This study, supported by the local education commission of Guangxi Zhuang Autonomous Region, was conducted between July and August 2024 in 12 public primary schools across two cities with a high concentration of the Zhuang ethnic group. Previous research indicated that children aged 10 years or older possess adequate reading comprehension skills to complete self-report questionnaires (44). In China, fourth-grade students are typically around 10 years old. Therefore, students in the fourth grade and above were included as qualified participants.

The aim of the study was presented to the participants and their guardians in a written format. This survey is intended to collect data about demographic information, self-reported PA, screen time, sleep duration and PF. Participants voluntarily completed the questionnaire using electronic devices (e.g., computers or mobile phones). Students without personal access to electronic devices were encouraged to use their parents’ equipment. If electronic devices were unavailable, students completed a paper-based questionnaire, and school staff manually converted into an electronic format. Only participants who provided consent were able to proceed to the formal data collection pages. All data were collected and analyzed anonymously. The study received ethical approval from Wuhan Sports University (No. 2025019).

A total of 10,170 participants completed the questionnaire. We excluded those who (i) completed the questionnaire in < 300 s, (ii) were neither Han nor Zhuang, or (iii) had implausible values for age, physical activity, screen time, or sleep. After considering the availability of the necessary variables, the final sample for analysis comprised 6,913 students, as shown in Figure 1.

All statistical analyses were carried out with the SPSS software (version 26, IBM Corporation, Armonk, NY, USA). Descriptive statistical methods were employed to present the basic characteristics of the participants. For continuous variables (e.g., age), data were reported as mean ± standard deviation (SD), while categorical variables (e.g., ethnicity and residence) were presented as counts and percentages.

Two sets of pairwise regression models were constructed. In the first set, all possible pairwise comparisons were conducted across levels of the number of 24-h movement guidelines met (0, 1, 2, or 3), resulting in six independent models (e.g., 3 vs. 2, 3 vs. 1, 3 vs. 0, etc.), with p-values adjusted using the Šidák method. In the second set, we examined all possible pairwise contrasts among the eight possible combinations of guideline adherence (e.g., all vs. sleep only, MVPA only vs. None, etc.), yielding 28 additional independent models. Given the large number of comparisons, the Holm–Bonferroni method was applied to balance rigorous error control with statistical power. To explore whether the association between guideline adherence and PF differed by ethnicity or residence, interaction terms were included in the primary models. Generalized linear models (GLMs) with normal distribution and identity link function were employed to estimate the aforementioned relationships (54–56). The unstandardized regression coefficient (B) represents the adjusted mean difference in PF scores between the compared groups, and statistical significance was defined as p < 0.05 after adjustment for multiple comparisons. Given the observed SD of PF scores (SD = 0.81, as shown in Supplementary Table S1), a B value of ≥ 0.405 (0.5 SD) was considered a moderate, potentially meaningful change, and ≥ 0.162 (0.2 SD) a small but detectable change, based on established benchmarks for patient-reported outcomes (57, 58).

The descriptive characteristics of the participants, classified by residence and ethnicity, are presented in Table 1. The sample consisted of 21.5% Han ethnicity (45.1% girls) and 78.5% Zhuang ethnicity (49.9% girls), with a mean age recorded of 11.6 years (SD = 1.0), and 63.0, 18.0, and 19.0% of participants lived in rural, suburban and urban regions, respectively. Regarding the exposure variables, 28.0% of the participants did not meet any guideline, while only 1.6% met all three guidelines. 2.4, 10.1, and 37.9% of participants met only the MVPA guideline, the screen time guideline, and the sleep guideline, respectively. A greater proportion of Han participants met all three guidelines, but a lower proportion met guideline for sleep duration. A greater proportion of rural participants met guideline for sleep duration, but a lower proportion met guideline for MVPA (see Supplementary Table S2 for details). Regarding the outcome variables, approximately 50% of the participants perceived themselves as “average.” A higher proportion of participants reported evaluations of “good” and “very good” compared to those reporting “poor” and “very poor.” Moreover, a greater proportion of urban participants reported evaluations of “very good” (as shown in Supplementary Table S2).

Figure 2 presents the pairwise comparison of the associations between the number of guidelines met and self-reported PF indicators. Meeting more guidelines was significantly (p < 0.001) associated with better PF indicators. Additionally, a consistent dose-effect pattern was observed, with more guidelines met corresponding to higher B values.

