Can Confirmation Bias Improve Group Learning?

O’Connor, Cailin; Gabriel, Nathan

doi:10.31222/osf.io/dzych

Cited by 11 publications

(6 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our study suggests there is a similar mechanism at play with CARL, at least in fully-connected networks. In the social case, similar results have been obtained by Gabriel & O'Connor with an opiniondynamics model [20], and independently from resource scarcity. The authors argue that a moderate confirmation bias actually serves group learning, and could have developed in the social environment in which humans evolved.…”

Section: Discussionsupporting

confidence: 82%

“…But it was not obvious that a confirmation bias would also lead to poorer performance in groups of two or three agents. Similar results, however, have been obtained by Gabriel & O'Connor with an opinion-dynamics model, and independently from resource scarcity [20]. The authors implemented a "strong confirmation bias", and showed that, for small group sizes, strongest biases harm correct consensus.…”

Section: Discussionsupporting

confidence: 74%

“…In particular, confirmation has been linked to the emergence of problematic phenomena, such as echo chambers and polarization [18], [19]. A recent opinion-dynamics study, on the other hand, claims that a moderate amount of confirmation bias improves group learning, whereas a "strong" confirmation bias harms group consensus and leads to polarization [20]. In opinion dynamics, however, confirmation can be modelled in many different ways, and studies still lack a unified theoretical setup [18].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Moderate confirmation bias enhances collective decision-making in reinforcement-learning agents

Bergerot,

Barfuss,

Romanczuk

2023

Preprint

View full text Add to dashboard Cite

Humans tend to give more weight to information confirming their beliefs than to information that disconfirms them. Nevertheless, this apparent irrationality has been shown to improve individual decision-making under uncertainty. However, little is known about this bias’ impact on collective decision-making. Here, we investigate the conditions under which confirmation bias is beneficial or detrimental to collective decision-making. To do so, we develop a Collective Asymmetric Reinforcement Learning (CARL) model in which artificial agents observe others’ actions and rewards, and update this information asymmetrically. We use agent-based simulations to study how confirmation bias affects collective performance on a two-armed bandit task, and how resource scarcity, group size and bias strength modulate this effect. We find that a confirmation bias benefits group learning across a wide range of resource-scarcity conditions. Moreover, we discover that, past a critical bias strength, resource abundance favors the emergence of two different performance regimes, one of which is suboptimal. In addition, we find that this regime bifurcation comes with polarization in small groups of agents. Overall, our results suggest the existence of an optimal, moderate level of confirmation bias for collective decision-making.

show abstract

Section: Discussionsupporting

confidence: 82%

Section: Discussionsupporting

confidence: 74%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Moderate confirmation bias enhances collective decision-making in reinforcement-learning agents

Bergerot,

Barfuss,

Romanczuk

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The impact of confirmation bias in groups on group learning and performance remains an open question since we investigated confirmation bias and learning in two different experimental paradigms. It has been suggested that moderate confirmation bias may enhance individual 35 and group performance 36 when learning, and that intermittent social information improves collective problem-solving 37 . Understanding the conditions under which group decisions and associated confirmation biases contribute to accuracy and learning is an important avenue for future research on this topic.…”

Section: Discussionmentioning

confidence: 99%

Group Affiliation influences Information Processing in Collective Decision-Making

El Zein,

Lefebvre

2023

Preprint

View full text Add to dashboard Cite

Group decision-making and group affiliation are typically addressed as separate research topics. Our research combines these two areas to comprehend how groups impact the processing and updating of information during collective decisions. We utilized a minimal group paradigm to establish various levels of group affiliations (first-level in-group, second-level in-group, and outgroup). Participants engaged in majority vote group decisions within these three different groups, along with solo decisions, in two pre-registered online studies on Prolific. Our studies applied established experimental paradigms in human individual behavior to explore group decision-making. In the first study, participants (N=380) performed a learning task that required them to select gambles (among two options) with higher probabilities of winning. In the second study, participants (N=300) completed a perceptual task where they judged the direction of random dots twice, enabling us to assess how initial group decisions influenced information updating based on new information. We observed general effects of group decisions, irrespective of group affiliation: group vs. solo decisions reduced the impact of recent events on information processing (Experiment 1) and updating (Experiment 2). Regarding the effects of group affiliation, participants tended to rely more on in-group judgments compared to out-group judgments only when facing uncertainty. These findings contribute to our understanding of the mechanisms by which groups influence information and misinformation processing.

show abstract

“…It may seem obvious that in-group bias should create divisions between groups, inhibiting consensus formation. However, in-group bias may also speed up the convergence of opinions within in-groups, thus allowing opinions to spread further across the network before polarization can take hold [ 50 , 51 ]. This suggests that the structure of social interactions may moderate the effects of in-group bias.…”

Section: Introductionmentioning

confidence: 99%

Social identity bias and communication network clustering interact to shape patterns of opinion dynamics

Steiglechner,

Smaldino,

Moser

et al. 2023

J. R. Soc. Interface.

View full text Add to dashboard Cite

Social influence aligns people's opinions, but social identities and related in-group biases interfere with this alignment. For instance, the recent rise of young climate activists (e.g. ‘Fridays for Future’ or ‘Last Generation’) has highlighted the importance of generational identities in the climate change debate. It is unclear how social identities affect the emergence of opinion patterns, such as consensus or disagreement, in a society. Here, we present an agent-based model to explore this question. Agents communicate in a network and form opinions through social influence. The agents have fixed social identities which involve homophily in their interaction preferences and in-group bias in their perception of others. We find that the in-group bias has opposing effects depending on the network topology. The bias impedes consensus in highly random networks by promoting the formation of echo chambers within social identity groups. By contrast, the bias facilitates consensus in highly clustered networks by aligning dispersed in-group agents across the network and, thereby, preventing the formation of isolated echo chambers. Our model uncovers the mechanisms underpinning these opposing effects of the in-group bias and highlights the importance of the communication network topology for shaping opinion dynamics.

show abstract

Can Confirmation Bias Improve Group Learning?

Cited by 11 publications

References 41 publications

Moderate confirmation bias enhances collective decision-making in reinforcement-learning agents

Moderate confirmation bias enhances collective decision-making in reinforcement-learning agents

Group Affiliation influences Information Processing in Collective Decision-Making

Social identity bias and communication network clustering interact to shape patterns of opinion dynamics

Contact Info

Product

Resources

About