同源主观数据是否膨胀了变量间的相关性——以战略人力资源管理研究为例

发布时间：2017-12-28 17:06

本文关键词：同源主观数据是否膨胀了变量间的相关性——以战略人力资源管理研究为例　出处：《武汉大学学报(哲学社会科学版)》2015年06期 　论文类型：期刊论文

【摘要】：研究变量之间的相关性是现代管理学实证研究的基本范式。然而,变量之间统计上的相关性极易受到变量的测量方法和数据类型的影响。特别是,基于同源数据的实证研究会夸大变量间的相关程度甚至带来虚假相关,这已经给管理学实证研究带来了严峻挑战。不幸的是,目前仍有大量的实证研究都在使用同源主观数据。本文以人力资源管理和企业绩效的关系为例,检验了在同一样本中,对变量的不同测量和数据类型如何导致变量间相关性结论的差异。结果显示,当自变量和因变量是"同源主观数据"时,自变量和因变量之间的相关程度最高;当自变量和因变量是"非同源主观数据"时,相关程度有所下降;当自变量采用主观数据,而因变量采用客观数据时,这种相关没有达到显著性。本文的研究例子表明,在管理学实证研究中一定要谨慎使用同源的主观测量数据。
[Abstract]:The study of the correlation between variables is the basic paradigm of the empirical study of modern management. However, the statistical correlation between variables is easily affected by the measurement method of variables and the type of data. In particular, empirical research based on homologous data will exaggerate the degree of correlation between variables and even bring about false correlation, which has brought serious challenges to management empirical research. Unfortunately, there are still a large number of empirical studies that use homologous subjective data. Taking the relationship between human resource management and enterprise performance as an example, this paper examines the difference between the different measurement and data types of variables in the same sample. The results showed that when the independent and dependent variables are "homologous subjective data", the independent and dependent variables between the highest degree of correlation; when the independent and dependent variables are "non homologous subjective data", related to the degree of decline; when the arguments are subjective data, and the dependent variable with objective data, this correlation did not reach significant of. The example of this study shows that the homologous subjective measurement data must be used carefully in the empirical study of management.
【作者单位】：中国人民大学劳动人事学院;澳大利亚新南威尔士大学商学院;
【基金】：中央高校基本科研业务费专项资金资助(11XNJ009)
【分类号】：F272.92
【正文快照】： 一、引言研究变量与变量之间的相关性是现代管理学实证研究的基本范式。学者们正是通过对各种变量进行测量然后检验变量之间的相关程度,从而不断构建和拓展管理学的理论知识体系。因此,对变量的测量和所获得的数据质量是决定管理学实证研究质量的重要因素。然而,在管理学实证

【共引文献】