Throughout human development, the architecture of the mind undergoes profound structural reorganization, shifting from generalized functional capacities in infancy toward specialized, modular cognitive faculties in adulthood. The age differentiation hypothesis stands as one of the cornerstone psychometric paradigms attempting to capture this systemic transition, postulating that cognitive abilities dissociate from a singular general factor into increasingly autonomous domains as individuals mature. Understanding whether, how, and why human intelligence bifurcates across chronological age provides profound insight into neural maturation, educational attainment, and lifespan developmental dynamics.
Age Differentiation Hypothesis
1. Concise Definition
The age differentiation hypothesis is a developmental psychometric proposition stating that the structural organization of human cognitive abilities transitions from a unified, relatively undifferentiated general intelligence factor (g) in early childhood into a collection of increasingly distinct, specialized, and autonomous cognitive abilities (such as verbal, spatial, and numerical faculties) as children mature into adolescence and early adulthood.
In theoretical psychometrics, this construct suggests that the proportion of variance accounted for by a higher-order general factor diminishes across age cohorts, accompanied by an attenuation in the intercorrelations among individual subtests measuring distinct cognitive domains. Conversely, lower-order group factors account for an increasing proportion of cognitive test variance over developmental time, reflecting greater specialization of cognitive functions.
Within broader lifespan psychology, the hypothesis is frequently examined alongside its conceptual counterpart, the age dedifferentiation hypothesis. While differentiation characterizes childhood and adolescence, dedifferentiation posits that in late adulthood and senescence, neurobiological degeneration leads to a secondary reintegration or collapse of distinct abilities back into a singular, highly correlated generalized profile of functional decline.
2. Etymology & Linguistic Origin
The term is a compound psychometric designation derived from developmental biology, philosophy, and classical Latin. The primary operative noun, differentiation, originates from the Neo-Latin differentiatio, stemming from the classical Latin verb differre, which combines the prefix dis- (meaning “apart,” “away,” or “asunder”) and ferre (meaning “to bear” or “to carry”). In classical usage, the root denoted the act of distinguishing, separating, or producing diversity among unified entities.
In the nineteenth century, embryologist Karl Ernst von Baer adopted “differentiation” to describe the biological development of organisms from homogenous, undifferentiated germ layers into specialized tissues and complex organs. The sociological and psychological application of the term was subsequently popularized by philosopher Herbert Spencer, who argued that all evolution proceeds from an indefinite, incoherent homogeneity to a definite, coherent heterogeneity.
The precise psychometric phrase “age differentiation hypothesis” emerged in twentieth-century psychological literature following early inquiries into the organization of mental traits. The formulation was formalized by American psychometrician Henry E. Garrett in the mid-1940s to describe the developmental fractionation of general intellectual ability into specialized cognitive mechanisms across chronological age.
3. Pronunciation & Grammatical Form
Pronunciation: /eɪdʒ ˌdɪf.əˌrɛn.ʃiˈeɪ.ʃən haɪˈpɒθ.ə.sɪs/
Part of Speech: Complex noun phrase (singular).
Plural Form: Age differentiation hypotheses.
Grammatical Variants: The adjectival and participial forms include age-differentiated (e.g., “an age-differentiated cognitive profile”) and the verbal process age differentiation (e.g., “cognitive performance differentiates with age”). In psychometric discourse, the term operates as a countable abstract noun phrase when referring to the formal theoretical proposition, and as an uncountable process noun when describing the observed phenomenon of age-related factorial divergence.
4. Detailed Conceptual Explanation
To fully grasp the age differentiation hypothesis, one must understand how cognitive psychologists conceptualize the internal architecture of human intelligence. In psychometric assessment, individuals who perform well on one type of cognitive test (such as vocabulary) generally tend to perform well on other, seemingly unrelated tests (such as abstract pattern reasoning or mental rotation). This pervasive positive correlation across heterogeneous intellectual tasks is known as the positive manifold, which Charles Spearman mathematically modeled as the g factor, or general intelligence.
