Ridderprint

Discussion 157 7.4 Implications 7.4.1 Implications for HRM Research While this dissertation was largely problem-, practice-, and data-driven, and not centered on a specifically theoretical framework, there are several implications for HRM research and the way science is conducted therein. 7.4.1.1 Collaborations First, there are many opportunities for collaborations between science and practice under the label of people analytics. Such collaborations have the potential to benefit the scientific field of HRM, the responsible scholars, the organizations and their employees, and even society as a whole. Contemporary organizations have vast, longitudinal datasets containing information about their employees and the way they are managed. More often than not, however, they lack the expertise to leverage the value of such data (Van der Togt & Rasmussen, 2017). Fortunately, scholars can help to identify a natural experiment in a training context (Van der Laken, 2017), to prevent the loss of information in case of multi- level issues (Chapter 3), to identify opportunities and limitations of social media for employees selection (Roth et al., 2016), to match applicants and vacancies via text mining (Kobayashi et al., 2017), and to explore employee retention as a survival problem (Chapter 6). In this sense, innovative collaborations between organizations and scholars can assure that the vast amounts of data contemporary organization collect produce knowledge and insights. With a balanced approach, people analytics projects may materialize benefits for the organization (more effective HRM), its employees (well-being and productivity), the scholars (data and publications), and society (innovation, labor participation, well-being, and productivity). 7.4.1.2 Alternative Research Second, I want to resonate the need for a divergence from purely explanatory studies. In contemporary management research, the balance between (scientific) rigor and (practical) relevance is off, highly in favor of the rigor dimension (Van Aken, 2004, p. 223). While emphasis on local problem-solving has been present in the form of Action Research (Eden & Huckham, 1996), in collaborative research (Rynes et al., 2001), and in mode 2 knowledge production (Gibbons et al., 1994), the last decades of management research have focused on explanatory, theory-driven research (Van Aken, 2004). People analytics has the potential to reintroduce a focus on practical relevance. Similarly, Yarkoni and Westfall (2017) argue that a redistribution of energies over explanatory and predictive studies may benefit psychological science and bridge the gap between scientific study and real-world application (p. 1114-1118). Currently, there seems to be a “ lack of understanding […] of the difference between building sound explanatory models versus creating powerful predictive models, as well as confusing explanatory power with predictive power ” (Shmueli, 2010, p. 289). While it may not be directly obvious, predictive modeling actually serves many valuable scientific functions: it helps to explore causal mechanisms in contemporary, large datasets; develop and optimize new measurement instruments; test, compare, and improve current theories;