People often have to decide if a claim is true (for example, that climate change is caused by human activity) even if they do not have the necessary knowledge or experience. It is common for people to turn to others to determine if a claim is true. When multiple sources agree, it can be a strong indicator of who to trust and what to believe (Mercier & Morin, 2019). In cognitive science, there is debate about what makes a “consensus” and how it should influence our decisions. Some models suggest that agreement among independent sources is more reliable than among sources who have communicated with each other (Whalen, Griffiths, & Buchsbaum, 2018; Bovens & Hartmann, 2004; Dietrich & Spiekermann, 2013). Recent empirical and modelling work (Pilditch, Hahn, Fenton, & Lagnado, 2020), however, suggests that having sources that are connected can sometimes provide more reliable information than having sources that are independent. This symposium brings together researchers from a range of disciplines (cognitive modelling, social network modelling, cognitive psychology, philosophy) to examine what makes a consensus persuasive, and it should guide our judgments and inferences.

Information initially presented as a likely cause of an eventbut turns out to be incorrect can affect people’s reasoningdespite being clearly corrected – a phenomenon known as thecontinued influence effect of misinformation. The presentwork extends previous findings showing that misinformationthat implies a likely cause of an adverse outcome is moreresistant to correction than misinformation that explicitlystates a likely cause. Participants either read a reportdescribing a fire or a crash. The difference between impliedand explicitly stated misinformation was replicated with thefire scenario, which has been commonly used in continuedinfluence research. There was little evidence of a continuedinfluence of misinformation for the (novel) crash scenario.The results constrain the generalizability of the continuedinfluence effect and suggest that corrections that clearlyinvalidate initial misinformation can be effective.

When and how do people detect and adjust for sample bias? The present work examines how people use a small, potentially biased sample to reason about the population from which the sample was drawn. We developed a task to explore when and how people generalize from selectively sampled evidence and which cues learners use to infer the presence of selection bias and adjust their responses accordingly. Participants sampled data that followed several different distributions (e.g., bimodal) and constructed their own frequency distribution of a larger “population”. Participants “filled in” missing data when sample distribution observations were sparse or were inconsistent with priors about distribution shape. Human responses were compared with several computational model predictions. An averaging model captured response curves better than a normative Bayesian approach (for some distributions). Results suggest that people draw on both prior distributional expectations and observed sample distributions when generalizing from potentially biased samples.

Witness credibility is important for establishing testimonial value.The story model posits that people construct narratives fromevidence but does not explain how credibility is assessed. Formalapproaches use Bayesian networks (BN) to represent legalevidence. Recent empirical work suggests people might alsoreason using qualitative causal networks. In two studies,participants read a realistic trial transcript and judge guilt andwitness credibility. Study 1 varied testimonial consistency anddefendant character. Guilt and credibility assessments wereaffected by consistency but not prior convictions. Study 2constructed a BN to represent consistency issues. Individualparameter estimates were elicited for the corresponding BN tocompute posterior predictions for guilt and credibility. The BNprovided a good model for overall and individual guilt andcredibility ratings. These results suggest people construct causalmodels of the evidence and consider witness credibility. The BNapproach is a promising direction for future research in legalreasoning.

Misinformation, and its impact on society, has become anincreasingly topical field of study of late. A body of literatureexists that suggests misinformation can retain an influenceover beliefs despite subsequent retraction, known as theContinued Influence Effect (CIE). Researchers have arguedthis to be irrational. However, we show using a Bayesianformalism why this argument is overly assumptive, pointingto (previously overlooked) considerations of reliability of, anddependence between, misinforming and retracting sources.We demonstrate that lay reasoners intuitively endorseassumptions that demarcate CIE as a rational process, basedon the fact misinformation precedes its retraction. Moreover,despite using established CIE materials, we further upturn theapplecart by finding participants show CIE, and appropriatelypenalize the reliabilities of contradicting sources.

The studies apply a Bayesian source credibility model to a legal setting to test epistemic influence of witnesstestimonies. The model amalgamates perceived witness trustworthiness and access to accurate information as independentelements that describe and predict the impact of the testimony of that particular witness.Across two studies, the model enjoys a good fit with observed posterior ratings of the likelihood of guilt (study 1: R2 = .867,study 2: R2 = .701). Study 1 (n = 101) employs different witness types and reports whilst study 2 (n = 102) employs differentwitness types, access to accurate information, and reports.The studies suggest the applicability of a Bayesian source credibility model in a legal setting to account for the impactof different witness types. We show that participants are sensitive to the type of witness and that different witnesses have apredictable impact on the perception of the testimony.

When reasoning about a claim, it makes sense to be more persuaded if lots of other people agree. But, there are many factors that make weighing the evidence behind a consensus complicated. For example, a consensus might be more or less informative depending on the type of claim, or whether each consensus member formed their opinions independently. These factors might also influence people differently depending on their own assumptions or preferences. In this study we used a mock social media paradigm to assess how persuaded people were by two factors: the presence of consensus (no consensus vs. consensus), and source independence (a consensus based on independent information sources vs. a consensus formed off shared, dependent sources). We varied these factors at both the group and individual level. At the group level, we assessed a third factor: whether people were influenced by the type of claim being reasoned about (we assessed 60 different claims divided into 4 categories). Almost everyone was more persuaded by consensus trials compared to no consensus trials. However, the strength of this effect was credibly stronger if the claim was likely to have a ground truth. We found that around one third of participants were sensitive to source independence. Of these, three quarters were more persuaded by a consensus based on independent sources, but the quarter who were more persuaded by dependent sources were persuaded just as strongly.

Is the nested sets approach to improving accuracy on Bayesian word problems simply a way of prompting a natural frequencies solution, as its critics claim? Conversely, is it in fact, as its advocates claim, a more fundamental explanation of why the natural frequency approach itself works? Following recent calls, we use a process-focused approach to contribute to answering these long-debated questions. We also argue for a third, pragmatic way of looking at these two approaches and argue that they reveal different truths about human Bayesian reasoning. Using a think aloud methodology we show that while the nested sets approach does appear in part to work via the mechanisms theorised by advocates (by encouraging a nested sets representation), it also encourages conversion of the problem to frequencies, as its critics claim. The ramifications of these findings, as well as ways to further enhance the nested sets approach and train individuals to deal with standard probability problems are discussed.