Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Jun 15;29(4):e70237. doi: 10.1111/desc.70237

How to Increase Children's and Adults’ Interest in Learning From Disagreement

Kirsten H Blakey 1,✉, Samuel Ronfard 1
PMCID: PMC13267442  PMID: 42294609

ABSTRACT

We often avoid talking to people who disagree with us, missing potentially valuable learning opportunities. Across four preregistered studies, we examined children's (6–11 years) and adults’ interest in learning from disagreement and whether it can be increased. In Study 1, we confirmed that children (6–11 years, N = 99, Canada) and adults (N = 100, USA and Canada) preferred to learn from those who agreed rather than disagreed with them across belief domains. In contrast, when asked what others should do, children (7–11 years, N = 83) and adults (N = 100) more often endorsed disagreement to increase learning, revealing a self‐other gap. Studies 3 and 4 tested two interventions aimed at increasing interest in learning from disagreement. In Study 3, explicitly teaching participants (96 children, 6–11 years; 100 adults) about the benefits of disagreement for learning reduced agreement seeking, particularly among older children and adults. In Study 4, asking participants (31 children, 10–11 years; 72 adults) to consider what others should do before making their own choices also reduced agreement seeking. To learn from disagreement, we cannot avoid it. We provide two means of increasing children's and adults’ interest in learning from disagreement by making epistemic benefits more salient.

Summary

  • Children and adults prefer learning from people who agree with them, despite recommending that others seek disagreement.

  • Explicitly teaching about the benefits of disagreement increased interest in seeking out opposing viewpoints.

  • Considering what others should do before making their own choice, increased participants’ engagement with disagreement.

  • Self‐distancing through third‐person reasoning helped both adults and children aged 10–11 prioritize epistemic benefits over social costs.

Keywords: belief, conflicting perspectives, disagreement, social learning

1. Introduction

We live in a society where disagreement is common, yet we rarely seek it out. Though disagreement can be challenging, both epistemically and socially, it offers valuable opportunities to reflect on our knowledge and beliefs. Gaining new perspectives or encountering unconsidered arguments can lead to the realization that our evidence is limited or biased, prompting us to seek additional information and update our beliefs. Despite an emerging understanding of these benefits when reasoning about other people (Ransom et al. 2025), children and adults tend to interact with, and seek information from, like‐minded individuals who share their norms, beliefs, and affiliations rather than engage with people who disagree with them (Blakey et al. 2021; Chalik et al. 2022; Hart et al. 2009; Mercier and Sperber 2011; Nickerson 1998; Smaldino and Velilla 2025; Stanovich et al. 2013). One reason for this discrepancy may be that when reasoning about disagreement for others, people naturally adopt a more reflective stance that makes epistemic benefits easier to see (Grossmann et al. 2021). When deciding for themselves, however, the interpersonal challenges of engaging with a disagreeing other may become more psychologically immediate, overshadowing those same benefits. We test this hypothesis across four studies. We first replicate the self‐other gap in seeking out disagreeing others, then test two interventions designed to increase the salience of epistemic benefits to see whether doing so shifts children's and adults' willingness to engage with someone who disagrees with them.

Recognizing the value of disagreement requires a sophisticated understanding of when and why people hold different views. This understanding is acquired across childhood, yet even by age 5, children hold nuanced views about when people are likely to disagree and about when disagreement is appropriate or expected (Banerjee et al. 2007; Danniels and Perlman 2021; Heiphetz et al. 2013; Wainryb et al. 2004). For example, they judge disagreements as more acceptable for ambiguous situations or matters of personal preference, than for moral questions. Five‐year‐olds also differentiate contexts where only one person can be correct from those where multiple perspectives may be equally valid, placing more weight on factual testimony than on personal opinions. That is, they understand that evidence constrains disagreement: contradictory claims cannot both be correct when there is decisive evidence (Confer et al. 2024; Yang et al. 2023). Despite this understanding, the ability to suspend judgments when faced with inconclusive evidence develops gradually between ages 5 and 9. As they get older, children are increasingly likely to maintain their initial belief when their own evidence is stronger, adopt another's belief when their evidence is weaker, and acknowledge that either of two conflicting claims could be valid when evidence is equally strong or absent (Langenhoff et al. 2023; Yang et al. 2023). It is also later in childhood—by 7 to 8 years—that children begin acknowledging that an ambiguous phenomenon can be interpreted differently by different people (Carpendale and Chandler 1996), an insight that helps them identify reasons behind individuals’ opposing beliefs and resolve disagreements by integrating alternative perspectives (Amemiya et al. 2021).

Yet even children who understand when disagreement is acceptable may not appreciate what can be learned from it. Disagreement has significant benefits, including reducing overconfidence and encouraging exploration in children aged 4 to 6 years (Langenhoff et al. 2024), and promoting critical thinking in adolescents and adults (Kuhn 2019; Mercier and Sperber 2011). However, there has been limited focus on children's reasoning about the epistemic value of disagreement. We know little about when and how children begin to recognize these benefits, or what factors shape their interest in engaging with disagreement. Recent work has begun to examine children's and adults’ reasoning about what others should do when navigating disagreement across belief domains (Ransom et al. 2025). Participants were asked to choose whether a protagonist should talk to a person who agrees or disagrees with them across four belief domains. Adults chose the person with a different view for ambiguous situations and personal preferences significantly above chance. Children aged 4 to 8 did not, but did choose the person with a different view more often in these domains than they did when the topic concerned factual knowledge or moral questions. While younger and older children endorsed learning from disagreement at similar rates, only children aged 7 to 8 years were able to articulate why doing so is helpful, recognizing that disagreement reveals the limits of our knowledge.

Previous research asked children to reason about what someone else should do in a third‐person context. Whether this reasoning extends to first‐person contexts remains an open question, particularly when the same paradigm is used and children's own beliefs are directly at stake. On the one hand, a large literature on children and adults has found that they tend to avoid information that contradicts their beliefs, preferring to seek confirming over conflicting information (Blakey et al. 2021; Chalik et al. 2022; Hart et al. 2009; Mercier and Sperber 2011; Nickerson 1998; Özkan and Ronfard 2026; Santhanagopalan et al. 2025; Smaldino and Velilla 2025), suggesting that they might show less interest in engaging with disagreement when reasoning about themselves than when indicating what others should do. On the other hand, children's responses in first‐ and third‐person contexts sometimes align, particularly when children are reasoning about evidence‐based beliefs and beliefs they are uncertain about. For instance, when facing disagreement, children aged 6 are more likely to revise their own beliefs about factual matters than personal opinions (Banerjee et al. 2007)—a pattern mirroring their reasoning about what other people should do (Heiphetz and Young 2017; Wainryb et al. 2004). In sum, it is currently unclear whether a self‐other gap exists when choosing to seek disagreement to learn more.

