Cognitive Ability and Decision-Making Biases

Whether smarter people make better decisions is a long-running question in cognitive psychology, and the answer is more nuanced than the popular framing suggests. Higher cognitive ability does not eliminate decision-making biases — it shifts which biases people are vulnerable to, and the relationship between IQ and rational thinking is much weaker than people generally assume. Stanovich and West (2008), in Journal of Personality and Social Psychology, demonstrated that intelligence and rational thinking are partially independent: high-IQ individuals show many of the classic decision-making biases (myside bias, anchoring, framing effects) at rates that are not much lower than low-IQ individuals. The recent work of Dawson (2025), in Personality and Social Psychology Bulletin, sharpens the picture in the financial-decisions domain — finding that higher cognitive ability does reduce optimism bias, but increases pessimism, and the net effect on accuracy is more modest than expected.

The honest summary: cognitive ability matters for some decision-making biases more than others, in some domains more than others, and “smart people” remain susceptible to systematic errors that the cognitive-bias literature has documented across the past four decades. Understanding which biases cognitive ability protects against — and which it does not — has practical implications for financial education, evidence-based decision-making, and the broader question of what intelligence actually buys.

The Stanovich-West rationality framework

Stanovich’s research program on rationality (Stanovich & West, 2008; Stanovich & West, 2013) frames the question explicitly. Intelligence as conventionally measured — by IQ tests assessing verbal, mathematical, and spatial reasoning — is what the field calls “Type 1” cognitive capacity. Rational thinking — the ability to override automatic responses, recognize cognitive biases, and apply normatively correct reasoning — is “Type 2” thinking, and it is partially dissociable from IQ. A person can score high on cognitive ability tests while being susceptible to decision-making biases, and vice versa.

The empirical evidence: across many studies of belief bias, myside bias, framing effects, and probabilistic reasoning errors, IQ correlates only weakly to moderately with bias resistance. The correlations are in the range of 0.20 to 0.40 for most biases; some biases (myside bias in particular) show essentially no relationship with IQ. Higher IQ helps with some biases — particularly those involving probabilistic reasoning where computational fluency matters — but does not provide general protection against the broader bias landscape.

The implication for cognitive-ability tests is structural. IQ tests do not measure rationality; they measure a related but distinct construct. Performance on a Wechsler battery or a Raven’s matrices is informative about general cognitive function but not directly informative about how a person will reason in a real-world decision under uncertainty. The dissociation matters because real-world consequences — financial decisions, voting choices, medical adherence — depend on rational thinking, not on IQ alone.

The Kruger-Dunning effect

One of the most-cited findings in this space is the Kruger and Dunning (1999) result, published in Journal of Personality and Social Psychology: people who lack ability in a domain also lack the meta-cognitive ability to recognize their own incompetence. Across multiple ability domains (logical reasoning, grammar, humor judgment), the lowest-performing quartile substantially overestimated their actual performance, while top performers slightly underestimated theirs.

The popular interpretation of “Dunning-Kruger” is often a caricature — that incompetent people think they’re geniuses while geniuses doubt themselves — but the empirical pattern is more measured. The lowest-performing quartile in Kruger and Dunning’s studies estimated themselves at around the 60th-65th percentile, not at the 90th. The effect is real but moderate, and it has been the subject of methodological debate (e.g., regression-to-the-mean explanations) over the past 25 years.

The practical takeaway is that meta-cognitive accuracy — knowing what you know and don’t know — is itself a cognitive ability, partially dissociable from object-level performance. People who are bad at a task are systematically less accurate about their own performance on it than people who are good at the task. This pattern is one of several reasons cognitive ability and decision-making accuracy are not synonymous.

What Dawson (2025) found about financial optimism

Dawson’s 2025 study analyzed data from over 36,000 individuals in a UK cohort, examining the relationship between cognitive ability (memory, verbal fluency, numerical reasoning) and accuracy of financial expectations against actual financial outcomes. The findings are more nuanced than the headline summary suggests.

The straightforward part: higher cognitive ability is associated with substantially more accurate financial expectations. Higher-cognitive-ability individuals were 22% more likely to align their financial predictions with actual outcomes, and lower-cognitive-ability individuals were 34.8% more likely to exhibit unrealistic optimism. The lower-IQ pattern of expecting better-than-actual outcomes is documented robustly.

The unexpected part: higher cognitive ability was also associated with a 53.2% increased likelihood of being overly pessimistic about financial outcomes. The pattern is not “high IQ = realistic, low IQ = optimistic” but rather “high IQ = either realistic or pessimistic, low IQ = optimistic”. The cognitive-ability gradient affects the direction of error in interesting ways, with high-ability individuals shifting from optimism toward pessimism rather than landing exactly on accuracy.

