Adolescence is a pivotal stage in terms of psychological, physiological, and social development, during which social experiences convey long-lasting impacts on an individual's growth (Kamis and Copeland, 2020). Adolescent social anxiety is a concern. It is defined as the persistent fear of social interactions due to severe evaluation apprehension (American Psychiatric Association, 2013). The level of social anxiety in Chinese adolescents had increased during the last two decades, plausibly due to societal changes such as loneliness, family support and complex interpersonal environments (Xin et al., 2022). Notably, intensive social anxiety may lead to social anxiety disorder (SAD) (Asher et al., 2017; Goodman et al., 2021), which has prevalence of 6.5 % among children/adolescents aged from 6 to 18 years, according to a meta-analysis (Polanczyk et al., 2015); onset of SAD often occurs during adolescence (Rapee et al., 2020). Nevertheless, Individuals may feel socially anxious and encounter difficulties in social settings without reaching the level of diagnostic disorder. Even in the non-clinical general adolescent population, a high level of social anxiety has profound effects on an individual's social interactions and interpersonal relationships (Morrison and Heimberg, 2013), educational achievement (Mou et al., 2022), career choice and development (Himle et al., 2020), internet gaming disorder (Marino et al., 2020), mental distress (Kraft et al., 2021), and life satisfaction (Dryman et al., 2016). A higher level of social anxiety in adolescents also increases the risks of depression and suicide in adulthood (Sareen et al., 2005; Woodward and Fergusson, 2001). Despite the significance, adolescent social anxiety has been an under-emphasized area of research compared to research on adolescent depressive symptoms (Lynch et al., 2021; Zhang et al., 2023b). Like other studies (Van Roy et al., 2009), the present study focused on the general adolescent population and investigated social anxiety in a dimensional instead of diagnostic manner.

Sex disparity in mental health is an important research topic. In general, the prevalence of mental distress (e.g., depression, anxiety, substance use and somatoform disorder) is higher among females than males (Andreasen et al., 2014; Bekker and van Mens-Verhulst, 2007). Such is also true regarding the level of social anxiety and prevalence of SAD (Barnett et al., 2021; Tang et al., 2022). For instance, the prevalence of SAD was 11.2 % versus 6.2 % among female and male adolescents in the U.S., respectively (Kessler et al., 2012). The sex difference in social anxiety may sustain the social inequity in mental health, as socially anxious female adolescents were at higher odds of having rumination (Jose et al., 2012), depression (Stein et al., 2001), self-injury, and suicidal ideation than male adolescents (Asher et al., 2017; Long et al., 2024); social anxiety may also lower academic and career achievements (Fang et al., 2025; Strahan, 2003). Thus, to reduce the sex disparity in mental health, it is important to explain the sex difference in the level of social anxiety. In literature, there is a dearth of such studies. One of the very few cross-sectional studies of the type found that masculinity mediated between sex and the level of social anxiety (Zentner et al., 2023). However, causal mediation relationships cannot be ascertained in such a cross-sectional study. To our knowledge, there are no longitudinal studies investigating mediations between sex and level of social anxiety. The present study partially filled out this knowledge gap.

This study firstly tested perceived stress as a potential mediator between sex and social anxiety. The gender role theory postulates that socially constructed gender roles would define specific behaviors and characteristics as being appropriate or inappropriate for males and females (Shields, 2008). Such gender roles prescribe higher expectations on females than males, in terms of emotional sensitivity (Fischer et al., 2018), fulfillments of family and social roles (Ji, 2023), academic performance (Anniko et al., 2019), and attention to appearance and body image (Rawat et al., 2015); such experiences and attributes may generate stronger stress among females than males (Cazassa et al., 2020; Slavich and Shields, 2018). These sociocultural expectations may also increase females' sensitivity and complexity toward stress perception, particularly in the areas of interpersonal relationships (Hampel and Petermann, 2006; Hankin et al., 2007). Additionally, the gender role theory implies that females have stronger social anxiety than males, as it postulates that females would have higher levels of emotional investments on interpersonal interactions (Eagly and Wood, 2016), concerns about quality of relationships (Umberson et al., 1996), and responsiveness to others' emotional needs (Bodenmann et al., 2015) than males, implying higher levels of perceived stress and social anxiety in female adolescents than male adolescents. Furthermore, a study indicated that perceived stress intensifies social anxiety via heightened negative emotions and lowered self-esteem (Farmer and Kashdan, 2015). It is thus plausible that perceived stress may mediate between sex and social anxiety among adolescents. Empirically, female adolescents were more likely than male adolescents to experience high levels of perceived stress (Piekarska, 2020; Thorsén et al., 2020), including but not limited to school-related stress (Kaczmarek and Trambacz-Oleszak, 2021).

Resilience is another potential mediator between sex and social anxiety. The gender role theory also postulates that specific norms and gender expectations may result in a sex difference in resilience (Eagly and Wood, 2016; Vera Gil, 2024). For instance, male adolescents are expected and trained to develop problem solving abilities and employ rational active coping strategies (Wilhsson et al., 2017) whereas females are socialized to express emotions and sensitivity and employ emotionally oriented coping strategies (Ptacek et al., 1992). According to the resilience theory, problem-focused coping strategies were positively associated with resilience whereas emotion-focused coping strategies were negatively associated with resilience (Labouvie, 1987; Lee et al., 2017). Hence, theoretically, males would be more than females to develop resilience. Empirically, a meta-analysis study also reported significantly higher resilience among males than among females (Gök and Koğar, 2021). Across countries, resilience was negatively associated with social anxiety (Jefferies et al., 2021; Lee et al., 2020; Pompili et al., 2024). Furthermore, coping strategies may affect resilience (Shing et al., 2016; Wood and Bhatnagar, 2015), which may in turn influence perceived stress (Ungar, 2011). Hence, as perceived stress was both negatively associated with resilience (Chandelkar and Shetty, 2019) and positively associated with social anxiety (Feng et al., 2019; Wang et al., 2022), it may mediate between resilience and social anxiety. A serial mediation was tested in this study.

Given the background, the present 3-wave longitudinal study investigated the associations between sex and the levels of perceived stress and resilience between male and female adolescents in China. Longitudinal associations between perceived stress/resilience and social anxiety were also investigated. In addition, mediation hypotheses between sex and social anxiety at Time 3 (T3) were tested. They included 1) H1: sex → resilience at Time 1 (T1) → social anxiety at Time 3 (T3), 2) H2; sex → perceived stress at Time 2 (T2) → social anxiety at Time 3 (T3), and 3) H3 (a serial mediation): sex → resilience at Time 1 (T1) → perceived stress at Time 2 (T2) → social anxiety at Time 3 (T3). Specifically, it is hypothesized that (1) females, compared to males, would have higher levels of perceive stress at T2 and social anxiety at T3; they would also have lower levels of resilience at T1 than males; (2) resilience at T1 would be negatively associated with perceived stress at T2 and social anxiety at T3; (3) perceived stress at T2 would be positively associated with social anxiety at T3.