Ridderprint

Discussion 145 7.3.1 Fragmented People Analytics Research Although not specifically substantiated in any of the chapters of this dissertation, I discovered that scientific research on people analytics and data-driven HRM has developed into multiple, mostly isolated streams of literature. The next section discusses these streams one by one, before demonstrating the ambiguity this fragmentation creates. 7.3.1.1 Conventional HRM research In general, most quantitative HRM research provides some form of evidence for the optimal management and organization of employees, thereby at least partially fitting the common definition of people analytics (e.g., Marler & Boudreau, 2017). However, some HRM studies appears more closely related to people analytics than others. Wright and Boswell (2002) divided conventional HRM research into four subcategories, based on the level of analysis (organization vs. individual) and the number of HRM practices considered in the study (multiple vs. single). In light of people analytics, particularly the level of analysis seems an important research characteristic. The level of analysis determines the granularity of the insights a research project produces. HRM research on the organizational level (e.g., research on strategic HRM, industrial relations, high-performance work systems, and isolated functions in Wright & Boswell, 2002) has explained howHRM implementation affects employees’ behaviors and cognitions and, in turn, their well-being and performance (e.g., Harter et al., 2002; Jiang et al., 2012; Van de Voorde, Paauwe, & Van Veldhoven, 2010). Such macro-level insights provide a general basis of evidence for best practices and are very relevant in the scoping phase of any people analytics project (e.g., Chapter 5). However, they infrequently provide direct evidence for local, contextual impact of HRM activities. HRM research aimed at the micro-level (e.g., research on psychological contract, employment relationship, functional HRM, and I/O psychology inWright & Boswell, 2002) may establish this local impact when large samples of a single organizational population are included. These studies can provide the local, evidence-based insights that hold direct value in HRM decision-making in practice (e.g., Chapter 6). Hence, particularly conventional micro-level HRM research closely fits the definition of people analytics. However, even for micro-level HRM research, the purpose will often starkly differ from that of people analytics. As discussed in Chapter 1, by tradition, the statistical modeling process will often be a testament to this. In sum, conventional HRM research only partially fits the common definitions of applied people analytics and some studies more than others. 7.3.1.2 People Analytics Literature Second, scholars in scientific and semi-scientific journals have discussed specifically people analytics (Marler & Boudreau, 2017) and big data initiatives in the HRM domain (e.g., McAbee, Landis, & Burke, 2017; LaValle, Lesser, Shockley, Hopkins, & Kruschwitz, 2011; Chapter 2). These studies can be considered the mainstream scientific people analytics literature as a search for “people analytics” on Web of Science or Google Scholar will result in these studies (see Marler & Boudreau, 2017). However, the focus of this stream of research seems more meta , discussing the what and how of people analytics rather than its application or empirical evidence of its value. For instance, a recurring