Are We Wielding this Hammer Correctly? A Reflective Review of the Application of Cluster Analysis in Information Systems Research
Are We Wielding this Hammer Correctly? A Reflective Review of the Application of Cluster Analysis in Information Systems Research
复制标题
DOI：
10.17705/1jais.00266
复制
发表时间：
2011-05
期刊：
影响因子：
--
通讯作者：
Venugopal Balijepally;George Mangalaraj;K. Iyengar
中科院分区：
文献类型：
--
作者：
Venugopal Balijepally;George Mangalaraj;K. Iyengar
研究文章VenuGopal Balijepally Prairie View A&M University vebalijepally@pvamu.edu George Mangalaraj西伊利诺伊大学G-Mangalaraj@wiu.edu Kishen Iyengar科罗拉多大学博尔德分校kishen.iyengar@colorado.edu聚类分析是一种强大的统计程序，用于提取数据和种群之间的自然配置。聚类分析具有在任何数据集中产生分组的看似无限的能力，具有超级技术的所有特征。然而，即使在数据中没有任何自然结构的情况下，该方法也会产生聚类，并且没有统计依据来拒绝数据中没有自然分组的零假设。因此，应用聚类分析的前提是健全的研究者判断和负责任的分析和报告。本文总结了一篇关于聚类分析在信息系统（IS）研究中的应用的反思性综述。在分析55个IS应用的基础上，讨论了聚类分析在使用中注意到的各种不足，并对今后的实践提出了建议。通过分析两个时间段的结果，突出了该技术应用的纵向趋势。
Research Article VenuGopal Balijepally Prairie View A&M University vebalijepally@pvamu.edu George Mangalaraj Western Illinois University G-Mangalaraj@wiu.edu Kishen Iyengar University of Colorado at Boulder kishen.iyengar@colorado.edu Cluster analysis is a powerful statistical procedure for extricating natural configurations among the data and the populations. Cluster analysis, with its seemingly limitless power to produce groupings in any dataset, has all the trappings of a super-technique. However, the method produces clusters even in the absence of any natural structure in the data, and has no statistical basis to reject the null hypothesis that there are no natural groupings in the data. Application of cluster analysis, therefore, presupposes sound researcher judgment and responsible analysis and reporting. This paper summarizes the results of a reflective review of the application of cluster analysis in Information Systems (IS) research published in major IS outlets. Based on the analysis of 55 IS applications of cluster analysis, various deficiencies noticed in its use are discussed along with suggestions for future practice. By analyzing the results over two time periods, longitudinal trends in the application of this technique are highlighted.