Why Simulate - AoM 2010


No comments posted yet


Slide 1

Slide 2

Slide 1

Why Simulate? Advantages and Pitfalls Sheen S. Levine, MIT c Prof. S. S. Levine

Slide 2

Simulate: Why and When Agent-based simulations are a way to investigate complex social and organizational phenomenon. Think about a simulation as a laboratory to test otherwise untestable proposition. Use it when data are rare or unobtainable, e.g., how cooperative tendencies of organizational members affect overall performance Use it when dealing with complexity, e.g., each member’s behavior is affected by the behavior of those around them Use it to explore possibilities, develop propositions for empirical testing, seeing which elements might matter

Slide 3

Three Common PitFalls

Slide 4

1: The Method is Cool. The Contribution is Cooler. The risk: the method becomes the substance Who will be interested in my research? Am I speaking their language? Which theory am I contributing to? Would theoreticians understand it? Would practitioners understand? “[a] paper must be clear to the modeler and it must also be possible for the non modeler to read the beginning, the end and understand the contribution to the question you are studying” R. Burton

Slide 5

2: Simple is Stunning The risk: an attempt to model the whole world Decide what’s in, and – importantly – what’s out How much can you achieve with few elements? “Scholars have described the world in various ways; the point is to simplify it”

Slide 6

3: Master the Mechanisms The risk: interesting results, but what generates them? Identify the mechanisms that generate your results Look for interactions, variance in effects “Social mechanisms are the ‘nuts and bolts’ processes by which cause and effect relationships in the social world come about” (Gross 2009:375)