1.1. The Present Research

In four preregistered studies, we examined children's and adults’ interest in seeking out disagreement for the purpose of learning. In Study 1, we examined children's (6–11 years) and adults’ interest in talking to someone who disagreed with their beliefs across four domains: factual knowledge, moral questions, personal preferences, and ambiguous situations. In Study 2, we compared choices participants made for themselves in the first‐person context to choices participants made for someone else across the same four domains in a third‐person context (Ransom et al. 2025).

Finding that participants were more likely to endorse talking to someone with a different belief when reasoning about someone else than about themselves, we tested two interventions designed to increase children's and adults’ interest in seeking disagreement for learning by increasing the salience of epistemic benefits. In Study 3, we explicitly taught participants (6–11 years and adults) about the benefits of disagreement for learning before asking them to decide whether to talk to someone who agreed or disagreed with them. In Study 4, we encouraged reflection on the benefits of disagreement in a less explicit intervention. We asked children (10–11 years) and adults to consider what someone else should do before making their own decisions—asking them to complete the third‐person trials before the first‐person trials. We hypothesized that third‐person reflection might prime epistemic thinking before the first‐person decision (Grossmann et al. 2021). In Studies 1 and 3, we also asked participants to explain some of their choices to better understand how recognizing the benefits of disagreement for learning develops between ages 6 and 11.

2. Study 1

2.1. Method

2.1.1. Participants

The final sample consisted of 99 children aged 6–11 years (M age = 9.08 years, SD = 1.74 years; 50 females, 49 males), and 100 adults (M age = 38.57 years, SD = 11.95 years, range 18 – 70 years; 49 females, 49 males, 2 gender non‐conforming). Five additional children were excluded due to task interference (n = 3) or technical issues (n = 2), and two additional adults were excluded for failing attention checks.

Children were ethnically diverse, with 94% of caregivers providing information: 38% White, 29% Asian, 9% Black, 1% Latin American, 23% identified as mixed, and 1% as other. Regarding caregiver education (97% reported), 14% of children had caregivers who did not attend university, 24% had one caregiver who did, and 63% had two caregivers with university education. Annual pre‐tax household income was reported for 66% of participants; among those who responded, 8% reported earning less than $50,000 CAD, 15% between $50,000 and $100,000 CAD, 17% between $100,000 and $140,000 CAD, and 60% over $140,000 CAD. For comparison, the median after‐tax income for Canadian families in 2023 was $74,200 CAD (Statistics Canada 2025).

The adult sample was also ethnically diverse, with 98% of participants providing information: 56% were White, 16% Asian, 15% Black, 7% identified as mixed, and 6% as other.

All children were recruited from the Greater Toronto Area (urban and suburban) and participated via Zoom with an experimenter. Adults from the US and Canada were recruited via Prolific and participated independently.

All studies were approved by the University of Toronto's Research Ethics Board (#37434) and conform to recognized ethical standards. Adult participants and children's parents provided written informed consent, while children provided verbal assent.

2.1.2. Stimuli and Procedure

All participants completed the same first‐person perspective trials. Participants were told that they would be asked about their thoughts on different topics. To make the statements in the vignettes more believable, participants were also told that children had previously answered the same questions.

In each trial, participants indicated which of two statements they believed, before watching a vignette of two children: one who agreed and one who disagreed with their belief. Participants were then asked who they would talk to if they wanted to learn more—the agreeing child or the disagreeing child (Figure 1A). Belief domain was manipulated across 16 trials (four per domain): 1) factual knowledge (e.g., germs are very small vs. very big), 2) moral questions (e.g., it is not okay vs. it is okay to hit people), 3) personal preferences (e.g., oranges vs. apples are the tastiest fruit), and 4) ambiguous situations (e.g., the purple rock vs. the orange rock is heaviest). We used the same beliefs as Ransom et al. (2025), see Table S2. After the last trial for each domain, participants were asked to explain their choice (explanations are analyzed in Study 3).

FIGURE 1.

FIGURE 1

Open in a new tab

Illustrative examples of a preference domain trial in A) a first‐person trial and B) a third‐person trial.

Trial order was randomized, with half of the trials featuring male and half female children. The child on the left was always introduced first. The side of the typical belief (factual and moral) and whether it matched the child's belief were counterbalanced across trials.

Participants also completed two interpretative theory of mind tasks: the integrating perspectives paradigm (Amemiya et al. 2021) and the ambiguous referential communication paradigm (Carpendale and Chandler 1996). These measures are analyzed in a separate paper.

2.1.3. Data Analysis

All analyses were performed using R (R Core Team 2024), with generalized linear mixed effects models (GLMMs) fit using lme4 (Bates et al. 2015). p‐values < 0.05 were considered statistically significant. We maintained ‘maximal’ random effects structures where possible, following Barr et al. (2013). Post hoc pairwise comparisons were made using emmeans (Lenth et al. 2019). We assessed the effect of the predictors of interest via likelihood ratio tests comparing the full model with a null model lacking the predictor variables. Odds ratios (ORs) and 95% confidence intervals (CIs) are reported throughout for comparisons of interest.

2.2. Results

Trials in which participants held atypical beliefs were excluded. This affected a small proportion of trials: for children we excluded 11 factual trials (2.8%) and 19 moral trials (4.8%), and for adults we excluded 10 factual trials (2.5%) and 25 moral trials (6.3%).

We fit two binomial GLMMs for participants’ choice of whom to talk to (disagreeing child 1, agreeing child 0), with fixed effects of age group, domain (factual, moral, preference, ambiguous), and their interaction, with a random intercept of participant ID to account for repeated observations. The first model only included children, with age groups: 6–7 years, 8–9 years, and 10–11 years. A model using continuous age in place of age groups is included in the supplementary materials. The second model included both children and adults, with age groups: younger children (6–9 years, combining the two younger groups from the first model, as there were no differences between them), older children (10–11 years), and adults.

2.2.1. Children

The model was significantly better than the null equivalent, χ 2(11) = 66.35, p < 0.001, though showed no interaction between age group and domain, χ 2(6) = 4.86, p = 0.561. A simplified model without the interaction did not show a difference in model fit, χ 2(6) = 4.92, p = 0.555, and revealed significant effects of age, χ 2(2) = 24.15, p < 0.001, and domain, χ 2(3) = 34.84, p < 0.001 (Figure 2). Post hoc analyses showed that the oldest children (10–11 years) were significantly more likely to choose the disagreeing child than 6–7 years (OR = 24.38, CI = [5.13, 115.79], p < 0.001) and 8–9 years (OR = 7.27, CI [1.61, 32.89], p = 0.006), with no significant difference between the younger age groups (p = 0.183). Children were significantly less likely to choose the disagreeing child in the factual domain compared to all other domains (ORs = 0.24–0.52, p ≤ 0.040), and in the moral compared to the ambiguous domain (OR = 0.46, CI [0.26, 0.84], p = 0.005).

