Computational Social Science for Sustainability

Foreword

This document is being regularly compiled, developed, and edited. Consider these course notes for Computational Social Science for Sustainability. Please send any feedback by email.

Last updated Introduction bullet points on 2024-12-17.

1. Introduction

  • A major challenge in promoting sustainability is to convince enough people to engage in some sustainable behavior, so that some sustainable goal (carbon-free electricity, for example) can be achieved. There exist psychological strategies for convincing people that some behavior is beneficial for an individual or an organization, such as the benefits of no-till and crop rotations in agriculture, which seem to outweigh the possible costs. Nudging and other popular ideas about promoting beneficial behaviors focus on the individual. The goal of commercial advertising is to convince people that certain products and services are valuable to buy and use. However, even if people learn perfectly what’s good and bad from others, computer models of social behavior suggest that a beneficial behavior will spread throughout a simulated population only a fraction of the time. That fraction, though, can be increased or decreased by manipulating the social networks that structure interpersonal interaction. Similarly, social network structure and the randomness of social interaction leads us to conclude that polarization is not inevitable in a group, given some initial set of group opinions.

  • To learn these lessons and to improve rigor in social and behavioral science, one must create models of interpersonal interaction. In computational models that we will study, social phenomena really emerge from simulated interpersonal interactions. A model is a simplified description of real-world phenomena. A computational model just uses a computer to solve equations, simulate, or statistically analyze some _phenomenon of interest. Social science models are most useful with a minimal number of empirically-motivated assumptions that still generate emergent social phenomena. The emergent social phenomena here are innovation diffusion, opinion and norm change, and widespread economic cooperation. These phenomena are emergent because they emerge from a series interactions between people, sometimes repeatedly between the same people, over time.

  • One of the learning goals of this course is to develop a theoretical understanding of how social phenomena emerge from social interactions over time. Social science theory is important because it identifies basic cognitive and social processes and behaviors and how they produce emergent social phenomena of interest (Kauffman 1970; Craver 2006). Interpersonal interaction can take several forms, but we focus on three categories: (1) social learning, (2) social influence, and (3) cooperation. Social networks constrain who interacts with whom in the real world, and so too in our computational models.

2. Computational models in R for sustainability social science

  • Review and expand on what is a model and why it’s useful.
  • A paragraph on computational modeling as a special form of mathematical modeling
  • Computation enables the rapid generation of simulated data given some interaction processes and social network structure.
  • Assumptions about social behavior and cognition become mechanisms in the model. Mechanistic modeling helps us avoid both misinterpretation of verbal models (Turner and Smaldino 2022) and overgeneralization in statistical models that do not sufficiently account for sources of variance in behavioral experiments, for example (Yarkoni 2022). Specifically, this approach identifies the components of a system and their capacities (Cartwright 1989), e.g. for memory or abilities. Given these capacities and

Programming in R

Object-oriented programming…in R!?

Social networks

Social networks are fundamental to modeling how behavior spreads in populations. As we will see ourselves in the next section, even if people learn perfectly well from one another with no social friction, they are not guaranteed to spread to every person in a simulated population.

3. Predicting the diffusion of sustainable innovations through social learning

4. Discourse as a force that drives opinion change and polarization

5. Promoting cooperation and institutions to achieve sustainability goals

6. A review and way forward

References

Cartwright, Nancy. 1989. Nature’s Capacities and their Measurment. Oxford: Oxford University Press.
Craver, Carl F. 2006. When mechanistic models explain.” Synthese 153 (3): 355–76. https://doi.org/10.1007/s11229-006-9097-x.
Kauffman, Stuart A. 1970. Articulation of Parts Explanation in Biology and the Rational Search for Them.” In PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association, 257–72.
Turner, Matthew A., and Paul E. Smaldino. 2022. Mechanistic Modeling for the Masses - commentary on Yarkoni, "The generalizability crisis".” Behavioral and Brain Sciences 45 (E33). https://doi.org/10.1017/S0140525X2100039X.
Yarkoni, Tal. 2022. The generalizability crisis.” Behavioral and Brain Sciences 45 (e1): 1–78. https://doi.org/10.1017/S0140525X20001685.