What the 2026 Study Changes About Openness and Cognitive Ability

What did the 2026 study measure?

The study connects personality tendencies reported by adults with their results on three assessed skills: literacy, numeracy, and adaptive problem-solving. Its cross-national label describes an analysis of public-use data from 12 countries, not a survey of every country or a comparison of national personality averages. That boundary matters: the findings summarize associations within defined adult samples, rather than establishing what openness means for any one person. In “Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five,” the researchers analyzed Cycle 2 of the OECD Programme for the International Assessment of Adult Competencies (PIAAC), collected in 2022–2023. The analysis covered 67,927 adults aged 16–65. Although 31 countries participated in the cycle, the article says 13 used an extended personality questionnaire that allowed facet-level analysis, and public-use files were available for 12. The OECD’s “PIAAC 2nd Cycle Database” is the repository for those files. Thus the analytic sample was shaped by which countries’ facet data were publicly available; it should not be described as a 12-country census of the world. Personality was assessed with the 30-item Big Five Inventory–2 Short Form (BFI-2-S). Its five broad domains each contain narrower facets: a facet is a more specific tendency nested within a broad trait. The questionnaire represents each of its 15 facets with two items. The study therefore compares both broad domains and these narrower measured tendencies with the three PIAAC skill results. It does not observe personality directly; it uses answers to brief self-report items as indicators. Nor does the word “cognitive ability” mean that the researchers administered a single, intelligence test. They treat literacy, numeracy, and adaptive problem-solving as related but separate indicators, and analyze each outcome separately. That design makes “cross-national” and “facet-level” useful descriptions, but bounded ones. It gives the study a broader adult and country base than a single-country student sample, while retaining short self-report measures and a particular set of assessed skills. Before applying a headline about openness and ability, a reader should keep those actual measures in view: the association concerns questionnaire-defined personality tendencies and PIAAC task performance across the available participating samples.

Sources: Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five; PIAAC 2nd Cycle Database

Which openness facets carry the pooled pattern?

The 2026 study’s pooled results make “openness is linked with cognitive ability” more precise: within its measure, Intellectual Curiosity and Aesthetic Sensitivity had modest positive associations with literacy, numeracy, and adaptive problem-solving; Creative Imagination had no statistically significant pooled association with any of them. The study analyzed nationally representative adult samples from 12 countries, totaling 67,927 people aged 16 to 65, using the 30-item Big Five Inventory–2 Short Form (BFI-2-S). Its three Open-Mindedness facets were each represented by two questionnaire items. These are patterns across people in the study, not evidence that a particular person’s interests reveal their skill level. (Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five.) Intellectual Curiosity concerns interest in ideas and wanting to understand them. Aesthetic Sensitivity concerns noticing or appreciating artistic and aesthetic qualities. Creative Imagination captures readiness to imagine possibilities and alternatives. They are related parts of the BFI-2-S Open-Mindedness domain, but the study estimates their associations separately. In simultaneous regressions, each coefficient describes a facet’s association with a skill while accounting for the other modeled facets. Coefficients for curiosity were about 0.09 across the three outcomes; those for aesthetic sensitivity were 0.12 for literacy, 0.09 for numeracy, and 0.10 for adaptive problem-solving. Creative Imagination’s estimates ranged from −0.01 to 0.00 and were not statistically significant. These standardized estimates are not points added to a test score or a ranking of people. The authors report similar patterns in bivariate correlations, although the joint model estimates each facet’s unique association. (Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five.) The distinction matters because an overall domain result can conceal different facet patterns. In pooled domain-level models, Open-Mindedness had positive coefficients for all three skills; at the facet level, the authors attribute that pattern primarily to curiosity and aesthetic sensitivity. “Openness predicts ability” can therefore sound broader than the result: it may imply that every feature grouped under openness has the same association. A more accurate summary is that, in these models, people reporting more curiosity or aesthetic sensitivity tended on average to have higher assessed scores in each skill. Creative Imagination did not show a significant pooled association. That null does not establish zero association in every country or for every person. The study reports cross-country variation, and its two-item facet measures are brief. The short-form validation, “Short and extra-short forms of the Big Five Inventory–2: The BFI-2-S and the BFI-2-XS,” describes these facets as narrower parts of five domains and notes measurement tradeoffs relative to the full inventory. Treat the labels as a specific measurement lens, not universal bins. When discussing the link, name the facet and skill, and distinguish a group average from a reflection about one person.

Sources: Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five; Short and extra-short forms of the Big Five Inventory–2: The BFI-2-S and BFI-2-XS

What do these facet labels mean—and how precise are they?