FIGURE 2.

FIGURE 2

Open in a new tab

Predicted probability of choosing the disagreeing child in each domain in the first‐person perspective, split by age group. Error bars represent 95% confidence intervals.

2.2.2. Adults and Children

The model fit significantly better than the null, χ 2(11) = 131.98, p < 0.001, with no interaction between age group and domain, χ 2(6) = 10.80, p = 0.095. The simplified model without the interaction did not differ significantly, χ 2(6) = 10.94, p = 0.090, and revealed significant main effects of age group, χ 2(2) = 29.22, p < 0.001 and domain, χ 2(3) = 84.32, p < 0.001. Post hoc comparisons showed that younger children were less likely to choose the disagreeing child than older children (OR = 0.08, CI [0.02, 0.27], p < 0.001) and adults (OR = 0.16, CI [0.06, 0.41], p < 0.001), with no significant difference between older children and adults (p = 0.326). Participants were also less likely to choose the disagreeing child in the factual domain compared to all other domains (ORs = 0.25–0.46, p < 0.001), and less often in the preference compared to the ambiguous domain (OR = 0.55, CI [0.38, 0.80], p < 0.001).

2.3. Discussion

When choosing whether to learn from someone who agreed or disagreed with them, both children and adults preferred agreement. Younger children (6–9 years) in particular rarely sought out disagreement. While the strength of this preference varied across domains, the preference for agreement was found in all of them.

As outlined in the introduction, previous work has shown that both children and adults endorse learning from disagreement when reasoning about what someone else should do (Ransom et al. 2025), at least when the topic relates to a personal preference or ambiguous situations. In Study 2, we directly compared first‐person and third‐person responses across age groups and domains to examine this self‐other gap more closely.

3. Study 2

3.1. Method

3.1.1. Participants

3.1.1.1. Third‐Person Perspective

The final sample of children consisted of 83 children aged 7–11 years (M age = 9.16 years, SD = 1.60 years; 41 females, 42 males). Data for 50 younger children aged 7–8 years, and 100 adults were obtained from an OSF repository (Ransom et al. 2024, Ransom et al. 2025). To ensure that age groups were comparable across studies, we tested an additional 33 older children aged 10–11 years (M age = 10.99 years, SD = 0.61 years; 16 females, 17 males).

The additionally recruited older children were ethnically diverse (100% reported): 27% White, 30% Asian, 9% Black, 30% identified as mixed, and 3% as other. Regarding caregiver education (100% reported), 21% of children had one caregiver who attended university, and 79% had two caregivers who did. Annual household income was reported for 76% of participants. Among those who responded, 12% reported earning between $50,000 and $100,000 CAD, 40% between $100,000 and $140,000 CAD, and 48% over $140,000 CAD.

3.1.1.2. First‐Person Perspective

First‐person perspective data were drawn from Study 1, comprising 33 younger children aged 7–8 years and 34 older children aged 10–11 years. (M age = 9.57 years, SD = 1.63 years; 34 females, 33 males), and 100 adults.

3.1.2. Stimuli and Procedure

The procedure for the third‐person perspective trials followed Ransom et al. 2025; Study 1). In each trial, participants watched a vignette featuring a protagonist who held a belief about a topic, and two other children: one who agreed and one who disagreed with the protagonist. Participants were told that the protagonist wanted to learn more about the topic and were asked whom the protagonist should talk to—the agreeing child or the disagreeing child (Figure 1B). The same beliefs and domains were used as in Study 1 (see Table S2).

Trial order was randomized, with half of the trials featuring male and half female children. The child on the left was always introduced first. The side of the typical belief (factual and moral) and whether it matched the child's belief were counterbalanced across trials.

3.2. Results

To compare how children's and adults’ choice of the disagreeing child differed in the first‐person compared to the third‐person perspective across belief domains, we fit a binomial GLMM with fixed effects of age group (younger children: 7–8 years, older children: 10–11 years, adults), belief domain (factual, moral, preference, ambiguous), perspective (first‐person, third‐person), and all possible interactions, with a random intercept of participant ID.

The model fit significantly better than the null, χ 2(23) = 871.48, p < 0.001, but did not show a three‐way interaction, χ 2(6) = 5.76, p = 0.450. Comparison with a simplified model including all two‐way interactions showed no significant difference in model fit, χ 2(6) = 5.84, p = 0.441. The simplified model revealed significant two‐way interactions between domain and perspective, χ 2(3) = 193.05, p < 0.001, and domain and age group, χ 2(6) = 26.02, p < 0.001 (Figure 3). Post hoc comparisons showed that participants were significantly less likely to choose the disagreeing child in the first‐person than third‐person context in all domains (ORs = 0.06–0.46, p ≤ 0.009) except the moral domain (OR = 0.66, CI [0.37, 1.17], p = 0.154). Adults and older children—but not younger children—were less likely to choose the disagreeing child in the factual than the moral domain (ORs = 0.35–0.52, p ≤ 0.026), and adults and younger children were less likely to choose the disagreeing child in the preference compared to the ambiguous domain (ORs = 0.52–0.60, p ≤ 0.018).

FIGURE 3.

FIGURE 3

Open in a new tab

Predicted probability of choosing the disagreeing child in each domain in the first‐person perspective (Study 1) and third‐person perspective (Study 2), split by age group. Error bars represent 95% confidence intervals.

3.3. Discussion

Compared to reasoning about what someone else should do, participants were less likely to endorse learning from disagreement when reasoning about themselves, except in the moral domain. This difference was most pronounced in the preference and ambiguous domains.

The contrast between reasoning about oneself and about others suggests that reasoning about someone else may have served as a form of self‐distancing, allowing participants to appreciate the epistemic benefits of disagreement without also weighing these against potential social costs, such as confrontation or exclusion that may be more salient when reasoning about themselves. This raises the question of whether making the epistemic benefits of disagreement more salient could increase their interest in engaging with it. If participants have some appreciation of these benefits but are failing to weigh them against social costs, then explicitly highlighting those benefits before participants make their choice might increase interest in seeking disagreement. In Study 3, we tested this directly with an intervention that explained the value of learning from conflicting views before participants chose whom to talk to.

