Simulation and Virtual Experimentation: Grounding with Empirical Data

Abstract

Herein, our purpose is to demonstrate (1) how simulation procedures can be developed and validated with existing empirical data and (2) how these procedures can be executed to conduct virtual experiments. To accomplish this purpose, we demonstrate how empirically collected data can inform simulation procedures to answer what-if research questions; the answers to which can, in turn, guide future empirical data collection. We discuss two examples to demonstrate the range of what-if questions that may be addressed via this approach. First, we provide guidance for developing simulation procedures that incorporate continuous data. At the end of the chapter, we describe how this approach was implemented in Kennedy et al.. In that research, we used available team-level, continuous, crosssectional data that had been collected via questionnaires to examine how project complexity impacts the curvilinear relationship between team communication and performance identified in Patrashkova-Vozdolska, McComb, Green, and Compton (2003). Virtual experiments were conducted by executing the simulation procedures under varying levels of project complexity to garner insights about the communication-performance relationship. Second, we demonstrate how to develop simulation procedures with discrete data. We then provide the example of this approach, from Kennedy and McComb, where we used transcribed and coded communication strings (i.e., discrete, longitudinal data) from a laboratory study to understand the relationship between team performance and when teams shift their conversations among different processes. Virtual experiments were conducted to ascertain what happens if certain process shifts occur earlier or later in the team's life cycle. Results of both studies inform theory about team communication and can be tested through laboratory and/or field experimentation (PsycINFO Database Record (c) 2017 APA, all rights reserved)