科研管理 ›› 2018, Vol. 39 ›› Issue (7): 139-150.

• 论文 • 上一篇    下一篇

基于云模型的企业社会责任评价模型及应用

孟斌,钮尓轩,匡海波,骆嘉琪   

  1. 大连海事大学 综合交通运输协同创新中心,辽宁 大连116026
  • 收稿日期:2017-06-15 修回日期:2017-10-19 出版日期:2018-07-20 发布日期:2018-11-06
  • 通讯作者: 匡海波
  • 基金资助:

    国家自然科学基金项目(71731003, 71672016, 71503199, 71431002, 71301017);长江学者和创新团队发展计划(IRT_17R13, IRT13048);国家社会科学基金项目(16BTJ017);辽宁省高等教育内涵发展专项资金资助项目(20110117201);大连市社科联重点课题(2016dlskzd042, 2017dlskzd034)。

An evaluation model and empirical research on CSR based on cloud model

Meng Bin, Niu Erxuan, Kuang Haibo, Luo Jiaqi   

  1. Collaborative Innovation Center for Transport Studies, Dalian Maritime University, Dalian 116026, Liaoning, China  
  • Received:2017-06-15 Revised:2017-10-19 Online:2018-07-20 Published:2018-11-06

摘要: 交通运输行业作为关乎国计民生的支柱性产业,在国家政策引领下加大重视企业社会责任实践是必然趋势。本研究以国际标准化组织的ISO26000、全球报告倡议组织的G4标准等权威机构报告及相关文献为基础,结合企业社会责任内涵,依据权威机构典型指标高频率原则,通过方差膨胀因子剔除反映信息重复的指标,通过主成分分析遴选出对企业社会责任评价结果影响显著的指标,构建了包括环境、人权等5个准则层,循环再造物料的使用、烟粉尘排放及减排等45个指标的企业社会责任海选评价指标体系。通过云模型构建交通运输行业企业社会责任评价模型。文章的创新与特色:一是通过建立某准则层内一个指标与其他所有指标的线性回归方程,求解反映指标相关性的指标方差膨胀因子VIF,剔除指标方差膨胀因子VIF大于阈值的指标并保留剩余指标,避免了指标反映信息重复。二是以方差贡献率为权重对因子分析中的因子载荷的绝对值进行加权,使指标信息含量的确定不但利用因子载荷体现第j个因子解释第i个指标信息的比重,还反映第j个因子解释原始指标群信息的比率,遴选对企业社会责任评价影响显著的指标,构建了对结果影响显著的企业社会责任评价指标体系。三是通过云模型构建交通运输行业企业社会责任评价模型,不仅考虑了企业社会责任概念的不确定性,也体现了交通运输行业的随机性与企业社会责任的模糊性之间的联系,在定性与定量之间形成相互映射的关系,通过云模型对企业社会责任评价指标进行赋权,测度了交通运输行业企业社会责任的履行情况。

关键词: 企业社会责任评价, 评价指标体系, 交通运输行业, 云模型

Abstract: As a pillar industry related to the people's livelihood, it is an inevitable trend for the transportation industry to increase the emphasis on corporate social responsibility practice under the guidance of national policy. This paper is based on the ISO26000 of the International Organization for Standardization, the G4 standard of the Global Reporting Initiative, and other relevant authoritative reports and related literature, then it combined with the connotation of CSR and the principle of typical indicators of high frequency in the authority agency to eliminate the indicators that reflect the repetition of information through the variance expansion factor and select the indexes of significant impact on the CSR evaluation through the principal component analysis. Hence five criterion layers (including the environment, human rights, etc.) and 45 indicators (including the use of recycled materials, smoke and dust emission and reduction, etc.) are constructed. The could model is used to construct the CSR evaluation model of transportation industry. The innovation and characteristics of this paper: First, by establishing a linear regression equation of one index and all other indicators, the index variance expansion factor VIF is used to solve the correlation index, and then remove the variance expansion factor which VIF is larger than the threshold and keep the remaining indicators to avoid the indicators reflecting the information repetition. Second, the absolute value of the factor load in the factor analysis is weighted by variance contribution rate, using the factor load to reflect the jth factor to explain the proportion of the ith indicator information, then screen significant indicators of the CSR evaluation and construct the CSR evaluation index system. The third is to construct a model of the performance evaluation for the CSR of the transportation industry based on the cloud model, it not only considers the uncertainty of the concept of the CSR, but also reflects the connection between the randomness of the transportation industry the vagueness of the CSR, it forms a mutual mapping relationship between qualitative and quantitative, and weights the evaluation index of the CSR through the cloud model, finally it measures the implementation of the CSR of the transportation industry.

Key words: CSR evaluation, evaluation index system, transportation industry, cloud model