4. Study 3

4.1. Method

4.1.1. Participants

The final sample consisted of 96 children aged 6–11 years (M age = 8.97 years, SD = 1.72 years; 54 females, 42 males), and 100 adults (M age = 36.45 years, SD = 12.6 years, range 18 – 70 years; 50 females, 50 males). Three additional children were excluded due to task interference (n = 1), technical issues (n = 1), or participation in Study 1 (n = 1). Two additional adults were excluded for failing attention checks. All children were recruited from the Greater Toronto Area and participated via Zoom with an experimenter. Adults from the US and Canada were recruited via Prolific and participated independently.

Children were ethnically diverse, with 94% of caregivers providing information: 36% White, 28% Asian, 4% Black, 2% Latin American, 1% Middle Eastern, 28% identified as mixed, and 1% as other. Regarding caregiver education (97% reported), 15% of children had caregivers who did not attend university, 28% had one caregiver who did, and 57% had two caregivers with university education. Annual household income was reported for 73% of participants. Among those who responded, 3% reported earning less than $50,000 CAD, 17% between $50,000 and $100,000 CAD, 24% between $100,000 and $140,000 CAD, and 56% over $140,000 CAD.

The adult sample was also ethnically diverse, with 98% of participants providing information: 60% were White, 15% Asian, 11% Black, 7% identified as mixed, and 6% as other.

4.1.2. Stimuli and Procedure

Participants began by watching a video‐based intervention which explained that people sometimes think the same things and sometimes think differently. It emphasized the potential value of learning from conflicting views and gave an example of the benefits of such learning (see supplementary materials for the full script). Following the intervention, participants completed the same first‐person trials as Study 1, with two additional trials per domain, resulting in 24 across four domains (6 per domain). After the final trial for each domain, participants explained their choice.

4.2. Results

Trials in which participants held atypical beliefs were excluded: for children we excluded 9 factual trials (1.6%) and 11 moral trials (1.9%), and for adults we excluded 16 factual trials (2.7%), and 46 moral trials (7.7%).

We fit the same two GLMMs as in Study 1, one including only children and one including both children and adults. The results replicated the earlier pattern: younger children (6–9 years) were less likely to choose the disagreeing partner than older children (10–11 years) and adults, and disagreement was chosen more often in the preference and ambiguous domains than in the factual and moral domains. Full results are in supplementary materials.

4.2.1. No Intervention vs. Intervention

To assess the impact of the intervention we compared choices of the disagreeing child in Study 1 (No intervention) and Study 3 (Intervention). We fit a binomial GLMM for with fixed effects of age group (6–9 years, 10–11 years, adults), domain, and study (Study 1: No intervention, Study 3: Intervention), and all possible interactions, with a random intercept of participant ID. The model was significantly better than the null, χ 2(23) = 574.77, p < 0.001, revealing a significant three‐way interaction between age group, domain, and study, χ 2(6) = 16.01, p = 0.014 (Figure 4). All age groups were more likely to choose the disagreeing child, in all domains, when they had received the intervention. Averaged across domains, adults’ choices increased from 0.35 to 0.62 (0.28 increase), older children's from 0.44 to 0.59 (0.15 increase), and younger children's from 0.17 to 0.27 (0.10 increase). Post hoc comparisons indicated that this difference was significant for all age groups in the factual (ORs = 3.80–5.25, p ≤ 0.007) and moral domains (ORs = 3.16–3.39, p ≤ 0.038). In the preference domain, it was significant for younger children (OR = 4.11; CIs [1.63, 10.41], p = 0.003) and adults (OR = 7.47; CI [5.26, 10.61], p < 0.001), but not older children (p = 0.233). In the ambiguous domain, only adults showed a significant effect (OR = 5.39, CI [3.81, 7.62], p < 0.001).

FIGURE 4.

FIGURE 4

Open in a new tab

Predicted probability of choosing the disagreeing child in each domain in the first‐person perspective without an intervention (Study 1) and following an intervention (Study 3), split by age group. Error bars represent 95% confidence intervals.

4.2.2. Explanations in Studies 1 and 3

We examined the explanations participants gave for their choices in the last trial in each domain in Studies 1 (No intervention) and 3 (Intervention). These explanations were coded based on three main themes: 1) correct explanations, related to what the participant thought to be true about the world (e.g., “It is a fact that dinosaurs are not alive now, so the other girl must be misinformed”), 2) alternate perspective explanations, referring to the value of considering a different perspective to one's own (e.g., “I think I should talk to her because she might have new information I don't have and can provide me with another perspective”), and 3) similarity explanations, focused on perceived similarities between the participants and their choice (e.g., “Because we have the same beliefs”). In total, 17.5% of explanations were excluded from analyses: 5.5% were coded as fitting multiple categories (5.7% of factual, 7% of moral, 4.1% of preference, and 5.1% of ambiguous explanations), and a further 12% were excluded because they did not fit any category (8.1% of factual, 8.6% of moral, 14.6% of preference, and 16.8% of ambiguous explanations). To assess interrater reliability, three researchers each coded the same 25% of adults’ explanations. Agreement—assessed using Fleiss kappa (Koo and Li 2016)—demonstrated very good agreement (kappa = 0.82). For this subset the codes from one coder were retained for analyses, while the remaining explanations were divided between the coders.

To test whether explanation type varied by study or age group, we fit a repeated‐measures multinomial model using the mclogit package (Elff 2022). Explanation type was the dependent variable, with study (No intervention, Intervention) and age group (6–7 years, 8–9 years, 10–11 years, adults), and their interaction as predictors, with a random intercept of participant ID to account for repeated observations. Model fit was evaluated using BIC values, with lower values indicating better fit (Neath and Cavanaugh 2012). Figure 5 shows the distribution of explanation types across studies and age groups. An additional model including domain as a predictor is provided in the supplementary materials.

FIGURE 5.

FIGURE 5

Open in a new tab

Proportion of each explanation type provided by each age group in Study 1 (No intervention) and Study 3 (Intervention). Values represent the number of explanations.

The model provided a better fit (BIC = 2218.29) than the null (BIC = 2263.20) and indicated a significant two‐way interaction between study and age group. Post hoc comparisons showed an age‐related shift from correct to alternate perspective explanations that was stronger in Study 3. Compared to Study 1, adults and 6–7 years were more likely to provide alternate perspectives explanations in Study 3 (p ≤ 0.046), while 10–11 years and adults were less likely to provide correct explanations (p ≤ 0.009). Only adults were less likely to mention similarity (p < 0.001).

