Summary: New research shows that people unconsciously adopt others’ tendencies toward prudence, impatience, or laziness.
Source: PLOS
People tend to unconsciously copy others’ tendencies for prudence, impatience, or laziness, new study in PLOS Computational Biology finds.
Behaviors described as “prudence,” “impatience,” or “laziness” are often considered stable personality traits that shape how individuals trade off risk, delay, and effort. Recent research, however, indicates these attitudes are more flexible: people’s preferences for risk, waiting, and effort shift in the direction of the attitudes they observe in others. This phenomenon—attitude alignment—has implications for social influence, decision making, and how groups form shared norms around effort and risk-taking.
Researchers Jean Daunizeau and Marie Devaine from INSERM in Paris combined mathematical modeling with cognitive psychology experiments to investigate how people learn about and from others’ hidden attitudes. The study recruited 56 participants who completed a series of decision-making tasks that involved choices between options varying in risk, delay, or required effort. Participants completed these tasks both before and after observing decisions made by fictional peers. Unbeknownst to participants, those peers were simulated by calibrated artificial intelligence agents programmed to display consistently prudent, patient, or lazy tendencies.
The results reveal two robust cognitive effects. First, participants displayed a false-consensus bias: they tended to assume that other people’s attitudes resembled their own, even without evidence. Second, participants showed a social influence bias: after observing others’ choices, their own attitudes tended to shift toward those they had observed. Intriguingly, these two biases interact in a non-linear way. Social influence increases with false-consensus when false-consensus is small, but as false-consensus becomes large it actually reduces social influence. Participants were largely unaware of these biases, suggesting the shifts operate at an implicit level.
Beyond the behavioral findings, the authors developed mathematical simulations showing that both the false-consensus bias and the social influence bias—and the specific way they interact—can be derived from a single computational learning mechanism. This mechanism allows people to infer other individuals’ hidden attitudes and then adjust their own uncertain beliefs about the best way to behave in comparable decision-making contexts. In other words, attitude alignment appears to reflect an adaptive, information-driven process for learning about the social environment rather than a mere drive to conform for social approval.
The study’s combined use of formal decision theory and experimental data supports a computational account of social cognition: by modeling how people infer hidden attitudes and integrate observed behaviors into their own decision-making, researchers can explain puzzling social biases in a precise, testable way. These results bridge theoretical models with empirical behavior and advance understanding of how social information shapes individual preferences for risk, delay, and effort.
The authors note potential clinical relevance and are extending this work to examine whether attitude alignment differs in populations with neuropsychiatric conditions, including autism spectrum disorder and schizophrenia. Differences in how people learn from and about others’ attitudes could influence social functioning and decision-making in these conditions.
Funding: This research was supported by the IHU-A-ICM and the French Ministère de l’Enseignement Supérieur et de la Recherche. The funders did not influence study design, data collection and analysis, the decision to publish, or manuscript preparation.
Competing interests: The authors have declared that no competing interests exist.
Source: Jean Daunizeau, PLOS. Original research: Devaine and Daunizeau, “Learning about and from others’ prudence, impatience or laziness: The computational bases of attitude alignment,” PLOS Computational Biology, published March 30, 2017 (doi:10.1371/journal.pcbi.1005422).
Learning about and from others’ prudence, impatience or laziness: The computational bases of attitude alignment
Individual attitudes toward costs such as risk, delay, and effort strongly influence goal-directed behavior and explain important differences between people. Learning about other people’s prudence, impatience, or laziness is therefore critical in social interactions, but the adaptive value of these attitudes in a given environment is often uncertain. The brain may therefore be tuned to gather information about how such trade-offs should be resolved, with observations of others’ behavior informing one’s own uncertain beliefs about optimal choices in similar decision contexts. In turn, the extent to which one learns from others depends on one’s ability to infer their hidden attitudes. From first principles of optimal inference, we derive computational predictions for how people learn about others’ attitudes. We predict two observable biases: (i) a tendency to overestimate similarity with others (false-consensus bias), and (ii) a tendency to align one’s attitudes with those observed in others (social influence bias). Our theory further predicts a non-trivial interaction between these biases. We validate these predictions experimentally by measuring participants’ attitudes both before and after they observed choices made by calibrated artificial agents acting as surrogate individuals.