA on weekdays. However, these associations were no longer significant after the school transition, contrasting with self-reported results. Discrepancies may reflect measurement differences: accelerometry assessed minutes/day by weekday type, while self-report captured days/week with at least 60 min of MVPA. The decline in accelerometer-based activity mirrored a drop in parental support, and may reflect decreased parental expectations (Zhang et al., 2018) under assumptions that schools would take greater responsibility for physical activity promotion during secondary school. A systematic review found that, while parental support remains important for children's physical activity across different ages, its influence weakens as children grow older (Gustafson and Rhodes, 2006). Our objective data support this trend but primarily for weekdays, likely reflecting that younger children are more commonly enrolled in structured physical activities during the week. These findings underscore the importance of helping parents maintain their role in supporting physical activity post-transition to mitigate expected declines.

4.1. Limitations and strengths

This study has limitations and strengths. We lacked accelerometry data for the pandemic cohort during elementary school, limiting objective comparisons across cohorts. Parenting practices were reported by one caregiver, potentially omitting the influence of other family members. While the Social-Ecological Model informed our approach, the study primarily focused on how specific external factors interact with the family system. Other relevant influences such as peer and neighborhood variables were not assessed and warrant further study. Nonetheless, the study has several strengths, including its longitudinal design, the inclusion of both pre-pandemic and pandemic cohorts, use of self-reported and accelerometer-based measures, and a focus on a critical developmental transition. Moreover, the pandemic cohort was recruited before any awareness of the pandemic, which resulted in valid cross-cohort comparisons due to comparable socio-demographic profiles. In addition, data were collected in the spring from all study participants, reducing seasonality effects. Finally, the demographic diversity of the sample also enhances generalizability.

5. Conclusions

Supportive parenting was associated with healthier screen-time and physical activity patterns during adolescence, particularly when behaviors were most negatively affected. In the non-pandemic cohort, most benefits were sustained (and even slightly stronger) after the transition. In the pandemic cohort, effects were more evident earlier, during peak restrictions. These findings highlight the value of supportive parenting in mitigating behavioral declines stemming from both developmental and environmental stressors. Interventions should equip parents with tools to sustain positive practices through major life transitions to promote adolescent health and well-being.

Disclosure of funding and conflicts of interest

The HABITs study was funded by the Canadian Institutes of Health Research via grant provided to LCM. OD-JG receives a postdoctoral fellowship from the Michael Smith Foundation for Health Research. LCM receives salary support from the British Columbia Children's Hospital Research Institute and the University of British Columbia. The authors have no additional funding and conflicts of interest to declare.

CRediT authorship contribution statement

Olivia De-Jongh González: Writing – original draft, Visualization, Formal analysis, Data curation, Conceptualization. Iyoma Y. Edache: Writing – original draft, Data curation. Claire N. Tugault-Lafleur: Writing – review & editing. Louise C. Mâsse: Writing – review & editing, Project administration, Funding acquisition, Conceptualization.

Ethics approval and consent to participate

Data for this study were collected as part of the HABITs longitudinal study approved by the University of British Columbia Research Ethics Board (H15–018). Participants signed informed consent forms prior to participating in this study.

Funding

The HABITs study was funded by the Canadian Institutes of Health Research via grant provided to LCM. OD-JG receives a postdoctoral fellowship from Michael Smith Foundation for Health Research. LCM receives salary support from the British Columbia Children's Hospital Research Institute and the University of British Columbia.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:(Louise C. Masse reports financial support was provided by Canadian Institutes of Health Research. Olivia De-Jongh Gonzalez reports financial support was provided by Michael Smith Foundation for Health Research. Louise C. Masse reports financial support was provided by BC Children's Hospital Research Institute. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.)

Appendix A. Supplementary data

Predetermined thresholds for determining and interpreting effect sizes in adolescent health behaviors

Data availability

The data used in this study are not publicly available. Data are however available upon reasonable request to the corresponding author, and with permission of the Research Ethics Board at the University of British Columbia.