In Study 1, children aged 6–7 years were more likely to provide correct explanations (p ≤ 0.038) and less likely to provide alternate perspective explanations (p ≤ 0.004) than all other groups. Children aged 8–9 years were less likely to give alternate perspective explanations than 10–11 years (p = 0.002) but not adults (p = 0.253), with no difference from either group in correct explanations (p ≤ 0.004). There were no significant differences between 10–11 years and adults (p ≥ 0.078). In Study 3, children aged 6–7 and 8–9 years were more likely to give correct and less likely to give alternate perspective explanations than 10–11 years and adults (p < 0.001). No differences between 6–7 and 8–9 years (p ≥ 0.732), or 10–11 years and adults (p ≥ 0.290). Similarity explanations showed minimal age‐related differences. See Table S3 for full comparisons.

4.3. Discussion

Explicitly teaching participants about the epistemic benefits of disagreement increased their interest in seeking it out. Older children (10–11 years) and adults not only chose disagreement more often, they preferred it. Although younger children (6–9 years) engaged with disagreement 60% more than in Study 1, they continued to prefer agreement. These changes in participants’ choice of whom to talk to were also reflected in their explanations, with those who received the intervention referring to the benefits of alternative perspectives more frequently than those who did not. There was also an age‐related shift away from correctness‐based explanations toward highlighting alternative perspectives in both studies.

Together these results suggest that even a brief explicit intervention can increase interest in learning from disagreement by making its epistemic value salient. Younger children's continued preference for agreement raises the possibility that the effectiveness of such interventions may be limited by children's pre‐existing understanding of the epistemic benefits of disagreement. In Study 1, younger children (6−9 years) rarely justified their decisions by appealing to those benefits. Children may need to first grasp why disagreement is valuable before reminding them of those benefits increases their motivation to seek out disagreement.

To examine whether a less explicit intervention would produce similar increases in seeking disagreement, Study 4 prompted older children and adults to reflect on the epistemic benefits without explicitly highlighting them. Participants first considered what someone else should do before reasoning about themselves. Based on past work with adults (Grossmann et al. 2021), we expected this self‐distancing to make the epistemic benefits of disagreement more salient before participants chose for themselves. We therefore focused on older children (10–11 years), who, unlike younger children, can already articulate why disagreement is epistemically valuable.

5. Study 4

5.1. Method

5.1.1. Participants

The final sample consisted of 72 adults (M age = 35.36 years, SD = 10.4 years; 37 females, 35 males) and 31 10‐ and 11‐year‐old children (M age = 11.14 years, SD = 0.58 years; 16 females, 15 males). One additional child was tested but excluded due to technical issues. Sample sizes were determined by power analyses conducted in G*Power (Faul et al. 2007) aiming for a power of 0.80, based on the observed effect size in Study 1. All children were recruited from the Greater Toronto Area and participated via Zoom with an experimenter. Adults from the US and Canada were recruited via Prolific and participated independently.

Children were ethnically diverse, with 90% of caregivers providing information: 36% White, 39% Asian, 4% Black, 4% Middle Eastern, 14% identified as mixed, and 4% as other. Regarding caregiver education (100% reported), 3% of children had caregivers who did not attend university, 23% had one caregiver who did, and 74% had two caregivers with university education. Annual household income was reported for 90% of participants. Among those who responded, 11% reported earning less than $50,000 CAD, 18% between $50,000 and $100,000 CAD, 21% between $100,000 and $140,000 CAD, and 50% over $140,000 CAD.

The adult sample was also ethnically diverse, with 96% of participants providing information: 59% were White, 13% Asian, 19% Black, 7% identified as mixed, and 1% as other.

5.1.2. Stimuli and Procedure

Half of the adults completed the first‐person perspective trials before the third‐person perspective trials, while the remaining adults and all children completed the third‐person trials then the first‐person trials. The trials used within each perspective were the same as Study 1 (first‐person) and Study 2 (third‐person), except participants were not asked to explain any of their choices.

5.2. Results

Trials where participants’ own beliefs were atypical (adults: 2 (0.3%) factual, 11 (3%) moral; children: 1 (0.4%) factual, 2 (0.8%) moral) were excluded.

5.2.1. Adults

We fit a binomial GLMM to investigate whether adults’ choice (disagreeing child 1, agreeing child 0) varied depending on the order in which they experienced the first‐person and third‐person trials. The model included fixed effects of domain, perspective and perspective order, and their interactions, with by‐participant random slopes for domain and perspective. The model was significantly better than null, χ 2(15) = 129.49, p < 0.001, and revealed a significant three‐way interaction between domain, perspective, and order of perspectives, χ 2(3) = 9.17, p = 0.027 (Figure 6). Post hoc comparisons showed that when participants completed the first‐person trials before the third‐person trials, they were significantly less likely to choose the disagreeing child in the first‐person trials in both the preference (OR = 0.09, CI [0.03, 0.21], p < 0.001) and ambiguous (OR = 0.18, CI [0.08, 0.42], p < 0.001) domains. However, when participants completed the third‐person trials before the first‐person trials, there were no significant differences between perspectives in their likelihood to choose the disagreeing child (ORs = 0.37–1.52, p ≥ 0.063).

FIGURE 6.

FIGURE 6

Open in a new tab

Predicted probability of adults choosing the disagreeing child across belief domains and perspectives, split by within‐subjects trial order. Error bars represent 95% confidence intervals.

5.2.2. Children

We fit a similar GLMM for children aged 10–11years, without order of perspectives predictor, and with a random intercept of participant ID. The model was a significantly better fit than the null, χ 2(7) = 171.61, p < 0.001), and revealed a significant two‐way interaction between domain and perspective, χ 2(3) = 8.21, p = 0.042, see Figure 7. Post hoc comparisons showed no significant differences between first‐person and third‐person trials in any domain (p ≥ 0.065). Children were more likely to choose the disagreeing child in the preference and ambiguous domains than in the factual and moral domains (ORs = 0.04–0.20, p < 0.001). In the factual and moral domains, participants chose the disagreeing child in the first‐person than in the third‐person trials, whereas the opposite pattern was observed in the preference and ambiguous domains, these differences were not significant, though likely underlie the significant interaction.

FIGURE 7.

FIGURE 7

Open in a new tab

Predicted probability of children (10–11 years) choosing the disagreeing child across belief domains and perspectives (within‐subjects). Error bars represent 95% confidence intervals.

5.3. Discussion

Asking participants to consider what someone else should do before reasoning about themselves increased their interest in seeking disagreement, similar to Study 3. Both older children (10–11 years) and adults who completed the third‐person task first were just as likely to engage with disagreement themselves as they were to endorse that someone else should. Eliminating the difference between first‐ and third‐person perspectives observed in Study 2. By contrast, adults who reasoned about themselves first replicated the pattern of results in Studies 1 (first‐person) and 2 (third‐person), being more likely to endorse that others should learn from disagreement than to seek it for themselves.