The three labels are operational definitions within one questionnaire: they name patterns represented by particular items, not universal categories that every personality theory divides in the same way. In the Big Five Inventory–2 Short Form (BFI-2-S), Open-Mindedness is represented by Intellectual Curiosity, Aesthetic Sensitivity, and Creative Imagination. A facet is a narrower tendency nested within a broader trait domain. In the 2026 study, each facet was measured with two self-report items on a five-point agreement scale. That lets researchers distinguish facets in large samples, but does not make each label a complete description of a person.

Familiar words can suggest more than the instrument measured. “Intellectual Curiosity” may sound like general intelligence, while “Creative Imagination” may sound like demonstrated creative achievement. Here, they are personality facets in a self-report measure. They do not directly test reasoning, knowledge, artistic skill, or the quality of a person’s ideas. The 2026 study relates those reported tendencies to separately assessed adult skills. The relation concerns how group-level measures vary together; the facet is not a substitute for a skill measure.

The BFI-2’s structure explains why this study uses three openness-related facets rather than one undivided score. The full inventory has 60 items covering five domains and 15 facets. Soto and John’s “Short and extra-short forms of the Big Five Inventory–2” describes the BFI-2-S as a 30-item version designed to retain broad coverage while shortening the questionnaire. Each facet receives two items. Their validation research found that this short form may be useful for examining facets in reasonably large samples, while the 15-item extra-short form should not be used for facet assessment. At the domain level, short forms retained much of the full inventory’s reliability and validity; at the facet level, the evidence favors care and adequate sample size. Brevity makes broad surveys more practical, but fewer items leave less room to represent each tendency.

This measurement tradeoff is why the labels should be read as a lens, not as finely resolved personal profiles. Two items can distinguish patterns well enough to test population-level associations in a large sample, yet give a less detailed basis for describing any one respondent. The 2026 study’s 67,927 participants strengthen its pooled analysis; that sample size does not turn a two-item facet into an exhaustive account of someone’s interests or habits.

The 2022 “Personality and Intelligence: A Meta-Analysis” makes the naming issue clear. It combined 272 effect sets and 162,636 participants across Big Five and HEXACO instruments and intelligence outcomes. It found that facets related to intellectual engagement and unconventionality were more strongly related to intelligence than other openness facets. Those terms overlap conceptually with parts of the BFI-2-S Open-Mindedness domain, but are not interchangeable labels for its three facets. Instruments differ in item content, facet boundaries, and outcomes. A responsible comparison preserves those distinctions: the meta-analysis establishes a facet-level precedent, while the 2026 study reports associations for this short-form measure and its three PIAAC skills. Similar wording invites comparison, not one-to-one translation.

Sources: Short and extra-short forms of the Big Five Inventory–2: The BFI-2-S and BFI-2-XS; Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five; Personality and Intelligence: A Meta-Analysis

What does ‘more variance explained’ actually mean?

In the 2026 cross-national study, a model containing all 15 Big Five facets accounted for more variation in assessed PIAAC skills than a model containing the five broad domains. Across 12 countries in the pooled analysis, the facet models accounted for about 10% of literacy and numeracy variation and 9% of adaptive problem-solving variation; domain models accounted for roughly 5–6%. This is a comparison between two ways of representing the full Big Five, not an estimate that openness by itself explains 10% of cognitive ability.

R-squared, often written R², is the proportion of differences in an outcome that a fitted statistical model accounts for in the data being analyzed. Here, the outcome was one of three assessed skills, and the predictors were either five broad personality domains or 15 narrower facets. These percentages summarize how closely each complete model tracked variation across the study sample, for each outcome considered separately. They do not say how much of a skill was caused by personality, nor do they give the probability that a particular person will perform at a certain level.

Why did the facet model fit better? A broad domain combines several narrower tendencies, and those facets need not relate to a particular outcome in exactly the same way. Combining them can blur distinctions that matter for a specific criterion; modeling them separately can capture more of those distinctions. The BFI-2-S validation study by Soto and John describes the instrument as hierarchical: broad domains contain narrower facets, while its short form is designed to assess both levels efficiently. The 2026 result suggests that this extra detail corresponded more closely to the assessed PIAAC outcomes in those data.

The study’s regressions entered domains or facets together to estimate shared variation and conditional associations. The authors say this setup does not imply that personality caused skill. A model R-squared is also different from an individual facet coefficient: the former summarizes the complete predictor set, while a coefficient describes one predictor’s association conditional on the others in that model. Neither statistic is the same as a simple correlation between two variables. The study reports that corresponding simple correlations were highly similar to its standardized coefficients, but that similarity does not turn coefficients into model-wide explained variance.