The age differentiation hypothesis proposes that the magnitude of this positive manifold is not static across the human lifespan; rather, it is systematically moderated by developmental age. In early childhood, the brain operates as a more globally integrated, generalized computational organ. An individual child’s performance on any intellectual task is heavily constrained by overarching systemic parameters, including global processing speed, working memory capacity, and fundamental executive control. As a result, cognitive tests administered to young children reveal high inter-item correlations, with the general factor g accounting for the dominant share of total variance across diverse test domains.
As the individual transitions through late childhood, adolescence, and early adulthood, the cognitive architecture undergoes systematic differentiation. As brain structures mature and environmental experiences (such as formal education and idiosyncratic interests) accumulate, specific cognitive faculties begin to decouple. Mathematical reasoning, verbal fluency, spatial orientation, and perceptual speed become increasingly independent from one another. Consequently, an adolescent may display exceptional competence in spatial-mechanical reasoning while exhibiting only average aptitude in verbal semantics, a divergence far less pronounced in early childhood.
Mathematically, differentiation manifests in factor-analytic models as a structural shift. When exploratory or confirmatory factor analysis is conducted across progressive chronological age strata, three distinct phenomena are predicted to occur: (1) the average correlation among lower-order factors declines; (2) the percentage of total variance explained by the first unrotated principal component or the general factor decreases; and (3) specialized group factors or domain-specific abilities account for a progressively higher percentage of unique variance.
Underlying this psychometric restructuring are profound neurobiological and experiential drivers. Neurobiologically, cerebral development involves initial overproduction of synaptic connections followed by prolonged synaptic pruning, progressive myelination of axonal pathways, and the segregation of large-scale functional brain networks. As localized cerebral regions achieve functional modularity, cognitive systems transition from diffuse neural recruitment to efficient, localized computation, providing the physiological basis for differentiated intellectual aptitudes.
5. Historical Development
The conceptual origins of the age differentiation hypothesis trace back to the pioneering debates surrounding the structure of human intellect in the early twentieth century. Charles Spearman, who established the two-factor theory of intelligence in 1904, recognized that intellectual functioning was dominated by g, but he also speculated about how ability levels and developmental stages might influence the strength of cognitive relations, formulating what would later be known as Spearman’s Law of Diminishing Returns (SLODR).
In the 1930s and 1940s, psychometric research expanded rapidly through the work of Louis Leon Thurstone, who formulated the Multiple Factor Analysis model. Thurstone argued that intelligence was composed of distinct Primary Mental Abilities (PMAs), including verbal comprehension, spatial visualization, and associative memory. This direct tension between Spearman’s singular g and Thurstone’s multiple independent factors set the theoretical stage for a developmental reconciliation: perhaps intelligence was Spearmanian in early childhood and Thurstonian in adulthood.
Henry E. Garrett formally articulated this developmental synthesis in his seminal 1946 paper, “A Developmental Theory of Intelligence,” published in the American Psychologist. Garrett reviewed existing correlational studies and proposed that cognitive abilities begin as a unified, undifferentiated pool of generalized energy and gradually crystallize into distinct patterns of ability over the course of physical and mental growth. Garrett argued that as a child matures, distinct verbal, numerical, and spatial aptitudes emerge as semi-autonomous functional entities.
During the latter half of the twentieth century, empirical investigations into Garrett’s hypothesis yielded conflicting findings. In 1970, Gunter Reinert published an exhaustive review providing tentative support for Garrett’s model, but noted that methodological inconsistencies between cross-sectional cohorts often confounded age effects with education and birth cohort effects. Around the same time, researchers such as Paul B. Baltes and K. Warner Schaie began exploring lifespan developmental trajectories, prompting investigators to look beyond youth differentiation to evaluate whether aging in late life initiated a reverse process: cognitive dedifferentiation.