6. General Discussion

To harness the epistemic benefits of disagreement, people must be willing to engage with those who disagree with them. Across four preregistered studies, we established that children (particularly younger children aged 6 to 9) and adults preferred to engage with someone who agreed with them, rather than someone who disagreed, in a learning context. This occurred despite the relatively trivial nature of the topics we selected, and is consistent with adult work using more consequential and personally meaningful topics (Taber and Lodge 2006), as well as confirmation bias work that finds preferences for seeking confirming over conflicting information (Nickerson 1998; Özkan and Ronfard 2026; Stanovich et al. 2013) and information avoidance (Chalik et al. 2022; Santhanagopalan et al. 2025). Crucially, using the same paradigm, we found that children and adults were less interested in engaging with disagreement when reasoning about themselves than when reasoning about what someone else should do.

Why were participants less interested in seeking out disagreement for themselves than for others? One might argue that children did not recognize the value of disagreement. Indeed, research in third‐person contexts suggests that children aged 7 to 8 are just beginning to express such understanding in their explanations (Ransom et al. 2025). However, this cannot explain why adults also preferred agreement, nor why both children and adults endorsed seeking disagreement more often when reasoning about others. This pattern suggests that the social costs of disagreement may be more salient when reasoning about oneself than about others. We return to this possibility below.

One alternative interpretation is that the preference for agreement might reflect a simple preference for similarity rather than the proposed epistemic and social factors. Similarity likely influenced participants’ choices to some degree, given that both children and adults tend to prefer interacting with like‐minded others (Chalik et al. 2022; Hart et al. 2009). However, similarity alone cannot account for the pattern we observed. Similarity was one of the categories that emerged from participants' explanations, but these references were relatively rare, and young children were more likely to refer to correctness than to similarity. Together, these findings suggest that the preference for agreement and the self‐other difference reflect decreased salience of the epistemic benefits of disagreement and increased salience of the interpersonal costs of engaging with it themselves.

Asking about who someone else should choose in the third‐person trials may have functioned as a form of self‐distancing, allowing participants to consider the epistemic benefits of a conflicting view without requiring them to weigh these against social costs. By contrast, asking participants who they would choose in the first‐person trials may have highlighted the potential interpersonal costs of disagreement—such as confrontation or social exclusion—more than epistemic benefits. Self‐distancing exercises have been shown to foster and increase intellectual humility over time in adults (Grossmann et al. 2021; Kross and Grossmann 2012; Porter 2025). Intellectual humility requires both recognizing the limitations of one's own knowledge and valuing others' perspectives (Musholt et al. 2026; Blakey and Ronfard under review). If social costs overshadow epistemic benefits when deciding for oneself whether to engage with others who disagree, then interventions that make epistemic benefits more salient—either through explicit teaching or through self‐distancing—might shift this balance. Accordingly, we assessed whether we could increase interest in engaging with disagreement by explicitly teaching about the learning benefits of disagreement (Study 3) and prompting self‐distancing through third‐person reasoning (Study 4).

Explicitly teaching or reminding participants about the benefits of talking to someone with a different belief increased their interest in doing so across all age groups. This effect was strongest in adults, who were more interested in seeking disagreement across domains after the intervention, with more selective increases for younger and older children. Following the intervention, adults and older children preferred disagreement. An alternative interpretation is that these changes reflect a desire to follow the “instructions” given in the intervention rather than genuine recognition of the epistemic benefits of disagreement. However, the results are not fully consistent with this account. If a desire to follow instructions were driving the effect, we might expect more consistent increases across ages and domains. Instead, younger children showed smaller increases and effects varied systematically by domain. Importantly, the intervention also changed the reasons participants gave for their choices, with participants referring to the benefits of different views more often in Study 3 than Study 1. Notably, despite being explicitly told about the epistemic value of conflicting perspectives, younger children continued to focus on correctness when reasoning about their choices. This provides additional evidence that appreciating others’ perspectives—and the epistemic understanding this requires—continues to develop in later childhood (Amemiya et al. 2021). It also suggests that reminders about the epistemic benefits of disagreement may be most effective for children who already possess this understanding.

In Study 4 we examined whether a less explicit intervention, prompting self‐distancing through third‐person reasoning, would produce a similar effect in adults and older children. Encouraging reflection on the epistemic benefits of disagreement by having participants reason about someone else before themselves, without explicitly describing those benefits, eliminated the self‐other gap for both older children and adults. Those who reasoned about others first were just as likely to engage with disagreement themselves as they endorsed others should be. This helps to dispel the suggestion that participants in Study 3 were simply following instructions rather than genuinely recognizing the epistemic benefits of disagreement, as a similar pattern emerged even when those benefits were not explicitly highlighted. As noted above, reasoning about someone else first may have activated a self‐distancing process, fostering a reflective mindset that persisted when participants subsequently considered their own choices. By implication, self‐distancing may be an effective intervention for improving children and adults’ intellectual humility and interest in learning from others (see Grossmann et al. 2021, for work with adults).

Across studies, participants were more interested in seeking disagreement when the topic concerned personal preferences and ambiguous situations than when it concerned factual knowledge or moral questions. This pattern is consistent with previous findings in third‐person contexts (Ransom et al. 2025) and prior work showing that people view disagreement as more acceptable when questions lack clear answers (Wainryb et al. 2004) or decisive evidence is absent (Yang et al. 2023). This raises an important question about when it is appropriate or rational to seek out disagreement for learning. Children and adults in our studies treated factual and moral domains as less appropriate contexts for seeking disagreement. However, even in the factual domain, engaging with disagreement can have epistemic value, not necessarily because the conflicting view is correct, but because it can highlight the limits of one's own knowledge, reveal where the different view has come from, or identify the kinds of evidence needed to address a question. At the same time, not all disagreement warrants engagement or belief revision; the credibility and expertise of the disagreeing party matters, as does the context in which the disagreement occurs (Morris et al. 2026).

Engaging with disagreement can be cognitively beneficial (Kuhn 2019; Langenhoff et al. 2025). Children and adults show some awareness of these benefits but appear to be held back by the potential social costs of disagreement. Their interest in learning from disagreement was lower when reasoning about their own beliefs than about what someone else should do. Two brief interventions—explicitly teaching the benefits of disagreement and prompting self‐distancing—increased interest in engaging with disagreement in older children and adults. Understanding how interest in disagreement develops, and how it can be fostered, may help reduce barriers to social learning and foster greater openness to engaging with those who think differently.

Author Contributions

Kirsten H Blakey: conceptualization, methodology, formal analysis, investigation, data curation, visualization, project administration, writing – original draft, writing – review and editing. Samuel Ronfard: conceptualization, funding acquisition, writing – review and editing, methodology.

