Social anxiety often reaches its peak during adolescence and may hold a noteworthy impact both at the social and academic levels. Though some adolescents easily learn to adequately cope with it, others become overwhelmed by social fears and end up avoiding social events or facing them with intense suffering. Adolescents with social anxiety disorder (SAD) seldom seek professional help and often become adults with SAD, making it necessary to timely intervene during early stages of the disorder and determine what works better, why, and for whom when trying to change the course of social anxiety in adolescence. Cognitive therapy (CT) has been the treatment of choice for adult SAD, but research within adolescent samples is scarce. More recently, third-wave cognitive approaches, namely compassion-focused therapy (CFT) and acceptance and commitment therapy (ACT), have also proven effective in treating mental disorders, but their efficacy towards adolescents' SAD remains unclear.

This research project intended to evaluate and compare the therapeutic efficacy of CT, CFT, and ACT as applied to adolescents with SAD, in comparison with waiting-listed, after intervention and over a 12-weeks and a 24-weeks follow-up, using a randomized parallel trial approach. This methodology will allow to test if intervened subjects diverge from the non-intervened SAD subjects, and if therapeutic change differs across intervention conditions. Therapeutic change process will be investigated, specifically if efficacy is mediated by change in the core constructs of each theoretical framework. Finally, moderators of change such the initial level of symptomology will be examined, as to determine which intervention works best for whom. The same procedures will be repeated for each intervention condition, namely recruitment, intervention, and assessment.

This research will provide evidence on which form of therapy may be the optimal choice in to intervene in SAD in general, and when dealing with diverse specific vulnerabilities associated with social fears. It will also add to the field of empirically validated therapies, with a specific focus on adolescence. From a societal perspective, the project will assist in empowering schools to contribute to shape how their students act and develop into socially apt adults.

Social anxiety is characterized by emotional reactivity to social events in which individuals believes they can be subject to the scrutiny of others. Although normative during adolescence, along with the growing importance attributed to acceptance and integration with peers, it can come to fulfill criteria for social anxiety disorder (SAD; McNeil, 2010), which is an intense and disabling experience for adolescents (Rao et al., 2007). SAD refers to excessive and persistent fear of being evaluated in social interactions, also with peers, which leads to those situations being faced with high levels of suffering or, if possible, avoided (APA, 2013). If not subject to specialized intervention, SAD usually has a chronic course and is associated with the emergence of other psychiatric disorders (Essau et al., 2004). In Portugal, it is estimated about 8% of adolescents in the community meet criteria for SAD, including intense anxiety and frequent avoidance of social situations, as well as severe impairment of academic and social functioning and quality of life. Most of these adolescents (roughly 94%) did not seek or were referred to specialized intervention (Vagos et al., 2021), which points to the importance of offering acceptable, practical, and effective forms of intervention for PAS.

With adults, the treatment of choice for SAD is Cognitive Therapy (CT; Mavranezouli et al., 2015) derived from the model by Clark and Wells (1995). This model proposes that social anxiety is maintained by cognitive biases (i.e., negative automatic thinking and self-focused attention) that support the practice of safety or avoidance behaviors, so that the intervention is aimed at making the focus of attention more flexible in order to be able to invest in and be exposed to social situations and extract evidence from them that may help adjust one's own thoughts and social expectations. Based on a review of evidence of the presence of various components of this model also in adolescents, Leigh and Clark (2018) propose that it could be applicable to the understanding and intervention in SAD in adolescents, contributing to better therapeutic results. In fact, previous work supports the therapeutic effectiveness (Leigh et al., 2021; Leigh & Clark, 2016) and applicability of this approach in different contexts (Creswell et al., 2021; Leigh et al., 2021).

While cognitive therapy focuses on modifying the cognitive content elicited by self-focused attention during social events as a means of preventing safety and/or avoidance behaviors in these situations (Clark & Wells, 1995), third-generation therapies (i.e., Compassion-focused Therapy - CFT and Acceptance and Commitment Therapy - ACT) emphasize the modification of the way the person relates to these cognitive contents. CFT proposes that this relationship be based on a compassionate attitude, given that psychopathology is based on a relationship with one's own internal experiences that is excessively critical and punishing, also in children and adolescents (Carona et al., 2017). There is previous evidence pointing to the association between self-compassion and social anxiety in adults (Werner et al., 2012) and adolescents (Gill et al., 2012), as well as to the effectiveness of CFT for SAD in adults (Boersma al., 2018; Gharraee et al., 2018). As for ACTC, its application with adolescents is also conceived, assuming that psychopathology results from an avoidant relationship with ones' own negative internal experiences (Hallibturton & Cooper, 2015). Still, along with the scant evidence linking its core constructs (i.e., psychological flexibility) to social anxiety, research on the efficacy of this intervention for SAD is scarce, although pointing to promising results with adults (Dalrymple, & Herbert, 2007; Khoramnia et al., 2020; Ossman et al., 2006). The applicability of these therapies to SAD in adolescence remains to be elucidated.

In view of the above, evidence on the efficacy of the therapeutic offer for SAD in adolescence is still scarce. So, it is relevant to explore the therapeutic results of CT, CFT and ACT, as well as to explore the mechanisms by which each one enables change and with whom it can be most effective. Previous research has pointed to the importance of changing core constructs in the CBT intervention with adults with SAD (e.g., Santoft et al., 2019), but evidence regarding the therapeutic approaches under study is scarce. In addition, and considering the relevance of the intervention being adjusted to the characteristics of the person receiving the intervention (Beck et al., 2010), it is important to explore the moderators of therapeutic change, namely with regard to the severity of symptoms within what is proposed to be the main psychological vulnerability for each of the approaches under analysis.

