The simple rules of a complex world: William Rand and Roland Rust

The purpose of this commentary is to explain that it is not useful to unnecessarily complicate a model. Striving for realism for its own sake does not advance understanding; however, making sure that a model provides valid insights is a useful goal.

The authors advocate that a standard should exist based on whether experts in a field think that a particular mechanism is necessary for the model to achieve the goals of validity and sufficiency.

The authors find that critiques that do not offer a more valid alternative model do not necessarily advance the production of science.

Decision makers need to understand the assumptions and limitations of the models that they are using, but they should also be educated on the basic concepts of modeling literacy, and develop an understanding that all models are necessarily incomplete, as to make a model a perfect reflection of the real world would not provide insightful generalizations.

Although the original paper provides some additional cases that should be explored in understanding the diffusion of information, the authors extend this paper by providing a standard that explains when it is necessary to examine additional extensions and when the original (less complex) model is sufficient.