While empirical research demonstrates that social media use positively correlates with academic performance when deeply integrated into instructional processes as a medium for knowledge construction and teacher-student interaction (Altalhi, 2020), additional evidence supports this relationship. Alshalawi’s (2022) quantitative study of Saudi undergraduate students further confirmed a positive association between social media usage frequency and academic performance, highlighting its potential value as a digital learning tool. However, the effects of social media use are context-dependent. When students multitask between social media and academic activities, cognitive resource conflicts produce negative outcomes (Malik et al., 2020; Shafiq and Parveen, 2023). Substantial research links high-intensity social media use with poor academic outcomes (Whelan et al., 2020; Evers et al., 2020) and shows that excessive use can impair attention and increase academic distraction (Boer et al., 2020). Experimental studies demonstrate that social media multitasking directly reduces test performance by increasing cognitive load (Lau, 2017). Meta-analytic evidence confirms these detrimental effects are robust across contexts (Appel et al., 2020). Based on this evidence, we hypothesize:

Social anxiety shows complex relationships with digital media use. The American Psychiatric Association defines this condition through two key dimensions: persistent fear or anxiety in social situations and excessive worry about negative evaluation by others (Lolk, 2013). Clinical research characterizes social anxiety as an anxiety disorder marked by anticipatory embarrassment and shame (Rasouli et al., 2022).

Problematic social media use disrupts both learning efficiency and real-world relationships while worsening anxiety symptoms by weakening offline social support (Alkis et al., 2017). Digital natives are particularly vulnerable to compulsive online behaviors that trigger emotional disorders including social anxiety and depression (Salmela-Aro et al., 2017). Social anxiety undermines university students’ psychosocial adaptation and relationship quality while predicting poor academic performance (Brook and Willoughby, 2015; Zukerman et al., 2019; Mou et al., 2024). Socially anxious individuals often use self-isolation and avoidance to manage their anxiety (Russell and Topham, 2012). These avoidance behaviors—withdrawing from classroom participation, avoiding questions, and reluctance to engage in discussions—produce maladaptive outcomes and absenteeism that ultimately harm academic performance (Mou et al., 2024). We therefore propose:

Fear of missing out (FoMO) occurs when individuals worry about missing relevant information or desired content when social media and smartphones are inaccessible (Casale et al., 2018). Research shows that problematic social media behaviors significantly influence adolescents’ FoMO levels. Studies find strong links between problematic social media use and FoMO, with more severe usage patterns associated with greater FoMO symptoms (Primack et al., 2017).

Student-focused research indicates that FoMO negatively affects academic performances (Tanrikulu and Mouratidis, 2023; Parizad et al., 2022; Kong et al., 2024). This relationship operates through specific cognitive mechanisms: when students lack effective strategies for managing classroom smartphone use and experience impaired decision-making, their learning focus deteriorates and academic performance suffers (Rosen et al., 2018). We therefore hypothesize:

Despite promises to “give people the power to build community and bring the world closer together,” social media often disrupts daily activities and relationships. Social media use links to negative mental health outcomes (Vannucci et al., 2017), with FoMO and social anxiety as key psychological mechanisms (Scheinfeld and Voorhees, 2022; Elhai et al., 2016).

FoMO amplifies uncertainty and anxiety in virtual social contexts (Oliveira et al., 2018; Zhu et al., 2024). These negative emotions drive pathological online compensation behaviors—frequent social media refreshing and excessive message checking—creating compulsive dependence on virtual social activities (Wang et al., 2021). This creates a vicious cycle where cognitive resources and emotional energy are depleted, leading to higher online social anxiety levels (Wu et al., 2025).