So the roughly doubled R-squared supports a limited conclusion: the 15-facet representation tracked these three group-level skill outcomes more closely than the five-domain representation in the analyzed samples. It does not mean openness explains 10% of anyone’s ability, and cannot predict an individual result. For an individual interpretation, keep the level of the evidence visible: name the measured tendency and task, and treat the model comparison as a reason for greater descriptive detail, not as a personal estimate.

Sources: Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five; Short and extra-short forms of the Big Five Inventory–2: The BFI-2-S and BFI-2-XS

Does the new study overturn earlier evidence?

No. The 2026 study refines an existing association between openness and cognitive performance; it does not overturn it. The 2022 meta-analysis, “Personality and Intelligence: A Meta-Analysis,” had already found that openness was the strongest Big Five correlate of intelligence overall. Its results also suggested that the connection depended on which part of openness and which kind of cognitive outcome researchers measured. The newer study makes that distinction more visible in a large adult sample, but its measures and questions differ enough that it should be read as an extension, not a direct replication. The 2022 synthesis brought together 272 effect sets and 162,636 participants. Across the reviewed Big Five dimensions, openness showed the strongest overall association with intelligence, with a reported corrected correlation of ρ = .20. The relationship was stronger for crystallized intelligence, which draws on acquired knowledge and learned skills, than for fluid intelligence, which concerns reasoning through problems less dependent on prior learning. At a narrower level, intellectual engagement and unconventionality showed stronger links than several other openness facets. Those results already cautioned against treating openness as one undivided tendency. The 2026 paper, “Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five,” asks a related question using another design and set of measures. It analyzes 67,927 adults in 12 countries and relates the BFI-2-S personality measure to PIAAC literacy, numeracy, and adaptive problem-solving. In its pooled models, Intellectual Curiosity and Aesthetic Sensitivity had modest positive associations across those skills, while Creative Imagination did not show a statistically significant pooled association. The paper also reports that facet associations varied across countries. These results offer a more specific picture of openness facets in assessed adult competencies, rather than a new estimate of the same intelligence outcomes reviewed in 2022. The distinction matters because the meta-analysis combined several instruments, including NEO, BFAS, BFI-2, and HEXACO measures, alongside varied general, fluid, and crystallized intelligence criteria. “Intellectual engagement” in that synthesis may overlap conceptually with “Intellectual Curiosity” in the BFI-2-S, but the labels do not establish identical item content or interchangeable measures. Likewise, PIAAC literacy and numeracy are not simply alternate names for crystallized and fluid intelligence. The newer study’s adaptive problem-solving outcome adds another defined competency, not a measure of every form of reasoning. So the fairest verdict is refinement. The broad positive openness–cognitive association remains plausible, and earlier evidence already pointed toward facet differences. The 2026 contribution is to show how three named BFI-2-S facets relate to three PIAAC skills in a cross-national adult dataset, while documenting variation across those countries. Because the studies differ in instruments, populations, and outcomes, their estimates cannot be compared as if one repeated the other under identical conditions. The practical conclusion is narrower: specify the facet, the cognitive outcome, and whether a claim describes a group average. Neither study makes an openness tendency evidence of any one person’s ability.

Sources: Personality and Intelligence: A Meta-Analysis; Personality and Intelligence: A Meta-Analysis; Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five

Are PIAAC skills the same thing as intelligence?