Figure 3 presents the pairwise comparison of the associations between the specific combinations of guidelines met and self-reported PF indicators. Compared to meeting none of the guidelines, any combination of guidelines met was significantly (p < 0.05) associated with improved PF indicators except for muscular strength in terms of “Screen only vs. None.” In the pairwise comparison of “Sleep only vs. Screen only,” “MVPA + screen vs. MVPA only,” “MVPA + sleep vs. MVPA + screen,” “MVPA + sleep vs. MVPA only,” “Screen + sleep vs. Screen only,” “All vs. MVPA + sleep,” and “All vs. MVPA + screen,” no significant (p > 0.05) enhancement was found with any PF indicators. In addition, the combinations containing MVPA guideline were significantly (B > 0.22, p < 0.05) associated with the majority of PF indicators enhancement compared to those combinations without MVPA guideline. Finally, by comparing the pairwise combinations of “Sleep only VS Screen only,” “MVPA + screen vs. MVPA only,” “MVPA + sleep vs. MVPA only,” “MVPA + sleep vs. MVPA + screen,” “Screen + sleep vs. Screen only,” “Screen + sleep vs. Sleep only,” “All vs. MVPA + screen,” “All vs. MVPA + sleep,” and “All vs. MVPA + screen,” it is found that meeting the screen guideline was significantly (p < 0.05) associated with more PF indicators.

Table 2 presents the associations between the number of guidelines met and self-reported PF indicators, compared to meeting none of the guidelines, stratified by ethnicity and residence. The data show that meeting any number of guidelines was significantly associated with enhanced PF indicators in both Han and Zhuang adolescents (p < 0.01). Additionally, meeting any number of guidelines was significantly associated with enhanced PF indicators in rural adolescents (p < 0.001). Meeting all three guidelines was significantly associated with enhanced PF indicators in adolescents from any residence (B > 0.44, p < 0.05). Meeting only one guideline was not significantly associated with enhanced PF indicators in both suburban and urban adolescents (p > 0.05). Meeting two guidelines was not significantly associated with enhanced flexibility in urban adolescents (p > 0.05).

Table 3 presents the associations between specific combinations of guidelines met and self-reported PF indicators, compared to meeting none of the guidelines, stratified by ethnicity and residence. The data show that among Han adolescents, meeting the screen time guideline alone or in combination with MVPA or sleep guidelines showed no significant association with the majority of PF indicators (p > 0.05 for most comparisons). Meeting the MVPA guideline alone was not significantly associated with general PF or flexibility (p > 0.05). Among Zhuang adolescents, no significant (p > 0.05) correlation exists between meeting the screen time or “MVPA + screen” guidelines and the muscular strength. In all Han and Zhuang participants, all other combinations are significantly (p < 0.001) correlated with all better PF indicators. In addition, any specific combinations of guidelines met was significantly (p < 0.001) associated with enhanced PF indicators in rural adolescents, except for muscular strength, which showed no significant association with adherence to the screen time only or “MVPA + screen” guidelines. In urban adolescents, meeting “Screen only,” “Sleep only,” and “Screen + sleep” combinations of guidelines were not significantly (p > 0.05) associated with all PF indicators.

In this sample, only 1.6% of adolescents met the overall guidelines. This result was similar to the findings of two other studies on Chinese children and adolescents (29, 46), but significantly lower than those reported in studies from other countries (31, 32, 59, 60). A nationwide survey in China reported that 12.4% of adolescents aged 13–22 fully adhered to all guidelines, whereas adherence among those aged 13–17 was markedly lower (3.7%). This disparity may be associated with intense academic pressures in primary and secondary education settings and may also reflect younger adolescents’ limited self-regulatory capacity (30). Because the present data were collected exclusively in Guangxi, the even lower adherence observed here might additionally reflect the region’s relatively underdeveloped economy and lower health literacy (40).

In terms of meeting individual movement guidelines, 2.4, 10.1, and 37.9% of participants met the MVPA guideline, the screen time guideline, and the sleep duration guideline, respectively. Only the prevalence of meeting the sleep duration guideline was higher than the world average (17.7%), while the proportions of meeting the MVPA and screen time guidelines were both lower than the world average (7.7% for MVPA and 19.5% for screen time). More details are shown in Supplementary Table S3. These findings suggest that the low overall adherence is driven primarily by non-compliance with MVPA and screen-time recommendations.

Taken together, the results highlight a critical public health concern that warrants immediate attention and targeted interventions. Future efforts should therefore focus on enhancing guideline-adherence awareness among younger adolescents and developing programs specifically targeting MVPA and screen time behaviors.