Funding

This research was supported by a Social Sciences and Humanities Research Council of Canada Insight Grant to Samuel Ronfard (435‐2021‐0170) and a grant from the John Templeton Foundation (62835).

Ethics Statement

This research received approval from the University of Toronto's Research Ethics Board (#37434). Adult participants and children's parents provided informed written consent, while children provided verbal assent.

Conflicts of Interest

The authors declare that there is no conflict of interest regarding the publication of this article.

Supporting information

Supporting Information: desc70237‐sup‐0001‐SuppMat.docx

DESC-29-e70237-s001.docx (622.9KB, docx)

Acknowledgements

We thank members of the Childhood Learning and Development Lab at the University of Toronto Mississauga for their help with stimuli creation, recruitment and data collection.

Data Availability Statement

All hypotheses, methods, and analysis plans were preregistered prior to data collection (Study 1: https://aspredicted.org/h45k‐wwwd.pdf; Study 2: https://aspredicted.org/xvvt‐4ym8.pdf; https://aspredicted.org/z5br‐tffw.pdf; Study 3: https://aspredicted.org/sbqb‐2hf9.pdf; Study 4: https://aspredicted.org/j59r‐g88m.pdf; https://aspredicted.org/b48p3x.pdf). Minor deviations are reported in Table S1. All materials and primary data are publicly available on OSF (first‐person perspective trials: https://doi.org/10.17605/OSF.IO/FTZ4H third‐person perspective trials: https://osf.io/tb6cu/). All analysis scripts are also publicly available on OSF (https://doi.org/10.17605/OSF.IO/FTZ4H).

