Real Research With Fake Data: A Tutorial on Conducting Computer Simulation for Research and Teaching

Abstract

Although many have recognized the potential value of computer simulations as a research tool, and there are many examples of research papers using simulation in several management journals, instruction on how to actually build and conduct computer simulations is absent from most doctoral education in management programs, as well as typically missing from research methods texts. The purpose of this paper is to provide an introductory tutorial on how to conduct computer simulations for both research and teaching. I walk the reader through the basic principles of simulation, showing both the math and techniques needed to create data based on either desired relationships among the simulated variables (i.e., based on correlations) or based on a desired model. The paper also introduces techniques to make data more “interesting”, including how to add skew or kurtosis, create multi-item measures with a desired level of unreliability, and how to incorporate mediated, moderated, and non-linear relationships. To facilitate the tutorial, the methods described in the paper are illustrated using Excel, Mplus, and R, so that a wide variety of people can begin to use this method and not just those who are experts in R or have a computer programming background. Supplemental files are provided that illustrate each example used in the paper. The goal of this paper is not to help make experts on simulation even better at their trade; rather, it is to open up to all readers the powerful potential of this research tool.