中文摘要：

采用跨文化（中国人）个性测量表 （ Cross - Cultural [ Chinese ] Personality Assessment Inventory,，简称CPAI-2）对182名服务企业员工进行本土化人格特质测查，并同时获得其直接上级主管对他／她工作绩效的评定。目的是探讨本土化人格特质与工作绩效之间的关系模式（线性或非线性）。结果表明：和谐性与工作绩效具有正向的线性关系；面子与工作绩效具有倒U曲线关系；人情与工作绩效具有正U曲线关系；其他本土化人格特质与工作绩效的关系均不显著。

英文摘要：

Chinese personality traits have unique differences from other cultures due to China＇s special geography and collectivist cultural background. Therefore, the philosophy of localization was accepted and advocated by more and more Chinese psychologists. Meanwhile, the nonlinear relationships between personality traits and job performance have been found in some studies. These results transformed the traditional top - down strategy into double strategies for the applications of personality tests in the psychology of personnel management. However, there was no research to combine the linear and nonlinear models to examine the relationships between indigenous personality traits and job performance in the Chinese work setting. In the current study, we explored the linear and/or nonlinear reationships between indigenous personality traits and job performance. The participants were 182 service employees from several service industries in Beijing, and the immediate supervisors of the respondents provided ratings of their job performance and returned independently to the interviewers. The ratings of job performance were self - compiled based on job analysis and in - depth interview. Confirmatory factor analysis found that the one - factor performance model had a better fit （χ2 = 27.79, df = 8, CFI = . 95, NF1 = . 93, RMSEA = . 07）, and the coefficient alpha was . 85. Meanwhile, the nine subscales from CPAI -2 were selected to assess the indigenous personality traits： face （FAC）, family orientation （FAM）, defensiveness （DEF）, graciousness vs. meanness （G_M）, veraciousness vs. slickness （V_S）, traditionalism vs. modernity （T_M）, renqing （REN）, harmony （HAR） and thrift vs. extravagance （T_E）. The mean Cronbach＇s coefficient for the entire set of personality scales was . 70 in the representative normal sample in Chinese mainland and Hong Kong. The data were analyzed with SPSS 15.0, and the main statistical methods were correlation analysis and hierarchical polynomial regression