No. In the 2026 study, literacy, numeracy, and adaptive problem-solving are related but distinct adult competencies used as indicators associated with cognitive ability. They provide evidence about performance on these assessed skills; they do not amount to a comprehensive measure of general intelligence or an individual IQ result. Literacy concerns understanding and using written information; numeracy concerns using mathematical information; adaptive problem-solving concerns reaching goals when a solution is not immediately available and conditions may change. The OECD’s “Assessment Frameworks for Cycle 2” describes adaptive problem-solving as more than finding an answer: a person may need to monitor whether the situation has changed, reconsider a plan, evaluate options, and adjust strategy. Its scope also differs from Cycle 1’s Problem Solving in Technology-Rich Environments. That earlier assessment focused on using digital applications to locate and manage information; Cycle 2 still uses technology-rich settings, but assesses adapting to changing problems rather than digital proficiency itself. The OECD’s “Adult Problem Solving Skills” overview further clarifies that adaptive problem-solving is not limited to digital environments. This change in task definition helps explain why “problem-solving” should not be treated as a single, timeless ability. The later assessment puts more emphasis on responding to shifting conditions and selecting among available resources. Yet it remains one defined competency, shaped by its task framework, rather than a direct reading of every form of reasoning people use. Literacy and numeracy also involve knowledge and practiced skills, not a complete intelligence measure. The 2026 article reports that the three PIAAC outcomes were strongly correlated on average, at r = .88, and that their personality-association patterns were similar. This helps explain why its authors present a common pattern across outcomes. Correlation does not make their content identical; each task still has distinct demands. Similar results across these three measures therefore support a claim about those assessed competencies, not every cognitive capacity. The 2022 meta-analysis provides a useful contrast, not a direct translation. It examined general, fluid, and crystallized intelligence across several personality instruments and found that some openness facets related more strongly to crystallized than fluid intelligence. Here, “crystallized” broadly concerns accumulated knowledge, while “fluid” refers to reasoning in less familiar problems. They are not exact equivalents of PIAAC literacy, numeracy, or adaptive problem-solving. A careful summary names the specific skill tested and stops there: this study links personality tendencies with assessed adult competencies at group level; it does not tell a reader their intelligence or predict how they will perform on a different task.

Sources: Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five; The Assessment Frameworks for Cycle 2 of the Programme for the International Assessment of Adult Competencies; Adult Problem Solving Skills; Personality and Intelligence: A Meta-Analysis

What does cross-country variation change?

The cross-country result changes how confidently we should describe the openness–skill link: the pooled association summarizes these samples, but it is not a fixed pattern that can be assumed to hold equally in every setting. In the 2026 study, associations between BFI-2-S facets and literacy varied across 12 countries. Variation was comparatively high for Intellectual Curiosity and Aesthetic Sensitivity, with standard deviations across country estimates of about 0.07–0.08. Deviations in Sweden contributed especially to this spread. Similar patterns appeared for numeracy and adaptive problem-solving; the supplement also found substantial variation for Creative Imagination. These are differences in estimated within-country links between a facet and an assessed skill. They do not show that people in one country are more open, capable, or imaginative than people elsewhere. Country-standardized scores cannot support national mean comparisons. The OECD’s PIAAC 2nd Cycle Database makes public-use files available, but the analyzed 12 countries do not represent every country or culture. “Cross-country variation” can refer to whether average facet levels differ between countries, or whether a facet’s association with a skill differs. The study addresses the second question. Researchers standardized personality facet and cognitive skill scores within each country before estimating associations. The public-use files contain country-standardized facet scores rather than item-level personality responses, so the researchers could not independently test measurement invariance across their 12-country sample. Measurement invariance means that a measure relates to the underlying construct comparably across groups. The international PIAAC consortium found at least partial metric invariance for broader Big Five domains in the full PIAAC data. That does not replace facet-level checks in this analytic sample. The study says associations may still be analyzed, but country mean levels should not be compared. PIAAC’s sampling, translation, and cross-cultural adaptation procedures are relevant strengths; the study describes nationally representative probability samples and standardized translation steps. But these procedures do not prove every item functions identically in every subgroup or language. They are separate from the specific limit on independent invariance analyses. The evidence supports neither dismissing the findings as incomparable nor treating them as culturally uniform. The variation does not identify its cause. Culture, education systems, language experience, learning opportunities, or measurement differences are possible questions for future research, not explanations established here. The authors also call potential cultural moderators speculative. A deviation centered in Sweden is a feature of these estimates, not evidence of national character or a cultural mechanism. The analysis covers countries with public-use facet data and brief two-item facet measures. For readers, the practical change is modest but important: treat the pooled pattern as a summary across these samples, not a universal constant. The study supports saying that openness-facet associations with assessed skills vary across the countries examined. It cannot tell an individual what their curiosity or aesthetic attention means for their own literacy, numeracy, or problem-solving, or explain a country-level difference. A claim that “people from this country are naturally more curious or better at reasoning” goes beyond the evidence.

Sources: Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five; PIAAC 2nd Cycle Database

Why might a facet and a skill move together without one causing the other?