References

  1. Blakey, K. H. , and Ronfard S.. Under review. “Interpretive theory of mind predicts greater willingness to seek out disagreement for learning”. [DOI] [PubMed]
  2. Amemiya, J. , Walker C. M., and Heyman G. D.. 2021. “Children's Developing Ability to Resolve Disagreements by Integrating Perspectives.” Child Development 92, no. 6: e1228–e1241. 10.1111/cdev.13603. [DOI] [PubMed] [Google Scholar]
  3. Banerjee, R. , Yuill N., Larson C., Easton K., Robinson E., and Rowley M.. 2007. “Children's Differentiation Between Beliefs About Matters of Fact and Matters of Opinion.” Developmental Psychology 43, no. 5: 1084–1096. 10.1037/0012-1649.43.5.1084. [DOI] [PubMed] [Google Scholar]
  4. Barr, D. J. , Levy R., Scheepers C., and Tily H. J.. 2013. “Random Effects Structure for Confirmatory Hypothesis Testing: Keep it Maximal.” Journal of Memory and Language 68, no. 3: 1–43. 10.1016/j.jml.2012.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bates, D. , Mächler M., Bolker B., and Walker S.. 2015. “Fitting Linear Mixed‐Effects Models Using lme4.” Journal of Statistical Software 67, no. 1: 1–48. 10.18637/jss.v067.i01. [DOI] [Google Scholar]
  6. Blakey, K. H. , Rafetseder E., Atkinson M., et al. 2021. “Development of Strategic Social Information Seeking: Implications for Cumulative Culture.” PLoS ONE 16, no. 8: e0256605. 10.1371/journal.pone.0256605. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Carpendale, J. I. , and Chandler M. J.. 1996. “On the Distinction Between False Belief Understanding and Subscribing to an Interpretive Theory of Mind.” Child Development 67, no. 4: 1686–1706. 10.2307/1131725. [DOI] [Google Scholar]
  8. Chalik, L. , Over H., and Dunham Y.. 2022. “Preschool Children Weigh Accuracy against Partisanship When Seeking Information.” Journal of Experimental Child Psychology 220: 105423. 10.1016/j.jecp.2022.105423. [DOI] [PubMed] [Google Scholar]
  9. Confer, J. A. , Schleihauf H., and Engelmann J. M.. 2024. “Children and Adults' Intuitions of What People Can Believe.” Child Development 95, no. 2: 447–461. 10.1111/cdev.13988. [DOI] [PubMed] [Google Scholar]
  10. Danniels, E. , and Perlman M.. 2021. “Because ‘Everybody Believes in Different Things’: Examining Tolerance of Divergent Preferences, Beliefs, and Morals in Kindergarten Students.” Social Development 30, no. 4: 973–993. 10.1111/sode.12520. [DOI] [Google Scholar]
  11. Elff, M. 2022. mclogit: Multinomial Logit Models, with or without Random Effects or Overdispersion (Version 0.9.6) [Computer software]. https://cran.r‐project.org/web/packages/mclogit/index.html.
  12. Faul, F. , Erdfelder E., Lang A.‐G., and Buchner A.. 2007. “G*Power 3: A Flexible Statistical Power Analysis Program for the Social, Behavioral, and Biomedical Sciences.” Behavior Research Methods 39, no. 2: 175–191. 10.3758/bf03193146. [DOI] [PubMed] [Google Scholar]
  13. Grossmann, I. , Dorfman A., Oakes H., Santos H. C., Vohs K. D., and Scholer A. A.. 2021. “Training for Wisdom: The Distanced‐Self‐Reflection Diary Method.” Psychological Science 32, no. 3: 381–394. 10.1177/0956797620969170. [DOI] [PubMed] [Google Scholar]
  14. Hart, W. , Albarracín D., Eagly A. H., Brechan I., Lindberg M. J., and Merrill L.. 2009. “Feeling Validated Versus Being Correct: A Meta‐Analysis of Selective Exposure to Information.” Psychological Bulletin 135, no. 4: 555–588. 10.1037/a0015701. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Heiphetz, L. , Spelke E. S., Harris P. L., and Banaji M. R.. 2013. “The Development of Reasoning About Beliefs: Fact, Preference, and Ideology.” Journal of Experimental Social Psychology 49, no. 3: 559–565. 10.1016/j.jesp.2012.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Heiphetz, L. , and Young L. L.. 2017. “Can Only One Person be Right? The Development of Objectivism and Social Preferences Regarding Widely Shared and Controversial Moral Beliefs.” Cognition 167: 78–90. 10.1016/j.cognition.2016.05.014. [DOI] [PubMed] [Google Scholar]
  17. Koo, T. K. , and Li M. Y.. 2016. “A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research.” Journal of Chiropractic Medicine 15, no. 2: 155–163. 10.1016/j.jcm.2016.02.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kross, E. , and Grossmann I.. 2012. “Boosting Wisdom: Distance From the Self Enhances Wise Reasoning, Attitudes, and Behavior.” Journal of Experimental Psychology: General 141, no. 1: 43–48. 10.1037/a0024158. [DOI] [PubMed] [Google Scholar]
  19. Kuhn, D. 2019. “Critical Thinking as Discourse.” Human Development 62, no. 3: 146–164. 10.1159/000500171. [DOI] [Google Scholar]
  20. Langenhoff, A. F. , Engelmann J. M., and Srinivasan M.. 2023. “Children's Developing Ability to Adjust Their Beliefs Reasonably in Light of Disagreement.” Child Development 94, no. 1: 44–59. 10.1111/cdev.13838. [DOI] [PubMed] [Google Scholar]
  21. Langenhoff, A. F. , Srinivasan M., and Engelmann J. M.. 2024. “Disagreement Reduces Overconfidence and Prompts Exploration in Young Children.” Child Development 95, no. 5: 1616–1627. 10.1111/cdev.14098. [DOI] [PubMed] [Google Scholar]
  22. Langenhoff, A. F. , Thompson B. D., Srinivasan M., and Engelmann J. M.. 2025. “Disagreement Drives Metacognitive Development.” Trends in Cognitive Sciences 29: 1097–1108. 10.1016/j.tics.2025.05.014. [DOI] [PubMed] [Google Scholar]
  23. Lenth, R. , Singmann H., Love J., Buerkner P., and Herve M.. 2019. Emmeans—Estimated Marginal Means aka Least‐Squares Means. https://cran.r‐project.org/package=emmeans.
  24. Mercier, H. , and Sperber D.. 2011. “Why Do Humans Reason? Arguments for an Argumentative Theory.” Behavioral and Brain Sciences 34, no. 2: 57–74. 10.1017/S0140525X10000968. [DOI] [PubMed] [Google Scholar]
  25. Morris, B. , Shaw A., and Liberman Z.. 2026. “Calibrated Deference: Children's Evaluations of Responses to Disagreement Across Knowledge Gaps.” Cognition 272: 106472. 10.1016/j.cognition.2026.106472. [DOI] [PubMed] [Google Scholar]
  26. Musholt, K. , Ronfard S., Rottman J., et al. 2026. “Developing Intellectual Humility: Questions, Dilemmas, and Future Directions.” Current Psychology 45, no. 7: 709. 10.1007/s12144-025-08608-y. [DOI] [Google Scholar]
  27. Neath, A. A. , and Cavanaugh J. E.. 2012. “The Bayesian Information Criterion: Background, Derivation, and Applications.” WIREs Computational Statistics 4, no. 2: 199–203. 10.1002/wics.199. [DOI] [Google Scholar]
  28. Nickerson, R. S. 1998. “Confirmation Bias: A Ubiquitous Phenomenon in Many Guises.” Review of General Psychology 2, no. 2: 175–220. 10.1037/1089-2680.2.2.175. [DOI] [Google Scholar]
  29. Özkan, F. E. , and Ronfard S.. 2026. “The Boundaries and Ontogeny of Myside Bias.” Journal of Experimental Psychology: General ahead of print. 10.1037/xge0001932. [DOI] [PubMed]
  30. Porter, T. 2025. “Chapter Three—Intellectual Humility: On Recognizing Our Limits.” In Advances in Experimental Social Psychology, edited by Gawronski B., Vol. 71, 135–179. Academic Press. 10.1016/bs.aesp.2024.10.005. [DOI] [Google Scholar]
  31. R. Core Team . 2024. “R: A Language and Environment for Statistical computing.” R Foundation for Statistical Computing. https://www.R‐project.org/. [Google Scholar]
  32. Ransom, A. , Blakey K. H., and Ronfard S.. 2024. Learning From Disagreement [Data set]. Open Science Framework. 10.17605/OSF.IO/TB6CU. [DOI] [Google Scholar]
  33. Ransom, A. , Blakey K. H., and Ronfard S.. 2025. “Children's Developing Understanding of the Value of Disagreement for Learning.” Child Development 96: 1373–1384. 10.1111/cdev.14253. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Santhanagopalan, R. , Risen J. L., and Kinzler K. D.. 2025. “Becoming an Ostrich: The Development of Information Avoidance.” Psychological Science 36, no. 7: 528–544. 10.1177/09567976251344551. [DOI] [PubMed] [Google Scholar]
  35. Smaldino, P. E. , and Velilla A. P.. 2025. “The Evolution of Similarity‐biased Social Learning.” Evolutionary Human Sciences 7: e4. 10.1017/ehs.2024.46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Stanovich, K. E. , West R. F., and Toplak M. E.. 2013. “Myside Bias, Rational Thinking, and Intelligence.” Current Directions in Psychological Science 22, no. 4: 259–264. 10.1177/0963721413480174. [DOI] [Google Scholar]
  37. Statistics Canada . 2025. Canadian Income Survey, 2023 . The Daily. Statistics Canada. https://www150.statcan.gc.ca/n1/daily‐quotidien/250501/dq250501b‐eng.htm. [Google Scholar]
  38. Taber, C. S. , and Lodge M.. 2006. “Motivated Skepticism in the Evaluation of Political Beliefs.” American Journal of Political Science 50, no. 3: 755–769. 10.1111/j.1540-5907.2006.00214.x. [DOI] [Google Scholar]
  39. Wainryb, C. , Shaw L. A., Langley M., Cottam K., and Lewis R.. 2004. “Children's Thinking about Diversity of Belief in the Early School Years: Judgments of Relativism, Tolerance, and Disagreeing Persons.” Child Development 75, no. 3: 687–703. 10.1111/j.1467-8624.2004.00701.x. [DOI] [PubMed] [Google Scholar]
  40. Yang, Q. T. , Sleight S., Ronfard S., and Harris P. L.. 2023. “Young Children's Conceptualization of Empirical Disagreement.” Cognition 241: 105627. 10.1016/j.cognition.2023.105627. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information: desc70237‐sup‐0001‐SuppMat.docx

DESC-29-e70237-s001.docx (622.9KB, docx)

Data Availability Statement

All hypotheses, methods, and analysis plans were preregistered prior to data collection (Study 1: https://aspredicted.org/h45k‐wwwd.pdf; Study 2: https://aspredicted.org/xvvt‐4ym8.pdf; https://aspredicted.org/z5br‐tffw.pdf; Study 3: https://aspredicted.org/sbqb‐2hf9.pdf; Study 4: https://aspredicted.org/j59r‐g88m.pdf; https://aspredicted.org/b48p3x.pdf). Minor deviations are reported in Table S1. All materials and primary data are publicly available on OSF (first‐person perspective trials: https://doi.org/10.17605/OSF.IO/FTZ4H third‐person perspective trials: https://osf.io/tb6cu/). All analysis scripts are also publicly available on OSF (https://doi.org/10.17605/OSF.IO/FTZ4H).

Articles from Developmental Science are provided here courtesy of Wiley

RESOURCES