Research consistently shows that higher social anxiety predicts lower academic performance over time (Kajastus et al., 2024; Ashshawareb et al., 2024; Abdulsalim et al., 2025). Socially anxious students struggle to integrate into academic environments because they fear not being valued in interpersonal relationships. These fears inhibit classroom participation and help-seeking from teaching assistants, professors, and other college staff (Brook and Willoughby, 2015). Therefore, we propose:

University environments systematically influence young adults’ cognitive development, psychological adjustment, and social adaptation (Thornberg and Charmaz, 2022). Teacher-student relationships form core interpersonal bonds in educational settings, establishing psychological connections through cognitive interaction, emotional bonding, and behavioral feedback (Wang et al., 2023). Attachment theory shows that early interactions with significant others create internalized working models that guide behavioral responses across contexts (Wang et al., 2023).

Research demonstrates that teacher-student relationship quality significantly affects students’ classroom engagement, academic performance, and intrinsic motivation (Pianta et al., 2012). Positive relationships enhance learning engagement and performance through emotional support and behavioral modeling while buffering against negative emotions like anxiety and depression. Teacher behaviors shape students’ development in cognitive, emotional, and behavioral domains through protective factors including mutual respect, trust, and emotional warmth (Veldman et al., 2013). Poor teacher-student relationships weaken learning motivation and academic self-efficacy, creating sustained declines in learning outcomes (Wang et al., 2024). We hypothesize:

We used stratified cluster sampling to ensure representative national coverage. From March 15 to April 1, 2025, we selected 120 cities across China using random sampling principles. The selection included provincial capitals across 23 provinces and 5 autonomous regions, as well as four municipalities directly under central government administration. Additionally, 2–6 prefecture-level cities were randomly selected from non-capital administrative regions within each province and autonomous region. To achieve optimal demographic distribution, quota sampling was implemented for participant recruitment. Research assistants were responsible for questionnaire distribution and collection among designated target populations, resulting in 3,775 distributed questionnaires. To ensure data quality, incomplete responses, surveys with excessively short completion times, and those containing extreme or anomalous values were systematically excluded. Following rigorous data cleaning procedures, 3,716 valid responses were retained, yielding an effective response rate of 98.47%.

The research assistants (RAs) recruited for this study were required to have at least a second-year undergraduate education in relevant fields such as psychology, communication, or education. RAs were selected based on their familiarity with the survey process, communication skills, and ability to effectively manage participant interactions. Prior to data collection, all RAs underwent a comprehensive training program that covered ethical research practices, appropriate survey management techniques, and quality control protocols. Data collection involved recruiting at least one RA or survey team in each sample city. Individual RAs were responsible for collecting 30 to 90 responses, while survey teams managed between 100 and 200 questionnaires. The survey was administered through the online platform Wenjuanxing¹ via one-on-one, face-to-face interactions. Participants accessed the surveys through the provided links, with RAs ensuring informed consent was obtained and assigning unique identification numbers to each response. In cases where participants had cognitive capacity but limited mobility, RAs conducted verbal, one-on-one interviews and recorded the responses on their behalf. Following data collection, two independent reviewers performed logical checks and data screening to ensure the accuracy and integrity of the data.

The final sample comprised 1,842 males (49.6%) and 1,874 females (50.4%), with 1,939 urban residents (52.2%) and 1,777 rural residents (47.8%). Age distribution included 655 eighteen-year-olds (17.6%), 746 nineteen-year-olds (20.0%), 751 twenty-year-olds (20.2%), 732 twenty-one-year-olds (19.8%), 739 twenty-two-year-olds (19.9%), and 93 participants over twenty-two years (2.5%), as detailed in Table 1.

Inclusion criteria: (1) aged 18–24 years; (2) Chinese nationality; (3) permanent Chinese residents (annual time abroad ≤1 month); (4) voluntary participation with informed consent; (5) ability to complete online surveys independently or with research assistant support; (6) comprehension of questionnaire item meanings.

Exclusion criteria: (1) impaired consciousness or psychiatric disorders; (2) concurrent participation in similar research studies; (3) unwillingness to cooperate.

The research team conducted comprehensive literature review to identify existing questionnaire designs and clarify research questions and relevant variables based on study objectives. Initial questionnaire drafts were developed with systematic item organization and categorization. Pilot testing was subsequently conducted to examine questionnaire validity and item appropriateness, with iterative refinements based on testing results culminating in the final instrument. The distributed questionnaire comprised two primary sections: demographic variables to assess participants’ basic characteristics, including gender, current educational level, household registration type, and age, specifically tailored for adolescent populations; and measurement scales for study variables addressing research-specific content.

