Integrating behavioral experimental findings into dynamical models to inform social change interventions

A team of researchers from the University of Zurich and ETH Zurich has published a significant paper in Nature Human Behaviour that bridges a major gap between two scientific disciplines. The study, led by Radu Tanase, René Algesheimer, and Manuel S. Mariani, integrates discrete choice modeling—a staple of economics and marketing used to understand individual decisions—with complex contagion theory from network science, which models how behaviors spread through social connections. This hybrid approach creates a method to estimate individual-level thresholds for adopting a new product or behavior, a variable previously difficult to quantify at scale.

The researchers validated their model through two controlled choice experiments, demonstrating its predictive power. When they fed these estimated individual thresholds into large-scale computational simulations, they made a critical discovery: state-of-the-art 'seeding' policies—strategies that target influential individuals to kickstart adoption—can be suboptimal. These network-based strategies, which focus solely on social connectivity, can be up to 20% less effective because they neglect the nuanced psychological and contextual drivers of individual choice captured by discrete choice models. The proposed experimental method corrects this by informing seeding algorithms with real behavioral data, leading to more efficient and effective interventions for promoting social good, public health campaigns, or sustainable technologies.