中文摘要：

以班级为单位选取1-6年级23个班级的小学生,通过最好朋友提名法,得到有互选朋友有效被试700人,采用同伴提名法、小学生友谊质量问卷、小学生人格发展教师评定问卷,运用多层线性模型（HLM）考察个体和班级两个水平上的同伴接纳、友谊质量对人格的影响,并在两个水平上检验友谊质量在同伴接纳对人格的影响上的多层中介效应。结果表明：（1）班级水平：班级平均友谊质量对外倾性、亲社会性、认真自控、情绪稳定性有预测作用;班级平均同伴接纳对情绪稳定性有预测作用。（2）个体水平：同伴接纳对人格5个维度均有直接影响;除情绪稳定性,友谊质量分别在同伴接纳对智能特征、外倾性、亲社会性、认真自控影响上有部分中介效应。

英文摘要：

The period in primary school is important for children’s personality development. Peers play an important role in personality formation and its development for children. We explored the effect of peer acceptance and friendship quality on the personality development of primary school students at individual and class levels by involving peer factor as a circumstance variable, centralizing the peer acceptance and friendship quality in the within-class group respectively and aggregating individual peer acceptance and friendship quality as two class-level variables. We examined how peer acceptance and friendship quality influenced the personality development in primary school students at both individual and class levels. Our research breaks the limitation of ignoring the influence of the class level social and psychological context effect in previous studies. We selected 700 participants as mutual friends with the class as a unit from 23 classes and 6 grades （1-6）, using the best friend nomination method. The data was collected with the peer nomination method and the primary school students’ friendship quality questionnaire and the teacher-assessment primary school students’ personality development questionnaire. According to the lower-level mediation model （1-1-1）, we respectively regarded the five dimensions of personality （intelligence, extraversion, pro-sociality, conscientiousness and emotional stability） as dependent variables in the Hierarchical Linear Modeling （HLM）, and examined the multilevel mediation effect of friendship quality on the relationship of peer acceptance and each dimension of personality by the multilevel mediation effect test procedures. The result indicates that for example intelligence, the average level of peer acceptance in class has no direct effect on intelligence（t=-0.090, p.05）. With the effects of the class level controlled, peer acceptance significantly predicted effects on intelligence in the path c （t=5.010, p.001）. Peer acceptance had significance pr