- Tension: The people most likely to dismiss expert opinion are the ones facing a decision where being wrong carries the most severe consequences.
- Noise: The discussion of expertise dismissal focuses on ignorance and tribal politics, missing the structural problem: high stakes make deferring to an expert feel psychologically unsafe.
- The Direct Message: When the decision feels too consequential to get wrong, the mind produces exactly the confidence needed to get it wrong.
To learn more about our editorial approach, explore The Direct Message methodology.
The conventional explanation for why people ignore expert advice involves one of two culprits: ignorance or ideology. The person either doesn’t know enough to recognize what expertise is worth, or they belong to a group that has decided, for tribal reasons, that a particular class of experts cannot be trusted. Both explanations are real and both are documented. Neither is the most interesting one.
The more interesting finding, and the one that most consistently appears in the research literature on influence and decision-making, is this: expert opinion is most likely to be dismissed not by people who are indifferent to the decision at hand, but by people who are deeply invested in it. The higher the perceived stakes, the stronger the psychological resistance to advice that runs counter to what the decision-maker already wants to do. This is not an edge case. It appears consistently enough across domains that it functions, in many high-stakes contexts, as a near-default feature of motivated cognition — not an anomaly.
The mechanism
The relevant psychological construct is motivated reasoning, the tendency to reason toward conclusions that serve a prior motivation rather than from evidence toward truth. It is not conscious. The person engaged in motivated reasoning experiences themselves as evaluating evidence normally. What has changed is the endpoint the evaluation is steering toward: a predetermined conclusion, shaped by what the person needs to be true.
What high stakes do to motivated reasoning is activate it more forcefully. Psychologist Dan Kahan’s research on identity-protective cognition found that when beliefs are tied to group identity or personal self-image, people process threatening information more defensively — and that this effect strengthens, not weakens, with greater cognitive ability. Smarter people, Kahan’s research suggests, are better at constructing rationalizations for conclusions they’ve already reached. Higher cognitive ability is not a reliable defense against motivated reasoning; in some contexts, it is a weapon in its service.
The stakes connection is specific. When a decision is low-stakes — say, choosing between two similar products — the decision-maker has relatively little invested in any particular outcome. Expert opinion is easy to absorb because being wrong costs little. As stakes rise, so does the decision-maker’s psychological investment in making the right call. And crucially, the most available evidence that they’re making the right call is their own prior conviction. The expert who arrives with a different view is not just providing information; they are implicitly suggesting that the decision-maker’s judgment is wrong. At high stakes, that implication is felt as a threat, and threats are defended against.
The financial evidence
Some of the cleanest evidence for this pattern comes from financial decision research, where the relationship between stakes, conviction, and expert dismissal can be measured with some precision. Research on what’s called the “action advice effect” found that clients who already have a strong view on a financial decision are significantly less likely to follow advice that contradicts their position — and that this resistance is specifically driven by the desire to confirm their own opinion rather than ignorance of the expert’s credentials. The expert’s qualifications are known and acknowledged. The advice is understood. It is then set aside in favor of what the client already believed.
This pattern is more severe in high-stakes decisions precisely because conviction is higher in high-stakes decisions. A person considering a small financial transaction may have mild preferences and can incorporate expert input relatively easily. A person making a major investment decision — one they’ve thought about, planned for, and emotionally committed to — has a higher prior conviction and correspondingly higher resistance to the expert who advises against it. The relationship between stakes and dismissal is not an accident of who makes high-stakes decisions. It is a feature of how the psychology of high-stakes decisions works.
What experts do in response
There is a second dimension to this problem. Experts who operate in high-stakes contexts are typically trained in epistemic precision, which means they communicate in the language of probability, uncertainty, and conditionality. The climate scientist says the evidence is consistent with, not certain about. The medical specialist offers options and their associated risks rather than a directive. The financial advisor hedges recommendations with caveats about individual circumstances. This is intellectually honest. It is also, psychologically, an easy target.
Research on risk communication has found that probabilistic expert advice — “there is an 80% chance this is the right course of action” — is processed very differently from confident assertions, and that in high-stakes contexts, people tend to discount hedged probabilistic framing while finding unqualified confident opinion more persuasive. The person who knows exactly what to do, and says so without qualification, is more persuasive than the person who has examined the evidence most carefully and communicates what it actually supports. The expert’s fidelity to uncertainty is weaponized against their influence.
The result is a feedback loop. High stakes activate motivated reasoning, which makes the decision-maker resistant to contrary expert advice. Experts respond to this resistance by emphasizing nuance and complexity, which makes their advice easier to dismiss. The decision-maker takes the expert’s uncertainty as confirmation that the expert doesn’t really know, and their own conviction as evidence that they do. They proceed with the high-stakes decision on the basis of exactly the judgment that expertise was supposed to check.
The social rationality of dismissal
Kahan’s research introduces a dimension that makes this problem more intractable, not less. He argues that in many high-stakes decisions — those involving climate, public health, economic policy — dismissing expert opinion is actually rational, from the perspective of the individual decision-maker’s real interests. An individual’s acceptance or rejection of scientific consensus on a contested issue is unlikely to change the outcome of that issue. But accepting the consensus, if it runs counter to the beliefs of their community, carries social costs: strained relationships, loss of belonging, reputational damage within a group that matters to them.
The calculation, Kahan suggests, is not “which belief is true” but “which belief serves me.” Given that social belonging is a more immediate and reliable source of welfare than being correct about a policy question, identity-protective cognition is not a failure of rationality. It is a form of rationality operating on a different objective function than the one experts assume they’re appealing to.
This framing helps explain why expertise is especially likely to be dismissed when stakes are highest: high-stakes decisions are most often the ones most entangled with identity. Climate is a high-stakes question that has become a proxy for values, politics, and group membership. Vaccination policy became high-stakes precisely because it became contested along identity lines. Financial decisions about major life events — homeownership, retirement — are high-stakes not just economically but psychologically, tied to narratives about who the decision-maker is and what kind of life they’re building. In every case, the stakes that make expert advice most important are the same stakes that make identity most salient, and identity-protective cognition most active.
What this means for influence
The practical implication of this research is uncomfortable for anyone who believes that better information leads to better decisions. It doesn’t, reliably, when the decisions are high-stakes and the decision-maker is already committed. The information environment is not the binding constraint. The psychological environment — the decision-maker’s level of prior conviction, the degree to which their identity is entangled with a particular outcome, the social costs of updating their view — is the binding constraint, and it is one that additional expert information may actively worsen by threatening the conviction it is trying to inform.
The research on what actually changes high-stakes decisions is less developed and less consistent than the research on dismissal, but it points toward a few things: trusted intermediaries who share the decision-maker’s identity and can translate expert consensus into group-compatible terms; approaches that reframe the decision so that updating one’s view doesn’t require losing face; and the slow work of building the kind of relationship in which expert advice is sought rather than received unsolicited. None of these are efficient. All of them are more effective, in high-stakes contexts, than simply providing more information.
The hardest part of the research to sit with is that expertise is not the problem. Expertise is correct, or as correct as it can be, in ways that matter enormously. What is not correct is the assumption that correctness, in a high-stakes context, is sufficient. The mind under pressure is a self-protecting system. It is not looking for the best available input — it is looking for the best available justification for what it has already decided to do.
