Clark和Wells (1995)关于社交焦虑障碍的认知模型认为，在（预期）进入社交情境时，社交焦虑个体的关于自我和他人的消极信念被激活，从而对社交情境有解释偏向，在接下来的认知过程中对自己的表现进行消极评价、对他人所呈现出的社交线索进行消极解释，而事前事后的一系列认知加工（比如，事前和事后反刍）和安全/回避行为等又会维持这种消极解释的倾向。但是，截至目前，还未有实证研究来支持有关于社交情境的解释偏向可以通过反刍来得以维持社交焦虑症状的假设。本论文聚焦于高社交焦虑青少年的反刍（包括事前和事后反刍），主要验证反刍在解释偏向和社交焦虑之间的中介作用。共设计了三个研究，分别验证反刍的中介作用、在经历不同类型的社交情境时反刍中介作用是否仍存在、从干预的角度验证反刍在维持社交焦虑症状起到的中介作用。 研究一，为了探索性地考察解释偏向、反刍和社交焦虑症状之间的关系：1）首先在1808名11-17岁青少年群体中通过问卷施测收集数据的方法，对解释偏向、事后反刍和社交焦虑症状进行中介效应检验（子研究一）；结果显示，事后反刍在解释偏向和社交焦虑症状之间起到部分中介作用。2）子研究二中，在实验情境下现场创建社交任务，分别招募11-13岁的31名和37名高低社交焦虑青少年进行5分钟演讲，并在预期进入演讲任务前对事前反刍和演讲任务结束后以及结束一天后这两个时间点（T1和T2）的事后反刍和社交焦虑水平进行测量；结果显示，解释偏向可以显著预测事前反刍，并完全通过事前反刍分别影响演讲任务后的事后反刍_T1和事后反刍_T2，并维持一天后的社交焦虑，即，事前和事后反刍在解释偏向和社交焦虑中起到完全中介作用。 研究二，进一步验证反刍的中介作用是否会因为经历不同类型的社交情境而有所不同（中介作用的情境差异性）。在实验情境下现场分别创建5分钟的演讲任务（即，“演讲组”）和5分钟的聊天任务（即，“聊天组”），招募高社交焦虑青少年61名（11-13岁），随机分配到演讲和聊天组，并在预期进入演讲任务前对事前反刍和演讲任务结束后以及结束一天后对这两个时间点（T1和T2）的事后反刍和社交焦虑进行测量。研究结果显示，在刚经历完社交任务后，演讲组青少年的事后反刍（T1）得分显著高于聊天组被试。有调节的中介分析结果则表明，演讲情境在解释偏向和事前反刍、以及在解释偏向和事后反刍_T1这两条路径上起到调节作用；即，只有在演讲情境中，解释偏向才会通过事前反刍得以影响事后反刍_T1，事前反刍起到完全中介作用；同样地，只有在演讲情境中，解释偏向才完全通过事后反刍_T1得以维持社交焦虑_T2。 研究三，从干预的角度验证反刍在维持社交焦虑症状中起到的中介作用。招募73名11-16岁高社交焦虑青少年随机分配到演讲组或聊天组，通过编制与日常学习生活情境相关的认知重评句的方式对青少年经历不同类型社交情境后所表现出的事后反刍进行有针对性地预防干预。结果显示，控制抑郁后，认知重评句可以有效缓解被试在经历社交事件后的反刍水平和社交焦虑症状；而且，事后反刍得分的减少显著预测社交焦虑水平的降低，验证了事后反刍维持社交焦虑症状的中介作用。但是，与预期不一致的是，干预效果并不存在显著的组别差异。 本论文考察了高社交焦虑青少年反刍在解释偏向和社交焦虑症状中的中介作用，分别从问卷和情境创建、情境差异性和干预的角度来对反刍的中介作用进行验证，将Clark和Wells的认知模型中相关假设在青少年群体中进行了验证，而且还确认了模型假设适用的范围，并从干预的角度再次验证并为干预青少年社交焦虑症状提供了启示。

Clark and Wells (1995)’s Cognitive Model for Social Phobia postulated that, when entering, more accurately, when expecting to enter social situation, high socially anxious individuals’ negative beliefs about themselves and others would be activated, making them explain the ambiguous social cues in a threatening manner. The tendency to negatively interpret would be termed as “Interpretation Bias” (IB). Following processes, including pre- and post-event cognitive processing, self-focused attention, overestimating negative appraisals from others and safety behaviors, would enhance and further maintain the inclination to biasedly interpret. However, so far, no experimental evidence was provided to examine whether interpretion bias related to social stimulus would maintain social anxiety level via rumination. As continuity and repetition characterize both pre- and post-event rumination, and the influence of rumination on social anxiety symptoms would continue for quite a while, hence, the present study focused on socially anxious adolescents’ rumination, including pre- and post- rumination. The main objectives were to examine the mediation effect of rumination between the inclination to biasedly interpret and social anxiety symptoms. To achieve this, three studies were conducted with questionnaires, experimentally establishing social situations (speech and/or interaction) and intervention, respectively. In Study 1, to firstly explore the association between interpretation bias, post-event rumination and social anxiety symptoms, 1808 adolescents ranging from 11 to 17 years were recruited and measured via a pack of questionnaires (sub-study 1). Results found that negative rumination after experiencing social events would partially mediate the inclination to negative interpret the ambiguous social details and the level of social anxiety. In sub-study 2, by experimentally establishing