According to the study’s theoretical framework, five primary variables were examined: social media use, teacher-student relationships, social anxiety, fear of missing out, and academic performance. To ensure adequate reliability and validity, items were primarily adapted from previously validated and well-established scales. Additionally, demographic questionnaires were included to examine variations across gender and geographic regions.

This study employed Harman’s single-factor test to systematically examine potential common method bias. Statistical analysis revealed that without factor rotation, six common factors with eigenvalues greater than 1 were extracted, with the first factor accounting for 32.364% of the variance. This value falls substantially below the critical threshold of 40%, indicating that the research data does not exhibit significant systematic common method bias and that the measurements demonstrate satisfactory discriminant validity. These findings provide methodological validity assurance for subsequent structural equation modeling analyses and ensure the reliability of inferences regarding inter-variable relationships.

This study utilized AMOS 26.0 statistical software to conduct confirmatory factor analysis on five core latent variables, with model fit indices detailed in Table 2. The analysis demonstrated that the five-factor model achieved a chi-square to degrees of freedom ratio (χ²/df) of 2.473 (below the critical value of 3), a root mean square error of approximation (RMSEA) of 0.034 (below the threshold of 0.08), and comparative fit index (CFI), Tucker-Lewis index (TLI), and goodness of fit index (GFI) values all exceeding 0.90. These fit indices not only fully satisfy statistical requirements but also demonstrate superior fit compared to alternative competing models, indicating robust discriminant validity in this study.

The means, standard deviations, and correlation coefficients for all variables are presented in Table 3. Results indicated that social media usage intensity demonstrated a significant positive correlation with fear of missing out (r = 0.172, p < 0.05), as well as significant positive associations with social anxiety (r = 0.165, p < 0.01) and teacher-student relationship quality (r = 0.156, p < 0.01), while exhibiting a significant negative correlation with academic performance (r = −0.131, p < 0.01). Regarding mediating variables, fear of missing out showed a moderate positive correlation with social anxiety (r = 0.173, p < 0.01), a significant negative association with academic performance (r = −0.361, p < 0.01), and a significant negative correlation with teacher-student relationship quality (r = −0.138, p < 0.05). Social anxiety demonstrated negative associations with both academic performance (r = −0.249, p < 0.01) and teacher-student relationship quality (r = −0.256, p < 0.01), while academic performance exhibited a significant positive correlation with teacher-student relationship quality (r = 0.234, p < 0.01). These correlation patterns preliminarily reveal complex pathways among variables, providing empirical foundations for subsequent structural equation modeling.

Based on the conceptual model, this study constructed a structural equation model incorporating social media usage, social anxiety, fear of missing out, and academic performance as core variables. In testing the hypothesized path relationships, social media usage demonstrated a significant positive predictive effect on fear of missing out (β = 0.212, p < 0.001) and social anxiety (β = 0.160, p < 0.001), while exhibiting a significant negative predictive relationship with academic performance (β = −0.124, p < 0.001), thereby supporting the first research hypothesis. Further analysis revealed that social anxiety exerted a significant positive effect on fear of missing out (β = 0.395, p < 0.001) and a significant negative predictive effect on academic performance (β = −0.143, p < 0.001), while fear of missing out similarly demonstrated a significant negative predictive relationship with academic performance (β = −0.268, p < 0.001). All hypothesized paths in the model achieved statistical significance, with path directions consistent with research expectations. Path coefficients and significance levels are detailed in Table 4 and Figure 1.