From the late 1990s through the modern era, the development of sophisticated statistical techniques—specifically multigroup confirmatory factor analysis (MGCFA) and local structural equation modeling (LSEM)—radically reshaped this field of inquiry. Psychometricians such as Timothy Salthouse, Ian Deary, and Elliot Tucker-Drob re-evaluated the differentiation hypothesis using large-scale, nationally representative, and longitudinal datasets, revealing that cognitive differentiation is far more nuanced, subtle, and context-dependent than originally envisioned by Garrett.
6. Theoretical Foundations
The age differentiation hypothesis rests upon multiple foundational frameworks drawn from psychometrics, cognitive neuroscience, and lifespan developmental theory.
From the psychometric perspective, the hypothesis is grounded in structural hierarchical models of intelligence, notably the Cattell-Horn-Carroll (CHC) theory. The CHC taxonomy organizes human intelligence into three strata: Stratum I (narrow, specific abilities), Stratum II (broad abilities such as Fluid Reasoning, Crystallized Intelligence, Visual Processing, and Short-Term Memory), and Stratum III (the general factor, g). The age differentiation hypothesis predicts that the structural path coefficients linking Stratum III (g) to Stratum II (broad abilities) weaken systematically as children grow older, giving Stratum II abilities greater statistical and functional independence.
In cognitive development, the hypothesis parallels Jean Piaget’s constructivist theory of cognitive growth. Piaget posited that cognitive development proceeds via the continuous reorganization of mental structures through assimilation and accommodation. While early infant sensorimotor operations reflect undifferentiated schemas, the formal operational stage reached in early adolescence permits abstract, domain-specific symbolic reasoning across multiple independent operational channels.
Raymond Cattell’s Investment Theory provides an explanatory behavioral mechanism for differentiation. Cattell posited that individuals are born with a generalized fluid intelligence (gf), a biologically based capacity for problem-solving and abstraction. Over time, individuals “invest” this general fluid ability into specific cultural, academic, and vocational pursuits. Through repeated practice and specialized exposure, these differential investments crystallize into distinct, domain-specific crystallized abilities (gc), driving psychometric differentiation across childhood and adolescence.
Finally, modern network theories of intelligence provide an alternative mathematical conceptualization. Network models propose that general intelligence does not reflect a physical, latent higher-order entity; rather, it represents an emergent property arising from mutualistic interactions among distinct cognitive processes during development. In early life, cognitive processes are deeply interdependent—deficits or gains in one system immediately constrain or accelerate others. Over developmental time, as these processes achieve structural maturity and efficiency, their mutualistic dependence stabilizes or diminishes, generating the psychometric illusion of a decoupling general factor.
7. Key Components, Types & Dimensions
The academic literature classifies cognitive differentiation along several theoretical and structural dimensions:
- Structural (Covariance) Differentiation: The systematic decrease in the magnitude of intercorrelations among lower-order cognitive tests or broad latent abilities as chronological age increases. This forms the classical foundation of Garrett’s hypothesis.
- Variance Differentiation: The shift in the proportion of variance explained by the general factor relative to specific factors. In variance differentiation, the eigenvalue of the first unrotated principal component decreases with age, while the eigenvalues associated with specialized group factors increase.
- Ability Differentiation (Spearman’s Law of Diminishing Returns – SLODR): A parallel dimension of differentiation based on intellectual capacity rather than chronological age. SLODR posits that individuals with higher overall intelligence exhibit a more differentiated cognitive profile (lower inter-subtest correlations) than individuals with lower overall cognitive functioning.
- Age Dedifferentiation: The reciprocal developmental trajectory observed in late adulthood, wherein the integrity of specialized cognitive systems deteriorates due to common biological aging mechanisms, causing previously independent cognitive abilities to become strongly intercorrelated once more.
- Domain-Specific Differentiation: The hypothesis that differentiation does not occur uniformly across the entire cognitive architecture, but rather unfolds asymmetrically within specific cognitive sectors (for example, academic-crystallized domains differentiating while perceptual-fluid domains remain stable).
