Although much as been written about “big data” and “big data analytics”1 and how these will revolutionize how decisions are made, I have a sense of unease and feel the need to not only question the veracity of this claim but also question the implications of what “revolutionizing how decisions are made” means. Big data analytics and its sister technology “machine learning” are touted as the next wave of technological advancement which will obviate the need for humans to deal with repetitive processes (they will be automated through robotic process automation—RPA) while traditional (and non-traditional) decision making will be transformed. Maybe. But such pronouncements sound suspiciously like prior technological waves which were also promoted as “revolutionary” and “transformative.” In this essay, I wish to explore some of the underlying assumptions associated with this inexorable drive to base decision making on big data and big data analytics. In my analysis, I will attempt to highlight some of the fallacies and misconceptions that lie behind big data and how these assumptions are dysfunctional. In particular, I want to explore how these dysfunctions might manifest themselves when it comes to practice. In doing so, I want to tie the hype associated with big data to the Information Systems (IS) academic field’s focus on—dare I say worship of—“theory.” Interestingly, my objection is not so much that big data does not generate theory but rather that both theory—as it is widely conceived by the academic community—and big data do not yield understanding/insight. They generate “something”—in the case of theory, supposed relationships between variables and in the case of big data, the behaviors of large populations—but this is not understanding. It is my belief that we have mistakenly adopted the view that knowledge is the same as understanding. Knowledge may be the product of our research, but this is not the same as understanding, which is how to use that knowledge in practice. It is something I have written about in my “Against Theory” paper (Hirschheim, 2019). This is a key notion which I will explore later in the essay. My contention is that academics continue to make the same mistakes, continue to embrace erroneous assumptions about what they should be doing with their research, what the products of the research should be, and how we can help drive practice. Big data is just the latest in a series of waves of technological advancements which will supposedly change what we do and how we do it. But will it? And if it will, what are its likely implications? I am hopeful this essay, which came about after reading Gary Smith’s (in press) insightful “Data Mining Fool’s Gold” paper, will help bring these issues to the fore and possibly allow for a meaningful dialogue to take place.

中文翻译：

---

对理解的攻击：大数据和理论如何使我们误入歧途：对加里·史密斯的数据挖掘愚人金的评论

尽管已经写了很多关于“大数据”和“大数据分析”1 以及它们将如何彻底改变决策方式的文章，但我有一种不安的感觉，觉得有必要不仅质疑这种说法的真实性，而且质疑这种说法的真实性。 “彻底改变决策方式”的含义。大数据分析及其姊妹技术“机器学习”被吹捧为下一波技术进步，它将消除人类处理重复过程的需要（它们将通过机器人过程自动化 - RPA 实现自动化），而传统的（和非传统的）决策将被改变。可能是。但这样的声明听起来很像以前的技术浪潮，也被宣传为“革命性的”和“变革性的”。在这篇论文中，我希望探讨一些与这种将决策建立在大数据和大数据分析基础上的无情驱动相关的基本假设。在我的分析中，我将尝试强调大数据背后的一些谬误和误解，以及这些假设是如何发挥作用的。特别是，我想探讨这些功能障碍在实践中如何表现出来。在此过程中，我想将与大数据相关的炒作与信息系统 (IS) 学术领域对“理论”的关注——我敢说是对“理论”的崇拜联系起来。有趣的是，我的反对意见与其说是大数据不会产生理论，不如说是理论——正如学术界广泛设想的那样——和大数据都不会产生理解/洞察力。它们产生“某物”——就理论而言，假设变量之间的关系，在大数据的情况下，大量人群的行为——但这不是理解。我认为我们错误地采纳了知识与理解相同的观点。知识可能是我们研究的产物，但这与理解不同，理解是如何在实践中使用这些知识。这是我在“反对理论”论文（Hirschheim，2019 年）中所写的内容。这是我将在本文后面探讨的一个关键概念。我的论点是，学者们继续犯同样的错误，继续接受关于他们应该在研究中做什么、研究的产品应该是什么以及我们如何帮助推动实践的错误假设。大数据只是一系列技术进步浪潮中的最新一波，据称它将改变我们的工作方式和方式。但会吗？如果会，它可能的影响是什么？我希望这篇文章是在阅读 Gary Smith（出版中）富有洞察力的“Data Mining Fool's Gold”论文后发表的，它将有助于突出这些问题，并可能允许进行有意义的对话。