Power analysis using G*Power 3.1 determined a minimum sample size of 54 participants (α = 0.05, power = 0.95, d = 0.80) (Faul et al., 2009), with 60 ultimately recruited to accommodate a 15% anticipated attrition rate (Brailovskaia and Margraf, 2020). By using the Short Video Addiction Scale (SVAS) in combination with the usage reports of short video applications on smartphones, a total of 63 college students who were addicted to short videos were recruited as participants from a university in Chengdu. Three students dropped out due to academic reasons before the start of the experiment, and finally, 60 students completed the entire experimental process. There are 7 diagnostic questions in the SVAS. If a participant gives affirmative answers to any 5 of these questions, they can be diagnosed as a severe short video addict, with a daily average usage duration of ≥ 4 h. All participants were randomly divided into an experimental group (17 males, 13 females) and a control group (15 males, 15 females), with an average age of 19.31 ± 1.36 years. This study strictly adheres to the ethical standards of the Declaration of Helsinki. All college students participating in this study have been informed of the research purpose, process, and potential risks, and have voluntarily signed a written informed consent form. This research has been approved by the Ethics Committee of Xihua University(XH250124-01).

A 2 (Group: Experimental Group/Control Group) × 2 (Time: Pre-test/Post-test) mixed experimental design was utilized. The group served as the independent variable, while the SVAS score constituted the dependent variable, and cognitive bias and self-control functioned as the mediating variables. The experimental group was engaged with the table tennis project, whereas the control group, aside from regular physical activities, did not take part in any other forms of physical exercise.

The experimental group undergoes a 16-week table tennis exercise intervention three times a week (Pan et al., 2017; Tsai et al., 2017). Participants are required to, in addition to normal physical exercises, begin practicing a 60-min table tennis session (10-min warm-up + 40-min formal training + 10-min relaxation) at 7:00 p.m. every Wednesday, Friday, and Saturday. Before each experiment, five participants are randomly selected to wear Polar watches to monitor their heart rates, with exercise intensity controlled at a moderate level (60–80% of the maximum heart rate, approximately 110–150 beats per minute) (Sperlich et al., 2011). The single-training module includes dynamic stretching in the warm-up stage, racket-holding swing practice, technical training (30 min) that encompasses basic movement strengthening, footwork training, multi-ball continuous practice, and a physical fitness module (15 min) that consists of special physical fitness cycles, core stability training, and relaxation and recovery through static stretching. Record the average weekly physical activity level of the subjects through the IPAQ-S. Conduct a t-test on the physical activity levels of the two groups before the formal start of the intervention (p = 0.61 > 0.05), demonstrating that there is no significant difference in physical activity between the two groups before the start. After the formal experiment begins, the physical activity level of the exercise group increases significantly [(2025.13 ± 279.18) MET-min/w], while the physical activity level of the control group remains stable [(1055.63 ± 235.96) MET-min/w] (for details, see Figure 2).

This study used SPSS 25.0 and AMOS for statistical analysis, recoded the reverse-scored questions, applied the analysis of variance method, conducted correlation analysis, and collinearity tests on the data. Subsequently, the SPSS add-in PROCESS add-in developed by Hayes (2013) was used to test the mediating model. A significance level of p < 0.05 was set for statistical significance.

Repeated measures analysis of variance (Table 3) showed that there were significant differences in the main effect of time for short video addiction [F_{(1, 58)} = 85.24, p < 0.01, η_p² = 0.60], cognitive bias [F_{(1, 58)} = 92.17, p < 0.01, η_p² = 0.61], and self-control [F_{(1, 58)} = 78.93, p < 0.01, η_p² = 0.58], indicating that after the experiment, each variable showed a significant trend of change over time. In terms of the group effect, short video addiction [F_{(1, 58)} = 21.35, p < 0.01, η_p² = 0.27], cognitive bias [F_{(1, 58)} = 28.74, p < 0.05, η_p² = 0.33], and self-control [F_{(1, 58)} = 22.86, p < 0.01, η_p² = 0.28] in the experimental group were significantly better than those in the control group. In the interaction term of time × group, there were significant differences for short video addiction [F_{(1, 58)} = 64.82, p < 0.05, η_p² = 0.53], cognitive bias [F_{(1, 58)} = 76.53, p < 0.01, η_p² = 0.57], and self-control [F_{(1, 58)} = 61.12, p < 0.05, η_p² = 0.51]. Further simple effect analysis was carried out. In the experimental group, the post-test scores of short video addiction [t₍₂₉₎ = 12.37, p < 0.001], cognitive bias [t₍₂₉₎ = 13.25, p < 0.001], and self-control [t₍₂₉₎ = 11.84, p < 0.001] were significantly better than those of the pre-test. There was no significant difference between the pre-test and post-test in the control group (p > 0.05). In the post-test, the scores of short video addiction [t₍₅₈₎ = 8.96, p < 0.001], cognitive bias [t₍₅₈₎ = 9.24, p < 0.001], and self-control [t₍₅₈₎ = 7.85, p < 0.001] in the experimental group were significantly better than those in the control group (Figure 3).