8. Examples & Illustrative Cases
To conceptualize the age differentiation hypothesis in practical terms, consider the differing cognitive profiles of children compared to older adolescents across standardized evaluations.
Case Illustration 1: Childhood Uniformity
A clinical psychologist administers the Wechsler Intelligence Scale for Children (WISC) to a sample of 6-year-old children. At this developmental baseline, an individual child who scores in the 85th percentile on the Verbal Comprehension Index is exceptionally likely to score within a very similar band (80th to 90th percentile) on the Visual Spatial Index, Fluid Reasoning Index, and Processing Speed Index. In this cohort, an overarching constraint—such as fundamental executive control or general sustained attention—dictates performance across disparate task modalities. When a factor analysis is conducted on the 6-year-olds’ scores, the first principal component accounts for upwards of 55% to 60% of total variance, reflecting an undifferentiated cognitive structure.
Case Illustration 2: Adolescent Specialization
The same battery or its late-adolescent counterpart (the WAIS) is evaluated in a cohort of 17-year-olds. At this stage, individual profiles exhibit pronounced intra-individual variability. A high-school student may score in the 98th percentile on the Visual Spatial and Fluid Reasoning subtests, demonstrating brilliant mechanical and abstract geometric reasoning, yet perform only in the 48th percentile on Verbal Comprehension and lexical retrieval tasks. Conversely, an avid reader may display extraordinary semantic fluency alongside modest spatial orientation skills. A factor analysis conducted on this older cohort reveals that the first principal component explains significantly less variance (e.g., 35% to 40%), while the unique group factors (Verbal, Spatial, Memory) account for substantially higher proportions of unique variance.
Case Illustration 3: Longitudinal Academic Differentiation
In an educational tracking study, elementary school students are monitored longitudinally from age 7 to age 16. In the early grades, academic performance in reading, arithmetic, and spatial geometry correlates at approximately r = .75. By age 16, following years of domain-specific academic instruction, divergent personal interests, and neurodevelopmental specialization, the correlation between advanced calculus achievement and advanced literary analysis drops to r = .35. This decoupling illustrates how environmental enrichment intersects with biological maturation to differentiate cognitive competencies.
9. Measurement & Assessment
Empirical evaluation of the age differentiation hypothesis requires rigorous psychometric and statistical methodologies designed to compare structural relationships across groups while avoiding methodological artifacts.
Historically, researchers relied on exploratory factor analysis (EFA) applied independently to cross-sectional age cohorts. Investigators calculated the average inter-test correlation coefficients within each age group, extracted the first unrotated principal factor, and evaluated whether the percentage of accounted variance declined across successive age strata. However, this approach was criticized for subjective factor extraction rules, inability to test for measurement equivalence, and susceptibility to sampling variability.
In modern cognitive psychology, the gold standard methodology involves multigroup confirmatory factor analysis (MGCFA). Researchers must first establish formal measurement invariance across age groups:
- Configural Invariance: Verifying that the same pattern of latent factors and indicator variables holds across all age cohorts.
- Metric (Weak) Invariance: Demonstrating that the factor loadings of individual indicator variables onto their respective latent factors are invariant across age groups. Testing the age differentiation hypothesis is statistically uninterpretable without metric invariance, as differences in correlations could simply reflect changing measurement properties of the test items themselves.
- Structural Invariance Evaluation: Once metric invariance is verified, researchers examine the structural parameters across age groups. Age differentiation is confirmed if the latent factor intercorrelations significantly decrease with age, or if the second-order factor loadings of broad abilities onto the general factor (g) decline significantly.
More recently, psychometricians have introduced Local Structural Equation Modeling (LSEM). Rather than dividing continuous age samples into arbitrary, discrete age brackets (which reduces statistical power and creates artificial boundaries), LSEM uses non-parametric kernel weighting to estimate continuous parameter trajectories across age. This allows researchers to track smooth, non-linear trajectories of factor loadings, factor variances, and factor correlations across the entire lifespan.