Based on the research’s conceptual model, this study hypothesized that social anxiety and fear of missing out would mediate the relationship between adolescent social media usage and academic performance, with analysis and verification conducted using AMOS 28.0. The bias-corrected nonparametric percentile bootstrap method was employed to examine the mediation effects of social anxiety and fear of missing out. Table 5 presents standardized estimates for each indirect path and 95% confidence intervals for mediation effects, with mediation effects considered significant when 95% confidence intervals exclude zero. According to Table 5, the 95% confidence intervals for all three indirect paths exclude zero, indicating significant effects for social anxiety mediation, fear of missing out mediation, and serial mediation. Specifically, the mediation effect value for social anxiety was 0.0075, for fear of missing out was 0.0165, and for serial mediation was 0.0041. Therefore, Hypotheses 2, 3, and 4 are all supported.

The PROCESS macro was utilized to examine the moderating role of teacher-student relationships in the association between fear of missing out and academic performance. Results indicated that the interaction term between teacher-student relationships and fear of missing out exhibited a significant negative predictive effect on academic performance (β = −0.175, p < 0.01), demonstrating that teacher-student relationships negatively moderate the relationship between fear of missing out and academic performance. Therefore, Hypothesis H5 was supported. To provide a more intuitive representation of the moderating effect of teacher-student relationships on the fear of missing out-academic performance relationship, this study followed recommendations by Aiken et al. (1991) by plotting relationships at one standard deviation above and below the mean to illustrate differential effects of fear of missing out on academic performance across varying levels of teacher-student relationship quality, as shown in Figure 2.

The current study demonstrates that social media use significantly and negatively impacts college students’ academic performance. Higher social media usage is associated with a reduced likelihood of achieving favorable academic outcomes, which aligns with findings from previous studies (Evers et al., 2020; Sampasa-Kanyinga et al., 2019; Giunchiglia et al., 2018). This negative relationship can be examined through various theoretical lenses. From a behavioral media use perspective, social media platforms are mobile and operate in real-time, allowing students to access them almost anywhere, including during class. This “always-on” nature of social media can undermine sustained attention and the cognitive processing necessary for effective learning (Al-Menayes, 2014).

From the media psychology perspective, constant information streams create patterns of “dynamic updating, immediate response, and constant switching,” leading to fragmented attention and cognitive overload (Shen, 2019). As a result, students’ working memory is occupied, making it difficult for them to focus deeply on course material. Furthermore, frequent late-night use of social media is closely linked to sleep disorders and academic burnout (Evers et al., 2020). According to the emotional exhaustion model, chronic sleep deprivation and fatigue drain students’ motivation and emotional regulation, ultimately leading to decreased academic engagement and poorer grades.

The study also finds that social anxiety and FoMO play significant mediating roles in the relationship between social media use and academic performance. According to cognitive-behavioral theory, individuals with social anxiety tend to have negative automatic thoughts about social interactions (Clark and Wells, 1995; Weidman and Levinson, 2015), making them more likely to turn to social media as a way to avoid real-world social pressures (Lee and Stapinski, 2012). However, this form of social participation does not alleviate social anxiety. Instead, it consumes cognitive resources, weakening students’ attention and executive control during learning tasks (Junco and Cotten, 2012; Oberst et al., 2017; Albulescu et al., 2024). Additionally, socially anxious individuals often engage in upward social comparisons on social media, leading to increased feelings of inferiority and negative self-worth, which further detracts from academic motivation and performance (Kajastus et al., 2024; Vogel et al., 2014; Jehi et al., 2024).

Further research indicates that excessive social media use intensifies anxiety and stress levels among college students, diminishing the positive impact of self-control on academic performance (Vogel et al., 2014). Thus, social anxiety emerges as a crucial mediating variable, influencing both the emotional and cognitive aspects of students’ academic performance.

FoMO, which is significantly linked to social media use (Alutaybi et al., 2020), exacerbates this problem. Students with high FoMO levels tend to worry about missing out on their peers’ experiences, which disrupts their focus and academic engagement (Chai et al., 2018). This, in turn, leads to poor academic performance and negative student outcomes (Hairul et al., 2025). This study is the first to validate this mediating pathway within the Chinese context.

