What kind of prediction? Evaluating different facets of prediction in agent-based social simulation

ABSTRACT

Researchers have become increasingly interested in the potential use of agent-based modelling for the prediction of social phenomena, motivated by the desire, first, to further cement the method’s scientific status and, second, to participate in other scenarios, particularly in the aid of decision-making. This article contributes to the current discussion on prediction from the perspective of the disciplinary organisation of agent-based social simulation. It addresses conceptual and practical challenges pertaining to the community of practitioners, rather than individual instances of modelling. As such, it provides recommendations that invite both collective critical discussion and cooperation. The first two sections review conceptual challenges associated with the concept of prediction and its instantiation in the computational modelling of complex social phenomena. They identify methodological gaps and disagreements that warrant further analysis. The second two sections consider practical challenges related to the lack of a prediction framework that, on one hand, gives meaning and accommodates everyday prediction practices and, on the other hand, establishes more clearly the connection between prediction and other epistemic goals. This coordination at the practical level, it is claimed, might help to better position prediction with agent-based modelling within the larger social science’s methodological landscape.

KEYWORDS:

• agent-based modelling
• social simulation
• prediction
• explanation
• complex systems
• social science

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. This term refers to the positivist assumption according to which prediction and explanation are structurally equivalent and could only be differentiated on pragmatic grounds (Hempel & Oppenheim, 1948).

2. Instead of trying to characterise the elements involved in everyday practices of prediction, normative accounts seek to provide an a priori rational reconstruction of the logic of prediction.

3. The two features are sometimes conflated in the complexity literature (e.g. Hooker, 2011; Mitchell, 2009). In social science, however, they are often kept separate because human action is regularly considered a key cause of reinforcement or entrenchment.

4. Polhill et al. (2021) have recently claimed that wicked systems i.e. those where changes produced by interaction introduce qualitatively metaphysical novelty into the system e.g. major technological breakthroughs, pose a more serious problem for prediction than complex systems. This belief seems justified by the fact that this metaphysical novelty, first, is not easy to model and, second, implies that the system is more contextually bounded and path dependent.

5. At least, from a methodological point of view. Complementary metaphysical and epistemological questions remain about whether complex social systems are at all predictable and whether we have the adequate and sufficient means to produce predictions about them (e.g. Martin et al., 2016). Practitioners have acknowledged these questions, but have refrained from offering a forceful positive or negative answer.

6. They also neglect the fact that Hempel himself argued that the instantiation of the symmetry thesis in inductive systematisation, the type of reasoning more likely to accommodate social explanation, is problematic, for these arguments are susceptible to what he calls ‘ambiguity of statistical prediction’ i.e. two equally sound arguments with true premises could be construed statistically supporting both the occurrence and non-occurrence of the explanans (Hempel, 2001).

Additional information

Notes on contributors

David Anzola

David is an associate professor in the School of Management at Universidad del Rosario (Colombia) and a member of the Innovation Centre. His research interests are in the philosophy of computational social science, complexity, digital technologies, and social theory and metatheory. He is currently working on the methodology and social epistemology of agent-based social simulation.

César García-Díaz

Cesar Garcia-Diaz is an associate professor of business administration at the School of Economics and Business of the Pontificia Universidad Javeriana (Bogota, Colombia). Initially trained as an engineer, he holds a PhD in Economics and Business from the University of Groningen (Netherlands). His research focuses on modelling the dynamics of organisational, economic, and social systems. He is interested in the link between agents' micro-level rules of behavior, structural interdependence and macro-level outcomes in a variety of settings (e.g., organisational dynamics, industry evolution, agricultural markets). His email address is [email protected]