The results of this study show that adhering to any number of the 24-h movement guidelines is more likely to lead to a higher level of PF, which suggests that sufficient physical activity, limited screen time, and appropriate sleep are related to the improvement of PF. However, previous studies (29, 31, 32) have indicated that meeting only one guideline, the screen time guideline, or the sleep duration guideline is not associated with PF improvement, which contrasts with the findings of our study. This discrepancy may be attributed to differences in experimental design or study population characteristics. Notably, although adherence to a single guideline was associated with higher PF, the effect size was small (B < 0.162), suggesting the association may be statistically detectable but of limited practical importance. Furthermore, this study reveals that meeting a greater number of the 24-h movement guidelines is increasingly and positively correlated with better PF indicators (i.e., 3 > 2 > 1 > 0). Meeting all three guidelines was associated with a PF gain of at least 0.44 points, exceeding the 0.405-point threshold for a moderate and meaningful improvement. These findings underscore the importance of adhering to all three 24-h movement guidelines (20) and provide empirical support for their holistic application in public health strategies.

In addition, all the combinations containing the MVPA guideline were significantly associated with PF indicators enhancement compared to those without the MVPA guideline (B > 0.22). This directly illustrates the leading role of meeting MVPA guideline in improving PF, which is consistent with the findings of Chen et al. (29), Cai et al. (30) and Tapia-Serrano et al. (31). Previous systematic reviews reported that there was strong, consistent evidence that MVPA was favorably associated with PF (20, 61). Specifically, MVPA enhances cardiorespiratory fitness via cardiorespiratory adaptation (62, 63), boosts muscular fitness (64, 65), and optimizes body composition (66, 67). These synergistic mechanisms collectively drive the significant association between MVPA and PF (20, 61). Moreover, increasing MVPA is likely to displace sedentary behaviors (including screen time) and improve sleep, thereby generating a synergistic effect that further enhances PF (30, 47, 68).

Interestingly, adherence to either the screen time or sleep duration guideline alone was significantly associated with improved PF. However, these associations tended to become non-significant or attenuated in pairwise comparisons of specific guideline combinations—a phenomenon that may reflect statistical interactions or mutual adjustment when other behaviors (e.g., MVPA) are concurrently considered. Notably, the associations with screen time adherence appeared more consistent across PF indicators than those with sleep duration adherence. Yet, the effect sizes for both guidelines were small (B < 0.18), suggesting limited practical relevance. Currently, controversies surround the relationships between screen time, sleep duration, and PF, respectively (22, 29, 30, 32). Some studies have reported a negative correlation between screen time and PF (45, 69, 70), whereas others have found a positive correlation between sleep duration and specific PF indicators (e.g., cardiorespiratory fitness, muscular fitness, and flexibility) (71). In detail, screen time (a major sedentary behavior) reduces the duration of MVPA, which may in turn impair PF (66). In contrast, sleep enhances PF by promoting nutrient absorption, regulating hormonal balance, modulating neural activity, and facilitating fatigue recovery—benefits that are especially prominent in children (20, 28).

Collectively, the data from the present study suggest that the effects of adhering to the 24-h movement guidelines on PF primarily depend on MVPA, followed by screen time and sleep duration. However, this is the first study of its kind to conduct pairwise comparisons among different combinations of meeting the 24-h movement guidelines. It is expected that more studies will be carried out in the future to clarify the relationship between meeting the 24-h movement guidelines and PF.

Minimal differences in the association between the 24-h movement guidelines and PF were observed across ethnic groups. This may be because all participants lived in the same geographic region. Despite differing ethnic backgrounds, shared environmental factors likely led to similar lifestyle patterns.