10. Applications & Practical Significance
The validity of the age differentiation hypothesis carries far-reaching practical ramifications for psychometrics, education, clinical neuropsychology, and personnel selection.
Test Design and Norming:
For developers of intelligence batteries (such as the Wechsler scales, the Stanford-Binet, or the Woodcock-Johnson batteries), the differentiation hypothesis dictates how scores should be aggregated and reported across development. If cognitive abilities are undifferentiated in young children, reporting a single Full-Scale IQ (FSIQ) score is psychometrically valid and clinically informative. However, if cognitive architecture differentiates significantly by late adolescence, relying solely on an omnibus FSIQ score obscures critical intra-individual strengths and weaknesses, necessitating the interpretation of broad index scores (e.g., Verbal Comprehension Index vs. Fluid Reasoning Index) for educational or clinical decision-making.
Educational Tracking and Curriculum Architecture:
Understanding the developmental timing of cognitive differentiation informs educational policy. In early childhood education, integrated curricula that link verbal, spatial, and numerical domains are theoretically congruent with the child’s unified cognitive structure. In contrast, secondary and post-secondary educational systems benefit from modular, differentiated curricula (e.g., specialized STEM tracks versus humanities concentrations) that align with the specialized, decoupled cognitive competencies emerging during adolescence.
Clinical and Neuropsychological Diagnostics:
In clinical neuropsychology, understanding whether an individual’s cognitive profile is expected to be uniform or differentiated is essential for identifying neurological impairment. A substantial discrepancy between verbal and non-verbal performance in a 6-year-old child might signal focal developmental anomalies or specific learning disabilities, precisely because such abilities are expected to be highly correlated at that age. In contrast, that same magnitude of discrepancy in an 18-year-old might represent completely normal cognitive differentiation and specialization.
11. Research & Empirical Evidence
The empirical literature surrounding the age differentiation hypothesis is characterized by intense debate, with decades of studies yielding mixed, contradictory, and highly nuanced findings.
Henry Garrett’s initial formulation in 1946 cited several early studies showing lower intercorrelations among cognitive tests in older adolescent samples compared to elementary-aged cohorts. In 1970, Reinert’s systematic synthesis lent support to the hypothesis, concluding that abilities appear more differentiated during adolescence than during early childhood. However, many early twentieth-century studies suffered from profound methodological limitations, including failure to control for range restriction, differences in test content across age groups, and reliance on cross-sectional designs that conflated age effects with educational cohort effects.
Beginning in the late 1990s, modern psychometricians subjected the hypothesis to rigorous tests using large, nationally representative standardization samples. A landmark study by Ian Deary and colleagues (1996) examined ability differentiation across multiple large-scale cohorts, finding evidence that while ability differentiation (SLODR) occurred under certain conditions, age differentiation was surprisingly weak or absent across early adolescence when tested with strict psychometric controls.
In a comprehensive study published by Juan-Espinosa and colleagues (2000, 2002), the researchers analyzed Spanish standardization samples of the WAIS across multiple age strata. Their findings demonstrated remarkable structural invariance of the g factor from early adolescence through mature adulthood, arguing that the fundamental architecture of human intelligence remains largely constant across the chronological lifespan, directly challenging Garrett’s classical assertions.
Conversely, significant support for developmental differentiation has been documented when researchers examine earlier developmental windows (such as the transition from toddlerhood to mid-childhood). In a seminal longitudinal analysis, developmental psychologist Elliot Tucker-Drob (2009) investigated cognitive data from early childhood through middle age. Tucker-Drob demonstrated that differentiation occurs prominently between the ages of 2 and 6 years, with the general factor explaining substantially less variance as specific cognitive capacities crystallize. However, from middle childhood through adulthood, the structure remains comparatively stable, suggesting that differentiation is an early developmental phenomenon rather than an ongoing adolescent process.
12. Cultural & Cross-Cultural Considerations
The manifestation and velocity of cognitive differentiation are deeply dependent upon sociocultural context, particularly the availability and structure of formal schooling.