Correlation analysis revealed significant associations among variables. To mitigate covariance interference, all predictor variables were standardized using z-score transformation before multicollinearity diagnostics. Results indicated tolerance values of 0.86 for table tennis intervention (VIF = 1.16), 0.35 for cognitive bias (VIF = 2.86), and 0.32 for self-control (VIF = 3.12). All tolerance values ranged from 0.32 to 0.86 (exceeding the > 0.1 threshold), and VIF values ranged from 1.16 to 3.12 (below the < 5 cutoff). Supplementary diagnostics confirmed a condition index of 18.7 (below the 30 threshold), minimum eigenvalue of 0.28 (above 0.01), and maximum variance proportion of 0.63 (below 0.90). Consistent with established multicollinearity standards (Cohen et al., 2003; O’brien, 2007), these results confirm the absence of significant multicollinearity, satisfying prerequisites for subsequent mediation analyses (Cohen et al., 2003).

Taking the group (exercise experimental group/control group) as the independent variable, short video addiction (change amount) as the dependent variable, and cognitive bias and self-control (change amount) as the mediating variables, the chained mediating effect test was conducted using the PROCESS plugin (Model 6, Bootstrap = 5,000) developed by Hayes. The results showed that table tennis exercise significantly and negatively predicted cognitive bias (β = –0.50, p < 0.01) and short video addiction (β = –0.25, p < 0.05), and significantly and positively predicted self-control (β = 0.40, p < 0.01); cognitive bias significantly and negatively predicted self-control (β = –0.30, p < 0.05), and significantly and positively predicted short video addiction (β = 0.40, p < 0.01); self-control significantly and negatively predicted short video addiction (β = –0.35, p < 0.01). For the results of regression analysis, please refer to Table 4; for the path model, please refer to Figure 4.

The Bootstrap method was used to resample 5,000 times to calculate the 95% confidence interval. Table tennis can directly affect short video addiction, with a direct effect value of β = –0.25 [95% CI (–0.40, –0.10)]. Cognitive bias and self-control play a partial mediating role between table tennis and short video addiction, with a total mediating effect value of β = –0.41 [95% CI (–0.55, –0.27)], accounting for 68.33% of the total effect. Specifically, the mediating effect consists of the following three paths: Path 1: The indirect effect value is β = –0.20 [95% CI (–0.30, –0.10)], accounting for 33.3% of the total effect, indicating that table tennis indirectly reduces addiction by decreasing cognitive bias. Path 2 : The indirect effect value is β = –0.16 [95% CI (–0.25, –0.06)], accounting for 26.67% of the total effect, indicating that table tennis inhibits addictive behavior by enhancing self-control. Path 3 : The indirect effect value is β = –0.05 [95% CI (–0.10, –0.01)], accounting for 8.33% of the total effect, revealing the continuous mediating mechanism of cognitive bias and self-control (Table 5).

Research shows that after 16 weeks of table tennis intervention, the short video addiction of college students has been significantly improved. Previous studies have indicated that physical exercise plays a positive role in alleviating mobile phone addiction and addictive social media use (Brailovskaia and Margraf, 2020; Zhao et al., 2022). The results of our experiment are consistent with these findings, demonstrating that a structured table tennis program can effectively reduce SVA scores. Specifically, the mechanism may be twofold. First, short video addiction is significantly associated with negative emotions such as depression and loneliness (Niu, 2023), and its essence is a virtual compensation in the absence of real-world social connections. Table tennis, as a team sport, creates high-frequency real-world social scenarios. It not only directly replenishes an individual’s psychological resources but also weakens the psychological dependence on virtual online social interactions by rebuilding real and intimate relationships, thus alleviating the psychological incentives for short video addiction at the root (Yang et al., 2025). Second, compared with traditional aerobic exercises, the skill training in racket sports (such as serving, spin control, and landing point prediction) can continuously stimulate an individual’s sense of challenge and provide immediate feedback. This engaging nature of table tennis may effectively compete for the time and attention that individuals would otherwise devote to short videos, thereby directly reducing usage time and dependency (Qian et al., 2025).

Research shows that after 16 weeks of table tennis intervention, the cognitive biases of college students have been significantly improved. This finding provides empirical support for Davis’s cognitive-behavioral theory model (Davis, 2001). This theory emphasizes the central role of maladaptive cognition in the mechanisms of addictive behavior and psychopathology. The results of this study reveal that table tennis can serve as an effective cognitive restructuring intervention. Longitudinal neuroimaging studies have shown that professional table tennis players are significantly superior to the general college student group in cognitive dimensions such as decision-making, visual attention regulation, and executive function. This is accompanied by an enhancement in the integrity of white matter fiber tracts in the brain and an improvement in neural conduction efficiency (Ji et al., 2024). The synergistic effect of this structural and functional change provides a biological basis for the improvement of cognitive biases (Padulo et al., 2016). Resting-state functional magnetic resonance imaging (fMRI) studies further confirm that table tennis training can optimize the functional connectivity networks in brain regions related to memory encoding, motor control, and visual information processing (Li et al., 2023). This aligns with existing research showing that sports participation can significantly reduce cognitive bias among healthy adults (Qi et al., 2024).

