Traditional DEA-based ranking techniques have some pitfalls such as ignoring the inherent differences among decision making units （DMUs）, or lacking a common weight-based ranking, etc. To overcome these obstacles, t...Traditional DEA-based ranking techniques have some pitfalls such as ignoring the inherent differences among decision making units （DMUs）, or lacking a common weight-based ranking, etc. To overcome these obstacles, the paper first examines the possible differences among all DMUs such as the technical efficiency difference, the preference structure difference and the within-group position difference. Based upon the above differences this paper induces an integrated ranking measurement which helps to give a fair and full ranking for all DMUs under evaluation. Following the three types of differences, this approach behaves greatly elaborately, accurately and reasonably. Finally, the results from the Olympics achievement evaluation approve the acceptability of this approach.展开更多
组织管理领域的多层次研究经常需要测量共享单位特性构念,常用方法是将单位内若干个体成员的评分聚合到单位层次,确保聚合后的分数具有充分代表性的统计前提是通过聚合适当性检验。聚合适当性检验的常用指标是组内一致性rWG和组内信度IC...组织管理领域的多层次研究经常需要测量共享单位特性构念,常用方法是将单位内若干个体成员的评分聚合到单位层次,确保聚合后的分数具有充分代表性的统计前提是通过聚合适当性检验。聚合适当性检验的常用指标是组内一致性rWG和组内信度ICC(1)、ICC(2),但目前学界对于这两类指标何者更优、rWG的原分布选择和数据清理、各指标的划界值等关键问题存在诸多争议。为此,首先对国内9份管理学、心理学期刊2014年以来发表的166篇包含聚合适当性检验的论文进行内容分析,并以Journal of Applied Psychology上的85篇论文为对比,查明常规实践中的共性问题,进而提出实践建议:(1)明确功能定位,将rWG作为聚合适当性指标,ICC(1)和ICC(2)分别作为效度、信度指标。(2)计算rWG时审慎选择原分布,排除组内一致性过低的组。(3)为各指标设定更加合理、有适度灵活性的划界值,停止使用武断、粗糙的经验标准。最后,强调研究者在模型构建和聚合决策中应加强理论考量,避免片面依赖统计检验结果。展开更多
基金supported partly by the National Natural Science Fundation of China for Innovative Research Groups(T0821001)the National Natural Science Fundation of China(70801056)University of Science and Technology of China Science Funds for Young Scholars.
文摘Traditional DEA-based ranking techniques have some pitfalls such as ignoring the inherent differences among decision making units （DMUs）, or lacking a common weight-based ranking, etc. To overcome these obstacles, the paper first examines the possible differences among all DMUs such as the technical efficiency difference, the preference structure difference and the within-group position difference. Based upon the above differences this paper induces an integrated ranking measurement which helps to give a fair and full ranking for all DMUs under evaluation. Following the three types of differences, this approach behaves greatly elaborately, accurately and reasonably. Finally, the results from the Olympics achievement evaluation approve the acceptability of this approach.
文摘组织管理领域的多层次研究经常需要测量共享单位特性构念,常用方法是将单位内若干个体成员的评分聚合到单位层次,确保聚合后的分数具有充分代表性的统计前提是通过聚合适当性检验。聚合适当性检验的常用指标是组内一致性rWG和组内信度ICC(1)、ICC(2),但目前学界对于这两类指标何者更优、rWG的原分布选择和数据清理、各指标的划界值等关键问题存在诸多争议。为此,首先对国内9份管理学、心理学期刊2014年以来发表的166篇包含聚合适当性检验的论文进行内容分析,并以Journal of Applied Psychology上的85篇论文为对比,查明常规实践中的共性问题,进而提出实践建议:(1)明确功能定位,将rWG作为聚合适当性指标,ICC(1)和ICC(2)分别作为效度、信度指标。(2)计算rWG时审慎选择原分布,排除组内一致性过低的组。(3)为各指标设定更加合理、有适度灵活性的划界值,停止使用武断、粗糙的经验标准。最后,强调研究者在模型构建和聚合决策中应加强理论考量,避免片面依赖统计检验结果。