The present work used a parallel clinical trial methodology (although different parallel conditions were collected sequentially over time) to assess the efficacy of three therapeutic approaches (i.e., CT, CFT and ACT) with adolescents with SAD, compared to a waiting list control group, and using four data collection times (i.e., pre-intervention/time 0, post-intervention/time 1, 12-weeks follow-up/time3, and 24-month follow-up/time 4). All participants included in the waiting-list control group were offered the opportunity to receive psychological intervention, after the waiting period. All procedures involved in this study (i.e., sampling, data collection to assess the efficacy of the intervention, and implementation of the intervention) were implemented after a positive position from the Ethics Committee of the institution hosting the investigation, authorization from the General Directorate of Education on the data collection protocol to be implemented, partnership with national secondary schools, authorization of legal guardians of adolescents, and consent of the adolescents themselves. Furthermore, this investigation complied with all applicable ethical and deontological requirements.

The sampling process included two phases. The first phase was the screening of participants. Collaboration protocols were established with 26 secondary schools so that all 10th and 11th grade students of a school were invited to fill out a self-report questionnaire that assesses social anxiety, thus contributing to homogeneity with regard to the age group and social and academic experiences of the potential participants. Students with scores on this questionnaire above the normative average in Portugal were invited to participate in an individual assessment through a semi-structured diagnostic interview, to verify compliance with the inclusion and exclusion criteria described elsewhere in this form. Having fulfilled these criteria, adolescents were invited to join the study and fill out selected primary and secondary outcome measures to assess the efficacy of the interventions. They were firstly allocated to the waiting list-control group; later, participants from the waiting-list as well as newly recruited participants were sequentially allocated to one of the intervention groups over time The waiting-list control group was recruited in the first year of the research. Fifty-two participants initially accepted to participate in the study, but only 41 adolescents responded to at least one assessment moment, with the other 11 refusing to participant before any assessment moment. The 41 participants were contacted at the four evaluation moments as stated above. Of these, 21 participants completed all assessment moments. Participants initially allocated to the waiting-list control condition (including the ones that did not complete the 4 assessment moments) were contacted to be re-evaluated in the beginning of the second year of the research. The ones still fulfilling the inclusion criteria were offered the possibility of receiving an intervention (i.e., of being allocated to an experimental condition). Of the 52 participants that initially accepted to participate in the study, we were able to contact and evaluate 30 of them - 28 still met criteria for inclusion in an experimental group, 1 did not present SAD symptoms and 1 was receiving psychological intervention. Of the 28 participants fulfilling inclusion criteria, 3 refused the intervention and 25 were randomly allocated to the CFT or CT condition (20 to the CFT and 5 to the CT condition), along with another newly recruited 20 participants (18 to the CT and 2 to the CFT interventions). At this moment, 22 participants had been allocated to the CFT condition (1 of which dropped out) and 23 were allocated to the CT condition (6 of which dropped out). At the end of the second year of the research project, 21 participants had completed the CFT intervention and 17 had completed the CT intervention. In the third year of the project 28 adolescents (3 screened in the second year of the project and 25 screened in the second year of the project) were recruited and assigned to the ACT and CT conditions (4 to the CT intervention and 24 to the ACT intervention) so that all conditions had the minimum number of participants necessary for data analysis. Two participants dropped out of the ACT condition. In total, 92 participants enrolled in the study. All participants in all conditions were followed and continue to be followed over the four defined moments. Additionally, participants were/will be evaluated to verify the presence or absence of SAD at the 24-month follow-up/time 4 using the same semi-structured diagnostic interview used to assess initial inclusion/exclusion criteria.

Interventions followed a structured brief manual (i.e., 10 weekly sessions and 2 booster sessions 1 and 2 months after the core program was completed) and were implemented in individual online format via videoconference, according to the time availability of each adolescent; the facilitator had adequate training in psychology and in the therapeutic approach to be implemented.

Collected data will be used to investigate the therapeutic efficacy of each for the experimental intervention conditions across the four assessment moments, on its own, in comparison with each other, and comparing to a waiting-list control condition. Psychotherapeutic efficacy is usually tested against waiting-list control groups, based on the assumption that if the intervention and control groups started similarly at pre-intervention and present differently at post-intervention, the intervention is responsible for the observed change. This procedure is well established in the literature but has seldom been applied to adolescent SAD or to the different theoretical approaches tackled in this trial. Processes of change will also be analyzed as reflected in therapeutic change in constructs that were considered transversal to the intervention (e.g., social anxiety) being mediated by change in core constructs for each intervention. In addition, the moderating effect of initial levels of interference and intensity of symptomatology will be explored across experimental intervention conditions.

Inclusion criteria for all experimental and the control group are:

• Aged between 15 and 18 years old at screening phase;
• Self-reported intensity of social anxiety higher than that reported by a normative comparable sample (Cunha et al., 2004);
• Main diagnosis of Social Anxiety Disorder as evaluated via the Mini-Kid (Rijo et al., 2016).

Exclusion criteria for all experimental and the control group are:

• Psychotic symptoms or suicidal risk as assessed via the Mini-Kid (Rijo et al., 2016);
• Being part of the special needs teaching system;
• Currently receiving psychological intervention/counselling.

IPD will not be shared as participants were guaranteed anonymity and confidentiality of their individual data. Datasets with alphanumerical identities given to each participant may be made available, either by publishing data or upon reasonable request directed at with the principal investigator.

The principal investigator is unaware of which participants are allocated to which condition; she is only aware of the inclusion criteria all participants must fulfill. The investigator and outcome assessor/data analyst are also unaware of treatment outcome and blind to the allocation condition.