Why pessimism among high-ability individuals? Several mechanisms are plausible. Higher-cognitive-ability individuals may have better awareness of downside risks and base-rate uncertainties, which can systematically tilt expectations downward. They may have more nuanced models of complex systems where outcomes depend on multiple uncertain factors. Or the pessimism may reflect a different kind of cognitive bias — a “negativity bias” in expectation formation that scales with the ability to imagine specific failure modes.

Domain-specificity of cognitive-ability effects on biases

Pulling together Stanovich’s rationality framework, the Dunning-Kruger meta-cognitive effect, and Dawson’s financial-decisions finding, a coherent picture emerges. Cognitive ability protects against:

• Probabilistic reasoning errors — base-rate neglect, conjunction fallacies, conditional-probability misapplications. Higher computational fluency helps directly.
• Some specific framing effects — when alternative framings make the formal structure visible, higher-IQ individuals are more likely to recognize equivalence.
• Optimism bias in domains where evidence is available — Dawson 2025 showed high cognitive ability reduces unrealistic optimism in financial predictions, perhaps by enabling better calibration against historical experience.

Cognitive ability does NOT protect well against:

• Myside bias — the tendency to evaluate evidence in line with prior beliefs. Stanovich and West (2013) showed myside bias correlates near zero with IQ; smart people are at least as susceptible.
• Belief bias in syllogistic reasoning, where the believability of a conclusion drives evaluation of validity.
• The Dunning-Kruger meta-cognitive errors, particularly when the task domain is not the test-taker’s expertise.
• Pessimism bias — Dawson 2025 found higher cognitive ability is associated with more pessimism, suggesting that the cognitive resources that protect against optimism may produce a counter-bias in the opposite direction.

Practical implications

For financial education and decision-support: the Dawson result suggests that interventions targeting numerical reasoning and probabilistic thinking are likely to reduce optimism bias more efficiently than generic “be more careful” appeals. But the same interventions may not address the broader bias landscape — myside bias, framing effects, and motivated reasoning operate through different cognitive substrates and require different intervention strategies.

For research on intelligence: the partial dissociation between IQ and rationality is one of the more important findings in cognitive psychology of the past two decades, and it has implications for how IQ tests are interpreted. High IQ does not guarantee good real-world judgment, and IQ tests should be understood as measuring cognitive capacity rather than the broader rationality construct that practical decision-making requires.

For self-assessment of decision-making: people who score high on IQ tests should not assume they are immune to systematic biases. The empirical evidence shows that high-IQ individuals exhibit most of the classic decision-making biases at rates only modestly lower than the general population, with some biases (myside, belief bias) showing essentially no IQ-related protection.

Where this fits in the broader cognitive-ability literature

The relationship between IQ and rational thinking sits at the intersection of psychometric intelligence research and behavioral decision research. Spearman’s (1904) general intelligence factor captures shared variance across cognitive tests but does not capture all cognitively-loaded behavior. The “rationality quotient” that Stanovich and colleagues have proposed as a separate construct is partially predicted by IQ but has substantial independent variance.

The honest reading is that human cognition has multiple systems that interact in complex ways, and reducing decision-making accuracy to a single IQ-based predictor misses important structure. Both cognitive ability and rational thinking matter for real-world outcomes; they correlate moderately but are not the same construct, and either can be high while the other is low.

Frequently Asked Questions

Does higher IQ make people more rational?

Modestly and selectively, yes. Stanovich and West (2008, 2013) document that IQ correlates with bias resistance at rates of approximately 0.20 to 0.40 for many biases, with myside bias showing essentially no IQ-related protection. Higher IQ helps with some biases — particularly probabilistic reasoning errors — but does not provide general protection against the broader bias landscape.

What is the Dunning-Kruger effect actually about?

It’s about meta-cognitive accuracy. Kruger and Dunning (1999) found that people who lack ability in a domain are systematically less accurate about their own performance than people who are skilled. The popular caricature — incompetent people think they’re geniuses — overstates the original effect, which showed lowest-quartile performers estimating themselves at around the 60th-65th percentile rather than at the top.

Why are high-IQ people more pessimistic about financial outcomes?

Dawson (2025) found that higher cognitive ability was associated with a 53.2% increased likelihood of pessimistic financial expectations. Plausible mechanisms include better awareness of downside risks, more nuanced models of uncertain complex systems, and a counter-balancing pessimism bias that scales with ability to imagine specific failure modes.

Can cognitive biases be trained away?

For some biases yes, for others no. Probabilistic reasoning errors can be reduced through targeted training in statistical thinking. Myside bias and motivated reasoning are more resistant; the cognitive resources required to override them are not the same resources that IQ tests measure, and generic education does not reliably reduce them. Domain-specific debiasing interventions are usually more effective than general appeals to careful thinking.

Does this mean IQ tests don’t matter?

It means IQ tests measure cognitive capacity rather than the broader rationality construct that real-world decision-making requires. They remain valid measures of what they measure, and they predict academic achievement, occupational success, and many cognitive outcomes well. But they do not directly measure decision-making quality, and high IQ does not immunize against the systematic biases the cognitive-bias literature has documented.