Theoretical frameworks like Self-Determination Theory (SDT) and Social Comparison Theory shed light on the mechanisms underlying these relationships. SDT suggests that social media environments often fail to satisfy students’ needs for autonomy, competence, and relatedness, prompting emotional compensation through excessive media use. This, in turn, diverts time from academic activities (Ryan and Deci, 2000; Deci and Ryan, 2008). Social Comparison Theory explains that upward comparisons on social media reinforce negative self-perceptions, amplifying FoMO’s detrimental effects on academic motivation (Vogel et al., 2014). Together, social anxiety and FoMO exert a negative influence on academic performance by depleting cognitive resources and reducing motivation.

The chain mediation analysis reveals that social media use influences academic performance through a series of mediating factors: social media use leads to social anxiety, which then leads to FoMO, and ultimately harms academic outcomes. This finding supports previous research (Dempsey et al., 2019; Wolniewicz et al., 2018) and extends the “emotional activation—cognitive resource displacement” pathway outlined in the I-PACE model (Brand et al., 2016).

Studies have shown that social anxiety directly predicts FoMO, with individuals exhibiting higher levels of social anxiety also showing a stronger concern about missing out on social media content (Dempsey et al., 2019). From a media psychology perspective, social anxiety can be seen as an immediate emotional response to perceived evaluative threats. Students with higher levels of social anxiety often monitor peer feedback closely and avoid negative evaluations, which intensifies their fear of missing out on relevant social information (Dempsey et al., 2019). FoMO then becomes a motivational amplifier, driving students to frequently refresh their social media feeds and switch attention between tasks, which compromises their ability to focus on academic work. While this model offers valuable insights, it may not fully capture the complexity of the emotional and cognitive processes involved, as it assumes a linear progression that may not apply universally across all individuals.

Positive teacher-student relationships (TSR) can be a crucial factor in satisfying students’ “relatedness needs,” as emphasized by Self-Determination Theory. When students feel respected, understood, and supported by their teachers, their intrinsic motivation is strengthened, and their sense of control over academic goals increases. This reduces the compulsive use of social media driven by FoMO and improves executive function and time management skills, reducing attention fragmentation. Empirical evidence shows that in the presence of strong TSR, the negative impact of FoMO on academic performance is significantly reduced (Li et al., 2022; Hornstra et al., 2023).

Positive TSR can be understood through the lens of James Carey’s ritual view of communication (Carey, 1975), where teacher-student interactions are seen as continuous, meaningful, and ritualistic rather than sporadic exchanges of information. Unlike the fragmented interactions common in online spaces, in-person interactions, such as classroom discussions, after-class tutoring, and daily greetings, help satisfy students’ relatedness needs, providing them with emotional security. This stability can counteract the cognitive depletion caused by FoMO, allowing students to better focus on their academic tasks. According to Feng (2022), the moderating effects of TSR unfold in three ways: emotional support, cognitive reappraisal, and resource reallocation.

First, emotional support helps students feel secure and valued, which can reduce the emotional triggers of FoMO. Second, cognitive reappraisal helps students challenge their fears of missing out on information, reducing their anxiety. Third, resource reallocation helps students better allocate their attention and time, thus reducing distractions caused by FoMO and enhancing their ability to engage in focused learning.

In practical terms, while improving TSR may help alleviate some of the cognitive and emotional burdens caused by FoMO, it does not fully counteract the negative effects on academic performance. To put this into perspective, this modest effect can be compared to the role of a supportive friend who helps to mitigate stress but does not eliminate it entirely. In educational settings, this implies that while fostering positive teacher-student relationships can certainly play a role in managing the adverse effects of FoMO, it might need to be combined with other interventions, such as developing self-regulation skills or digital detox practices, to more effectively address the issue. Educational and psychological interventions aimed at strengthening teacher-student relationships should focus on providing emotional support and fostering trust and understanding between students and teachers. However, the modest nature of this effect indicates that these interventions should not be seen as a panacea for all the challenges posed by FoMO. A more holistic approach is needed, where TSR interventions are complemented by strategies that directly address time management, media literacy, and emotional regulation.