校园欺凌是严重威胁学生健康发展的全球性问题。为探索青少年述情障碍特质与欺凌和受害行为的关系，并考察学校氛围对两者关系的调节作用，本研究采用整群方便取样，采用校园欺凌量表、多伦多述情障碍量表与特拉华校园氛围量表对北京市两所中学的初中生进行为期两年的追踪测查，结果发现：(1) 对青少年述情障碍特征进行潜在剖面分析，统计指标支持三剖面模型，分别定义为识别描述障碍组(25.1%)、典型障碍组(45.0%)和功能良好组(29.9%)。识别描述障碍组在情感识别困难、情感描述困难维度上的得分最高，在外向性思维维度上的得分显著低于典型障碍组，与功能良好组不存在显著差异；典型障碍组各个维度得分均属于中等水平，但在外向性思维得分上最高；功能良好组各个维度得分均属于最低水平，不存在典型的述情障碍特征。(2) 性别和年级会影响青少年述情障碍剖面的分布，但母亲受教育水平不影响青少年述情障碍剖面的分布。基于性别的检验发现，相较于功能良好组，女生比男生更可能归入识别描述障碍组，尽管这一差异边缘显著；基于年级的检验发现，相较于典型障碍组，初一学生比初二学生更可能归入识别描述障碍组。(3) 以变量为中心的分析发现，在控制性别、年级后，情感描述困难显著正向预测一年后的欺凌行为，情感识别困难、外向性思维显著正向预测一年后的受害行为。以人为中心的分析进一步发现，识别描述障碍组、典型障碍组青少年的欺凌行为均显著高于功能良好组，识别描述障碍组青少年的受害行为显著高于典型障碍组和功能良好组。上述结果说明识别描述障碍组青少年风险最高，参与欺凌/受害行为最为严重，紧随其后的是典型障碍组。(4) 述情障碍剖面与欺凌/受害行为间的关系受到学校氛围的调节作用。就欺凌行为而言，典型障碍组青少年感知到的学校氛围能够显著负向预测一年后的欺凌行为，在识别描述障碍组、功能良好组中则不存在类似的负向关系。就受害行为而言，识别描述障碍组、典型障碍组青少年感知到的学校氛围能够显著负向预测一年后的受害行为，在功能良好组青少年中则不存在类似的负向关系。上述结果符合素质-压力模型，说明消极的学校氛围对典型障碍组的欺凌/受害、对识别描述障碍组的受害具有重要影响。

School bullying is a global problem that threatens the positive development of adolescents. To investigate the relationship between alexithymia and bullying/ victimization, as well as the moderating role of school climate, this study adopted the random cluster sampling method to conduct a two-year longitudinal assessment. And students from two middle schools of Beijing completed the School Bullying Scale, Toronto Alexithymia Scale, and Delaware Campus Climate. The results showed that: (1) The model fit indices supported that there were three distinct groups of alexithymia among adolescents, namely identify and description disorder group(IDD, 25.1%), typical alexithymia group(TA, 45.0%) and well-functioning group(WF, 29.9%). The IDD scored highest on the difficulty identifying feelings(DIF) and difficulty describing feelings(DDF), while the scores of externally oriented thinking(EOT) in IDD group were significantly lower than those in TA group, but there was no significant difference of EOT between IDD and FW group. (2) Gender and grade predicated the profile of alexithymia, but the mothers’ educational level had no such effect. Girls were more likely than boys to be classified into the IDD group, rather than the WF group; Grade-seven students were more likely than grade-eight students to be classified into the IDD group, rather than the TA group. (3) After controlling for gender and grade, the variable-centered analysis found that DDF significantly positively predicted bullying behaviors one year later, while DIF and EOT significantly positively predicted victimization one year later. The person-centered analysis further found that bullying in the IDD and TA group were significantly higher than the WF group, and victimization in the IDD group was significantly higher than the TA and WF group. These results indicated that adolescents in the IDD group have the highest risk of engaging in bullying and victimization behaviors, followed closely by the TA group. (4) The relationship between the profiles of alexithymia and bullying/victimization was moderated by school climate. In terms of bullying, the perceived school climate by IDD group significantly negatively predicted bullying behaviors one year later, but there was no similar negative relationship in the TA and WF group. In terms of victimization, the perceived school climate by IDD and TA group significantly negatively predicted victimization one year later, but there is no similar negative relationship in the WF group. These results supported the diathesis-stress model, indicating that negative school climate has an important effect on bullying/victimization in the IDD and TA group.