Avoiding costly mistakes in groups: The evolution of error management in collective decision making

Abstract

Individuals continuously have to balance the error costs of alternative decisions. A wealth of research has studied how single individuals navigate this, showing that individuals develop response biases to avoid the more costly error. We, however, know little about the dynamics in groups facing asymmetrical error costs and when social influence amplifies either safe or risky behavior. Here, we investigate this by modeling the decision process and information flow with a drift–diffusion model extended to the social domain. In the model individuals first gather independent personal information; they then enter a social phase in which they can either decide early based on personal information, or wait for additional social information. We combined the model with an evolutionary algorithm to derive adaptive behavior. We find that under asymmetric costs, individuals in large cooperative groups do not develop response biases because such biases amplify at the collective level, triggering false information cascades. Selfish individuals, however, undermine the group’s performance for their own benefit by developing higher response biases and waiting for more information. Our results have implications for our understanding of the social dynamics in groups facing asymmetrical errors costs, such as animal groups evading predation or police officers holding a suspect at gunpoint.

Beyond collective intelligence: Collective adaptation

“…by a frequency-based or model-based strategy; [ 168 ]). In addition, individuals can use diverse state-based strategies [ 40 ], including confidence in own judgements [ 169 ] and amount of personal information [ 170 ] to decide whether to use social information in the first place.…”

“…One example is Tump et al . 's [ 170 ] combination of drift-diffusion cognitive models of individual and social learning with an evolutionary process that selects parameters of drift-diffusion models that are best adapted to particular group structures and problems with different costs of errors. Another example is Cooney et al .…”

“…To decide which strategy to use in different social and problem contexts, individuals and groups use metacognitive cues about their own and others' past and likely future performance [247,253]. One such cue can be the speed of response of different group members: those who have better individual information are likely to react first, allowing the others to learn socially from them, which can be particularly advantageous when costs of errors are high or when the group situation is competitive [170]. Strategy choice can also be affected by cultural norms (e.g.…”

“…Because collective adaptation involves an interaction of cognition, social environments and problem structure, modelling it is likely to require connecting several frameworks, each useful for a particular component or time scale. One example is Tump et al's [170] combination of drift-diffusion cognitive models of individual and social learning with an evolutionary process that selects parameters of drift-diffusion models that are best adapted to particular group structures and problems with different costs of errors. Another example is Cooney et al's [292] work on modelling the evolution of sociality that combines epidemiological models, a replicator equation and adaptive dynamics models to study evolution on three different time scales.…”

A cognitive computational approach to social and collective decision-making

“…For example, experimental research on individual police officers in shooting simulators has shown that the decision to shoot is typically made faster than the decision not to shoot, explained by a bias in the start point (Pleskac et al, 2018). Such asymmetries in decision timing-which appear in various social contexts (Tump, Pleskac, et al, 2022)-are predicted to have consequences on which decisions (and potential biases) are amplified in a social context, because early-arriving social information typically exerts a stronger influence on the collective (Tump, Pleskac, et al, 2022;Tump, Wolf, et al, 2022). The prediction at the collective level is thus that having multiple police officers in a shooting simulator will increase the likelihood to shoot.…”