However, adherence to the screen time, “MVPA + screen” or “Screen + sleep” guidelines was not associated with improvements in most PF indicators among Han adolescents, a pattern not seen in Zhuang adolescents. Several explanations may account for this disparity. First, this may indicate differences in screen behavior patterns between the two groups. Among parents of Han adolescents in China, expectations for their children’s academic achievement are relatively higher (72, 73), and this parental expectation may contribute to increased screen-based learning behaviors and longer weekend screen time among Han adolescents (20, 74). Such differences could weaken the beneficial effects of adhering to screen guidelines on PF, primarily by elevating adolescents’ academic pressure and reducing their engagement in PA (20, 75). Second, given the relatively small effect size (B < 0.19) of screen behaviors on PF, this association is highly susceptible to interference from other confounding factors. For instance, although both groups reside in the same broader region, Zhuang adolescents were more likely to live in rural areas and experience lower socioeconomic conditions (42). In these settings, lower parental education and limited household resources may restrict digital access and structured academic activities (72, 73), while traditional customs, such as participation in festivals, folk singing, and communal dancing, are more likely to be preserved (43). These communities also tend to maintain a distinct dietary culture (41, 42). In contrast, Han adolescents may face higher academic demands and calorie intake due to urban residence and higher socioeconomic status (76, 77), which could indirectly suppress PF despite guideline adherence. While these socioeconomic gradients were not measured in our study, they may underlie the observed differences in the association between movement behaviors and PF, suggesting that the findings reflect not just cultural variation, but also social context.

In summary, although unmeasured socioeconomic factors may contribute to contextual differences, our findings indicate that adherence to screen time, “MVPA + screen” or “Screen + sleep” guidelines does not confer the same PF benefits among Han adolescents as observed in their Zhuang peers. This disparity likely reflects differences in behavioral context, such as socioeconomic status, parental education, family structure, academic pressure, and cultural practices, and highlights the need for integrated, context-sensitive approaches that move beyond isolated movement behavior recommendations.

Another finding of this study is that significant differences exist in the relationship between meeting the 24-h movement guidelines and multiple PF indicators across residence groups. Among rural adolescents, adherence to any combination of guidelines was significantly associated with higher PF level. In contrast, among urban adolescents, no significant association was observed between adherence to screen time guideline, sleep guideline, or their combination and improvements in the majority of PF indicators. First, a large body of research has demonstrated the existence of urban–rural disparities in adolescents’ screen behaviors and sleep quality (71, 74). Urban adolescents often face higher academic expectations, which may contribute to increased screen-based learning activities, sedentary behaviors, and poorer sleep quality (71, 78, 79). Therefore, it is reasonable to hypothesize that the beneficial effects of adhering to screen time and sleep guidelines on PF may be weakened among urban adolescents—a phenomenon potentially driven by their elevated academic stress and reduced levels of PA (66, 71). Furthermore, urban adolescents in China generally have higher socioeconomic status than their rural peers, which contributes to more diverse lifestyles and potentially greater exposure to PF-related risk factors (34). For instance, urban adolescents tend to have a higher intake of calorie-dense foods (76), engage in less active commuting (e.g., walking or cycling) (34), and experience greater academic pressure (33, 51). Collectively, the influence of these factors may weaken the association between adherence to screen time and sleep guidelines and PF among urban adolescents.

However, due to the lack of data on these potential confounders, we are unable to investigate the underlying mechanisms. Therefore, future research is needed to examine how screen behavior patterns and sleep quality relate to PF in this population. It should also explore how these associations vary by socioeconomic status, with the aim of informing equitable, evidence-based interventions aligned with the 24-h movement guidelines. Furthermore, although the associations of adherence to screen time, sleep duration, and their combination with PF differed significantly between urban and rural areas, the effect size differences were relatively small (B < 0.19), suggesting limited practical significance.

This study has some inherent limitations. First, the cross-sectional research design cannot infer causal relationships between variables. Longitudinal tracking or experimental studies are required to determine the directionality of the association between meeting the 24-h movement guidelines and PF. Second, self-reported data may be affected by subjects’ recall bias and social desirability. Objectively measured devices can avoid such biases. Third, the use of convenience sampling may limit the representativeness of the sample, as it may not fully reflect the characteristics of Han and Zhuang adolescents. Future studies should employ more rigorous sampling methods (e.g., stratified random sampling) and larger sample sizes to enhance the generalizability of the findings. Finally, some important confounding factors (such as dietary habits, socioeconomic status, parental education, peer support, family structure and built environment) were not included in the investigation, which may limit the interpretation of the research results. Nevertheless, this study also has several strengths that need to be emphasized. First, it is the first large-sample exploration to focus on Han and Zhuang adolescents, expanding the scope of the population for understanding adherence to the 24-h movement guidelines. Second, it is the first pairwise comparison of the number and combination of adhering to the 24-h movement guideline. The relationships between different numbers and combinations regarding the 24-h movement guidelines and PF are more intuitive and reliable. Third, it is the first exploration of the impact of ethnicity and place of residence on the relationship between adhering to the 24-h movement guidelines and PF, providing more valuable clues for the promotion and optimization of the 24-h movement guidelines.