A statistical link does not tell us which way influence runs. In “Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five,” researchers analyzed data from 67,927 adults aged 16–65 in 12 countries. They modeled literacy, numeracy, and adaptive problem-solving as outcomes and personality facets as predictors, accounting for the other modeled facets. The authors say this regression estimates shared variance and unique associations; it does not assert that personality causes skill. They cannot establish whether a tendency came before a skill, followed it, or developed alongside it. Several pathways are plausible, but remain interpretations, not findings tested by that study. Someone who often seeks explanations or reads beyond an immediate task may accumulate practice relevant to literacy or numeracy. Conversely, a person who already handles a skill comfortably may find related activities more rewarding and choose them more often. Both a measured tendency and skill could also reflect conditions such as access to education, language experience, or time. These possibilities can coexist. The observed association cannot establish their relative importance or rule out other explanations. The match between a tendency and a task matters. Intellectual Curiosity, as measured in this study’s short questionnaire, concerns a narrower aspect of openness than the broad domain label. Literacy and numeracy assessments capture particular adult competencies; performance can reflect accumulated knowledge and opportunities to practice as well as how someone approaches ideas. Repeated engagement with idea-rich activities could therefore coincide with stronger performance on some assessed tasks. It does not mean curiosity produces a fixed gain, or that every curious person will perform well. The study did not measure these proposed practice pathways, so this account is not a reported result. “Personality and Intelligence: A Meta-Analysis” offers comparison, not a causal answer. Its synthesis of 272 effect sets and 162,636 participants found openness to be the strongest Big Five correlate of intelligence overall, with stronger links to crystallized than fluid intelligence. It combines different personality measures and cognitive criteria, and does not identify a mechanism. The 2026 PIAAC analysis adds a large cross-national adult sample and specific skill outcomes, while retaining the same limit: association is not temporal or causal evidence. The conclusion is modest. The results support describing small, facet-specific group associations and asking what experiences or task demands might accompany them. They do not show that raising openness improves cognitive performance, that skill training changes a personality tendency, or that this pattern predicts one reader’s ability. Distinguishing pathways would require evidence that tracks change over time or directly tests a proposed mechanism. Until then, treat causal stories as research questions, not advice about what a person should change.

Sources: Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five; Personality and Intelligence: A Meta-Analysis

How can a reader apply the evidence without turning it into a verdict?

Use a three-part check: name the tendency, the task, and whose evidence is being described. ‘Openness is linked with cognitive ability’ leaves all three vague. Is it intellectual curiosity, aesthetic sensitivity, or creative imagination? Is the outcome literacy, numeracy, or adaptive problem-solving? Does the claim describe a sample average or one person? The 2026 study, ‘Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five,’ found modest pooled associations for some measured facets across these skills. That supports a precise group-level description; it does not let a reader infer an individual’s skill from a personality label.

Consider an illustrative observation: someone repeatedly chooses speculative fiction. That is evidence of a preference for imaginative activity. By itself, it does not establish that the person often seeks unfamiliar ideas, attends closely to aesthetic qualities, or performs at a particular level on a literacy or numeracy task. These are different observations and outcomes. The BFI-2-S validation study, ‘Short and extra-short forms of the Big Five Inventory–2,’ treats Open-Mindedness facets as distinct parts of a broader domain; the broad label cannot stand in for every narrower tendency.

A one-time choice may reflect the book at hand, available time, familiarity, or company. Repeated choices across settings give a more useful basis for describing a tendency, though they still do not measure cognitive skill. The pooled openness domain had a modest positive relation, and the facet pattern was broadly similar across all three assessed skills. The study does not erase the association; it narrows its meaning. For self-understanding, ask: ‘What do I repeatedly choose to explore, and what direct evidence do I have about this specific task?’

Sources: Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five; Short and extra-short forms of the Big Five Inventory–2: The BFI-2-S and BFI-2-XS

What is a useful next step?

When someone says openness is linked with cognitive ability, make the sentence more precise before deciding what it means: name the facet, name the skill, and say whether the evidence describes a group average or an individual observation. For example, the 2026 study found a modest pooled association between Intellectual Curiosity and literacy, numeracy, and adaptive problem-solving. That result describes patterns across adults in the study; it does not tell you how any one person will perform. The study’s title, “Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five,” is a useful reminder that both sides of the claim need a name. For your own reflection, ask: What do I repeatedly choose to explore, and what direct evidence do I have about this specific task? Separate the answer from a label. Choosing to investigate a difficult question again and again is an observable pattern; it is not, by itself, a measure of literacy or reasoning. Consider whether the pattern recurs across settings. If you want to consider how this tendency sits alongside other personality tendencies, the private Context Profile offers a reflection exercise. It is not a cognitive measure, a career selector, or a validated assessment. Explore the personality profile learning library at /topics for more ways to describe everyday patterns carefully.

Sources: Unpacking the personality–cognitive ability link: a cross-national facet-level analysis of the Big Five

Questions readers ask