Computer Simulation in Social Sciences. A Logic of Enquiry.

With different fortunes and oscillating enthusiasm, computer simulation (CS hereafter) has supported theoretical investigation in managerial disciplines since the 60’s without really becoming part of mainstream approach to research (Harrison et al., 2007). In the attempt to further corroborate the role of CS in the repertoire of research strategies available to social scientists, the aim of the present book is to investigate the logic of inference that underpins simulation studies and to provide some hints for developing and testing models. A number of scholars in organization and management theory (OMT hereafter) have already produced appreciable efforts to build familiarity between CS and their community of practice. The efforts testify that an incentive exists to include CS among the legitimated panoply of methods that are employed by researchers to conduct research in OMT (Secchi and Neumann, 2016). To support researchers in using CS, for example, Davis, Eisenhardt and Bingham (2007) describe a roadmap that clarifies the steps necessary to design a CS study to develop theory. On similar veins, to captures the logic that underpins CS and to appreciate its potentials, avoiding looming pitfalls, Harrison et al. (2007) articulate a process model that guides researchers in using CS for theory development. In Harrison et al. (2007), an attempt is made to explain how CS support deductive and inductive inferences. In their point of view, CS is a third way to produce inferences that, on the hand, <<renders irrelevant the deductive problem of intractability>> and, on the other hand, <<partially overcomes the empirical problem of data availability>> (Harrison et al., 2007, p. 1230) that often afflicts inductive inferences. To push further this movement of increasing awareness of the potentials of CS, this book elaborates on the way different types of inferences may be generated when developing theory using CS. More precisely, the book makes four points. Firstly, it suggests that considering a third type of inference – abduction (Peirce, 1955a; Burks, 1964; Fann, 1970) – is fundamental to capture the value added by CS in theory development. To traduce the considerations contained in this study book into a usable theoretical frame for researchers, I propose two idealtypes of CS-based studies. To be sure, an idealtype crystallises what is essential about a phenomenon (Swedberg, 2005, p. 119) and may provide a safe point of departure to analyze and conceive of a simulation study. Secondly, the argument presented in the book clarifies than rather than employing a particular way of inference, a third way beside induction and deduction, CS employs the three type of inferences – abduction, deduction, induction – in different points of the research process and with different emphasis depending on the research question that animates the study. Thirdly, the book invites researchers to associate computer simulation and field research. To do so, the second chapter sketches a guideline to define a research strategy that integrates CS and field research. Finally, the book offers a few considerations and some specific techniques to address the issue of testing computer models that were generated to explore phenomena in the domain of the social sciences. The book is organizes as follows. In the first chapter, I address the logic underpinning computer simulation studies. I start by briefly defining CS. Then, I describe how and why scholars in social sciences came about to employ computer simulations for their research. Following, I explain how deductive and abductive inferences are the main output of CS-based theorization. Based on a selected small sample of articles in OMT that adopt CS-based research strategy, I build two idealtypes of CS-based studies. The studies subsumed to the first idealtype group emphasise the use of deductive inference whereas the studies that belong to the second idealtype more clearly adopt abductive inferences. I describe the two idealtypes grounding on the description of the sample of articles and I suggest how, typically, in most CS-based studies, deductive and abductive inferences interact. I conclude the first chapter by proposing how inductive inferences, as well, may emerge from a CS-based study. In the second chapter, I describe how computer simulation can be coupled with field-based research. The idea is to describe the logic for integrating three fundamentals steps: theorising from a field-study, theory formalization and theory development. The last step is obtained by building and nurturing a dialogue between synthetic behaviours, which are generated by the means of simulation experiments, and empirically observed behaviours. In the last chapter, I advance some principles for testing simulation models. In addition, I show typical testing techniques as used in System Dynamics. This latter is a discipline that has developed a robust procedure to tests computer models. A number of considerations close the book.