social task, i.e., 5-min self-introducing speech, 37 high and 31 low socially anxious adolescents (age range is 11 to 13) were recruited to assess their anticipatory rumination, post-event rumination (T1 and T2) and state social fear (T1 and T2) before, right after and one-day after the speech task. The results demonstrated that, the tendency to negatively interpret the ambiguous social cues would significantly predict pre-rumination when expecting to enter the speech situation and post-event rumination after finishing the speech task, both at Time 1 and Time 2, respectively, and the post-event rumination would further maintain the social anxiety level one-day after the social task. Namely, pre- and post-event rumination would fully mediate the relationship between interpretation bias and social anxiety. In Study 2, the main aim was to examine whether the mediation effect would show any difference in two different and typical social situations, i.e., 5-min speech v.s. 5-min interaction. To achieve this, 61 high socially anxious adolescents ranging from 11-13 years were recruited and were measured their pre-event rumination, post-event rumination (T1 and T2) and state social anxiety (T1 and T2) before, right after and one-day after the social task. The moderated mediation effect results displayed that, social situation type moderated the relationship between interpretation bias and pre-event rumination and the association between interperation bias and post-evenet rumination at Time 1. Namely, only in speech task, the tendency to negatively interpret could significantly trigger post-event rumination right after the task fully via pre-event rumination. In addition, the post-event rumination at Time 1 only played the role of full mediator in the relationship between interpretation bias and social anxiety one-day after the speech. In Study 3, the mediation effect of rumination was again exmianed by conductiong a pilot intervention targeting post-event rumination. 73 adolescents (11-16 ages) with high level of social anxiety symptoms, by developing typical social-related realistic thoughts based on the rationale of Cognitive Behavioural Therapy, i.e., cognitive construction technique, were recruited and randomly assigned to either imaged speech or interaction situation group. The intervention results showed that, after receiving cognitive construction, their level of post-event rumination and social anxiety was significantly improved even after controlling for depression. Moreover, the decrease of post-event rumination could significantly predict the reduction of social anxiety score. However, unexpected to hypothesis, the intervention effect did not show any difference between two social situations. The current research investigated the meditaion role of rumination between the relationship interpretation bias and social anxiety symptoms among high socially anxious Chinese adolescents via questionnaires and experimentally establishing social tasks, situational difference of the role of rumination and intervening. This paper not only examined the related hypothesis from the cognitive model of Clark and Wells in youth sample, and also confirmed the application scope of the model. Additionally, to examine the role of rumination and to further provide enlightenment with respect to intervening anxious adolescents’ social anxiety symptoms.