Computational Models of Social Behavior

Prof. Braumoeller made each [class meeting] more interesting than the previous. He has a good sense of humor, but is still serious enough that all the work gets finished. — participant from the U.S.

This course provides participants with an overview of models of social behavior and an introduction to the computational tools used to implement them. We begin with simple dynamic models of behavior, such as the Lotka-Volterra predator-prey model and Richardson's arms race, and work through more nuanced categories of models, such as agent-based, network, feedback, and contagion/diffusion models. We touch on behavioral models captured in common statistical models (e.g., models of selection and endogeneity) and end with a discussion of how to engage in multi-model thinking. Throughout the course, we discuss canonical social science models, with an emphasis on simulation and application.

This course is a standalone course that does not require any prior background or knowledge in programming.

Dates

This course was offered in 2021.

Instructor

Bear F. Braumoeller, Ohio State University

Detailed Description

The core knowledge of any discipline is made up of theoretical models. In the natural sciences, we have classical thermodynamics in chemistry and the Bohr model of the atom in physics. Similarly, political science has the median voter theorem, sociology Granovetter's threshold model of collective behavior, ecology the Daisyworld model, and the Prisoner’s Dilemma is prevalent throughout the social sciences. In the field of international relations, theoretical models like the security dilemma and offense-defense theory are pervasive in the languages of both the academic and policy communities. Theoretical models create a shared, interrelated set of ideas that help us understand the world.

While explanations in the social sciences generally involve both theories and tests of those theories, we tend to teach the testing much more than we teach the theorizing. Because the most basic models used for statistical testing are concrete and widely used, we have increasingly adopted them as theoretical models, even though they often do not reflect a very coherent theory. When they do not reflect a concrete theory, their results are not particularly meaningful. Having a diverse catalog of theoretical models on hand can serve as a counterweight to generic empirical models and help people to think about social science explanation in new ways. Showing the details of how these models are constructed gives social scientists the building blocks they need to tailor new theoretical models to their own needs.

This course is an introduction to and survey of theoretical models in the social sciences. We will survey a wide range of models, including network models, agent-based models, system-dynamics models, feedback, contagion, and diffusion. In the course of doing so, participants will be familiarized with some of the most canonical behavioral models in the social sciences. The goal of the course is to give students the ability to translate a wide range of theoretical intuitions into formal computational models. For that, we will use simulation as a tool to codify the assumptions of models and interpret the often-surprising implications of even very simple interactions. By the end of the course, participants will have both a variety of theoretical lenses through which to view phenomena of interest and a basic working knowledge of the tools needed to create their own original models.

Prerequisites

There are no prerequisites for this course. However, some familiarity with programming languages, such as Logo and/or R, would be helpful.

Requirements

Participants are expected to have access to an internet-connected computer. Access to data, temporary licenses for the course software, and installation support will be provided by the Methods School.

Core Readings

Page, Scott E. 2018. The Model Thinker: What You Need to Know to Make Data Work for You. 1^st edition. New York, NY: Basic Books.

Wilensky, Uri, and William Rand. 2015. An Introduction to Agent-Based Modeling: Modeling Natural, Social, and Engineered Complex Systems with NetLogo. Cambridge, MA: MIT Press.