Abstract: Purposes：The transition to the adolescent is an important stage of children's personality development. When children enter junior middle school and face new developmental tasks， they need to adapt to both physiological and psychological changes， which make this stage an important research topic. Studies on junior middle school students' personality differences can be conducted with two different groups of methods: variable-centered vs. person-centered. The variable-centered methods focus on subjects' factor scores differences. Person-centered methods refer to within subjects' personality types. There are many personalities within subjects， but there are limited typical personality types at the overall level. Person-centered methods provide the profile plot， and include additional information beyond variable-centered methods. They supplement each other. This study is to examine whether the three personality types of adolescents can be replicated in Chinese context. Procedures：Personality development scale for Chinese junior middle school students has proved its reliability and validity by pilot measures， EFA from a sample of 583 subjects， and CFA from a sample of 592 subjects. The Internal consistency reliability， test-retest reliability construction validity and criteria-related validity with NEO-FFI of the scale are ideal as well. Using the Chinese culture as the context， 13-to-15-year-old Chinese children's personality structure is divided into five factors: agreement， intellectual characteristics， conscientiousness， extraversion， and emotional stability， with each including some secondary traits. The structure of the scale is coincided to the Five-Factor Model. Higher factor scores correspond to higher personality development level. In the present study， personality development scale for Chinese junior middle school students is used to measure a sample of 3602 subjects from 14 junior middle schools in the city. And five factors' Cronbach's Alpha coefficients range from .773 to .874. Methods：Latent class analysis and its fit indices such as AIC， BIC，adjusted BIC， entropy， LMR and BLRT output by Mplus are used to identify junior middle school students' personality types. ANOVA and multiple comparisons are used to describe the characteristics of each personality type. Cross-tabulation Chi-square test and multinomial Logistic regression are both used to test and model the development characteristics of personality types using SPSS. Results：According to Robins' theory of self-control and adaptation and the posterior probability of subjects allocation by LCA， junior middle school students' personality can be optimally divided into three types: undercontrolled， overcontrolled and resilient. The resilient personality type is the majority(85.90%)，and has the highest scores on all five factors. The overcontrolled personality type has the second highest factor scores on agreement， intellectual characteristics and conscientiousness， moderate extraversion and low emotional stability. The undercontrolled personality type has the lowest scores on most fators. The resilient personality type proportion has a significant trend of decreasing with upgrading while the other two types increase. The resilient personality type proportion of girls is higher than that of boys significantly. The interactive effect between grade and gender on personality types development is not significant. Conclusions：According to the self-control and adaptation theory and based on Five-Factor Model of personality， the rationality of three junior middle school students' personality types are identified under the context of Chinese culture； The resilient personality type is the majority(85.90%) and adaptive， while the undercontrolled(6.7%) and overcontrolled(7.4%) personality types both have high risk of adaptation. There are significant grade and gender effects on personality types.

Key words: Junior Middle School Students, LCA, Personality Types, Development Characteristics

摘要： 目的：划分初中生人格类型及考察初中生人格类型的年级与性别特点。方法：本研究使用初中生人格发展自我评定量表对某地区3602名初中生被试施测，使用潜在类别分析对初中生人格类型进行划分，使用无序多分变量的Logistic回归考察初中生人格类型的年级与性别特点。结果：依据人格类型划分的相关理论和潜在类别分析的拟合指数，初中生可划分为低控型，过度控制型和适应型三种人格类型，其中适应型人数占大多数。方差分析结果表明适应型在初中生人格五维度得分均显著高于另两类；过度控制型有低情绪稳定性和低外倾性，同时有中等程度的亲社会性、智能特征和认真自控水平，低控型人格类型在大部分人格维度得分均较低。随年级增长，初中生适应型人数比例有显著下降趋势，过度控制型和低控型比例有所上升。性别差异方面，女生人格类型的适应型人数比例显著高于男生，过度控制型和低控型比例则显著低于男生。结论：依据自我控制和自我适应理论，初中生三种人格类型的划分得到重复验证，人格类型有显著的年级和性别特点。

关键词: 初中生, 潜在类别分析, 人格类型, 发展特点