This study confirmed that after 16 weeks of structured table tennis intervention, the self-control ability of college students was significantly improved, and this finding is consistent with the existing research results (Liu et al., 2012). The high-frequency decision-making needs and dynamic attention allocation mechanism in table tennis continuously exercise an individual’s working memory and cognitive flexibility (Elferink-Gemser et al., 2018). When dealing with multi-source information, table tennis players have a faster cognitive switching speed than ordinary people. This improvement in cognitive flexibility helps individuals to carry out more efficient self-regulation when facing temptations (Wang and Qiu, 2025a; Wang Y. et al., 2024). In addition, as a typical team competitive sport, table tennis’s characteristics of competition and cooperation prompt participants to continuously adjust their behavioral strategies to adapt to environmental changes. While enhancing social cognitive abilities, it indirectly strengthens self-control abilities (Yu et al., 2022; Zhou et al., 2023). These findings, together with intervention studies on team sports such as basketball (Haoran et al., 2023; Wei et al., 2024), jointly reveal that dynamic interactive sports can effectively improve individuals’ self-control levels through the dual paths of cognitive training and social adaptation.

This study found that cognitive bias and self-control play a significant chain-mediating role between table tennis exercise and short video addiction, thus supporting our hypothesis H4. The path analysis indicates that physical exercise reduces cognitive bias, which in turn enhances self-control ability, ultimately leading to reduced addiction. From the perspective of the limited resource theory, correcting cognitive bias requires consuming the limited psychological resources dominated by the prefrontal lobe. Continuous table tennis exercise systematically expands the reserve of psychological resources by enhancing the neural plasticity of the prefrontal cortex (Clarke et al., 2018). The results of this study are consistent with the conclusions of previous studies. Twelve weeks of moderate-intensity aerobic exercise significantly reduces the participants’ attentional bias toward addictive stimuli. At the same time, it enhances self-control ability by increasing the delay discount rate, directly verifying the effectiveness of the dual-mediation path (Cui and Yang, 2024; Yang et al., 2022). In addition, some research points out that in the digital media ecosystem, the algorithm recommendation mechanism reinforces addictive behaviors through immediate rewards (Ke et al., 2023; Kim et al., 2024; Wang and Qiu, 2025b). However, the cognitive restructuring triggered by physical exercise can weaken individuals’ neural adaptability to the immediate rewards of short videos. At the same time, the enhancement of self-control ability effectively blocks the positive feedback loop of addictive behaviors through the behavioral inhibition mechanism (Pindus et al., 2025). This dual mechanism reveals the unique value of physical exercise in the prevention and treatment of behavioral addiction in the digital age.

This study offers several important implications. Practically, the findings provide a feasible and effective non-pharmacological intervention for universities to address the growing issue of short video addiction among students. Integrating structured racket sports, such as table tennis, into physical education curricula or extracurricular activities could serve as a preventative and remedial measure to enhance students’ mental wellbeing and self-regulation skills. Moreover, for mental health practitioners, the identified chain mediation pathway (cognitive bias → self-control) suggests that intervention programs could target cognitive restructuring and self-control training simultaneously for better efficacy. Theoretically, this research extends the cognitive-behavioral model of addiction by empirically validating it in the context of short video addiction and introducing physical exercise as a key antecedent. The demonstrated chain mediation effect enriches our understanding of the underlying psychological mechanisms, highlighting a sequential process where exercise first ameliorates maladaptive cognition, which in turn bolsters self-regulatory capacity, ultimately reducing addictive behaviors.

While this study provides evidence for the effectiveness of table tennis intervention, several limitations should be noted when interpreting the results, which also point to directions for future research. Firstly, the generalizability of the findings is limited by the sample, which was drawn from a single university. Future studies should recruit participants from multiple universities across different regions and cultural backgrounds to enhance the external validity of the findings. Secondly, the reliance on self-reported measures for key variables (e.g., short video addiction, self-control) might be influenced by social desirability. Incorporating objective measures, such as behavioral tasks for assessing cognitive bias and self-control, or using smartphone analytics to track actual usage time, could strengthen the validity of future research.

Thirdly, the design lacked an active control group. The observed effects might therefore be attributed to the extra social interaction and attention received by the experimental group, rather than the specific attributes of table tennis. Including an alternative exercise group in future designs would help isolate the unique effect of table tennis.