已有研究表明青少年攻击和自伤行为有着普遍的联系以及共发性。然而，前人研究常 常独立考察攻击或自伤行为，缺乏对两者共发类型以及随时间发展变化的探究。本研究通 过纵向大样本数据，描绘了青少年攻击和自伤共发类型的个体异质性。在此基础上，研究 进一步从学校生态系统的角度出发，聚焦于同伴因素（受欢迎、友谊质量）、教师因素（教 师公正、师生关系）和学校氛围对青少年攻击和自伤的发生及发展变化的预测作用。 

研究选取了河南省郑州市 2 个区县的 7 所高中共 47 个班级的学生进行追踪测试，时间间隔为 6 个月。T1 时间点收到有效问卷 2601 份，其中男生 1339 人 (51.5%)，平均年龄 13.96 岁 (SD=0.62)；T2 时间点共收到有效问卷 2537 份，其中男生 1300 人 (51.2%)，平均 年龄 14.42 岁 (SD=0.62)，被试流失率为 2.4%。 

研究一考察了两个问题：1a 运用横向研究，采用潜剖面分析方法(LPA)，考察青少年 攻击和自伤行为的共发类型；1b 采用协变量多元 logistic 回归模型，考察学校因素对青少 年攻击和自伤类型的预测作用。结果发现，青少年攻击和自伤行为有 3 个类别：低症状者 (低自伤低攻击)、高攻击-中等自伤者、以及高症状者(高自伤高攻击)，其中低症状者比例为 89.4%，高攻击-中等自伤者比例为 8.4%，高症状者比例为 2.2%。学校因素中，师生关系和 学校氛围是影响青少年攻击自伤共发类型的重要因素。 

研究二在研究一的基础上，考察纵向的发展变化：2a 采用潜转变分析方法(LTA)，对青 少年攻击和自伤共发类型的转化模式进行考察；2b 采用协变量多元 logistic 回归模型，考 察学校因素对青少年攻击自伤类型转化的作用。

结果发现，攻击和自伤类别在半年时间内保持一致，但 3 个类别具有不稳定性和可转化性，其中高症状者和高攻击-中等自伤者具有 较弱的稳定性。师生关系是影响青少年攻击自伤共发类型转化的重要因素，支持性师生关 系是低症状和高攻击-中等自伤者的保护性因素，冲突性师生关系则是低症状者的风险因素。 好的学校氛围对于高攻击-中等自伤者有保护作用，高的友谊质量对低症状者有保护作用。 研究结果有助于深入理解我国青少年群体攻击和自伤行为在同一时间、不同时间的存 在类型和发展变化，为青少年问题行为的动态监测和干预提供了一定的依据。此外，确定 学校层面的重要预测因素，为学校教育工作提供针对性的预防和干预方案提供了指导意义。

It has been shown that adolescent aggression and self-harm are generally associated and co-occur. However, previous studies have examined aggression and self-harm independently, lacking an exploration of the types of co-occurrence and development over time. The combination of cross-sectional and longitudinal analysis, variable-centered and individual-centered approaches were used to depict individual heterogeneity in the co-occurrence types of aggression and self-harm in adolescents through longitudinal large sample data. Based on this, the study further focused on the predictive role of peer factors (popularity, friendship quality), teacher factors (teacher justice, teacher-student relationship), and school climate on the occurrence and developmental process of adolescent aggression and self-harm from the perspective of school ecosystem.

A total of 47 classes of students from 7 high schools of Zhengzhou City, were selected for follow-up testing over a 6-month interval. 2601 valid questionnaires were received at T1, of which 1339 (51.5%) were boys with a mean age of 13.96 years (SD=0.62); 2537 valid questionnaires were received at T2, of which 1300 (51.2%) were boys, with a mean age of 14.42 years (SD=0.62). The subject attrition rate was 2.4%.

Study 1 examined two questions: 1a used a cross-sectional study with latent profile analysis (LPA) to examine the co-occurrence types of adolescent aggression and self-harm; 1b used logistic regression model to examine the role of school factors on the aggression and self-harm types. Results revealed three categories of co-occurring adolescent aggression and self-harm: low symptom (low self-harm, low aggression), high aggression-moderate self-harm, and high symptom (high self-harm, high aggression), with the highest proportion of low symptom group followed by high aggression-moderate self-harm group. Teacher-student relationships and school climate were important factors influencing the co-occurrence type of aggression and self-harm in adolescents.

Study 2 examined longitudinal developmental changes based on Study 1: 2a used latent transition analysis (LTA) to examine the transformation patterns of three types; 2b used logistic regression models to examine the role of school factors on the transformation of adolescent aggression and self-harm types. The results found that the aggression and self-harm categories remained consistent over the six-month period, but the three categories were unstable and transformable, with lower stability of high symptom group and high aggression-moderate self-harm group. Teacher-student relationship was important factor influencing the transformation of the types of aggression and self-harm. Supportive teacher-student relationship was protective factor of the low-symptom and high aggression-moderate self-harm groups, while conflicting relationship was risk factor of low symptom adolescents. Good school climate was benefit for high aggression-moderate self-harm adolescents. High friendship quality had a protective effect on low symptoms group.

The study contributes to the understanding of the types of aggression and self-harm that exist and develop at the same time and across time in Chinese adolescents, providing the basis for dynamic monitoring and intervention of adolescent problem behaviors. What’s more, identifying important predictors at the school level provides some guidance and insight for school educators to provide targeted, individualized prevention and intervention programs for students with co-occurring behaviors.