Cross-cultural research indicates that formal schooling acts as a primary catalyst for cognitive differentiation. In societies with universal, compulsory education systems that emphasize distinct academic disciplines (such as advanced mathematics, grammar, visual arts, and literature), children are systematically trained to cultivate specialized mental habits. This systematic cultivation accelerates the psychometric decoupling of cognitive faculties. In cultures where children receive uniform, non-specialized training or where educational attainment is limited, cognitive performance across diverse domains tends to remain significantly more unified, exhibiting persistent high inter-subtest correlations into adulthood.
Furthermore, cultural differences in the holistic versus analytic conceptualization of intelligence influence how abilities differentiate. In many Western cultures, intelligence is conceptualized analytically, emphasizing the isolation and decontextualization of distinct mental mechanisms. In contrast, many East Asian and African traditional intellectual frameworks emphasize holistic problem-solving, social competence, and practical coordination. When Western psychometric batteries are translated and applied in non-Western cultural settings, the expected patterns of factor differentiation frequently fail to replicate, underscoring that the psychometric emergence of modular cognitive factors is not purely an endogenous biological inevitability, but rather a biocultural development.
13. Criticisms, Debates & Limitations
The age differentiation hypothesis has remained one of the most vigorously contested paradigms in developmental psychometrics, facing several severe theoretical and methodological critiques.
The Confound of Educational Selection and Range Restriction:
One of the most profound criticisms centers on sample composition across age cohorts. In many cross-sectional studies, older adolescent samples (often drawn from high school or university environments) exhibit significant cognitive selection and range restriction compared to younger, unselected elementary school cohorts. In statistics, restricting the range of an evaluated sample mathematically attenuates correlation coefficients. Critics point out that what appears to be psychometric differentiation in older groups is often a statistical artifact produced by examining a more homogenous, highly educated cohort.
Difficulty Establishing Metric Invariance:
A persistent psychometric challenge is that identical tests are rarely appropriate for individuals of vastly different ages. A test suitable for a 4-year-old (e.g., matching colored blocks) is far too simple for a 16-year-old, while the adolescent test (e.g., matrix reasoning) cannot be solved by the young child. Consequently, developmental batteries must alter item composition across age levels. These alterations introduce qualitative shifts in the constructs being measured, making it difficult to establish the true metric invariance necessary to prove that factor correlations have fundamentally dissociated.
The Ability-versus-Age Confound (SLODR Interference):
As originally demonstrated by Spearman, individuals of higher overall cognitive ability tend to have lower intercorrelations among cognitive subtests (Spearman’s Law of Diminishing Returns). Because average cognitive performance increases continuously across childhood and adolescence due to neurological maturation and education, older children exhibit higher absolute ability levels than younger children. Methodologists argue that observed age differentiation effects may simply represent ability differentiation in disguise, reflecting raw performance gains rather than an authentic chronological restructuring of mental faculties.
14. Related Terms & Distinctions
A rigorous understanding of the age differentiation hypothesis requires distinguishing it from closely related developmental and psychometric concepts:
- Spearman’s Law of Diminishing Returns (SLODR) / Ability Differentiation: While age differentiation posits that cognitive abilities dissociate as a function of chronological age, ability differentiation posits that abilities dissociate as a function of overall intellectual level. Under SLODR, high-IQ individuals possess more differentiated profiles than low-IQ individuals, regardless of age.
- Age Dedifferentiation: The theoretical inverse of age differentiation, referring specifically to the late-life process wherein cognitive abilities, sensory faculties, and neurological systems re-converge and become more highly correlated due to age-related biological degeneration in senescence.
- Cattell’s Investment Theory: A developmental theory explaining how differentiation occurs behaviorally. It posits that fluid intelligence (gf) is invested over time into specialized cultural and educational learning, creating distinct crystallized abilities (gc).
- Factorial Invariance: A psychometric prerequisite and statistical condition wherein the mathematical properties (factor loadings, intercepts, residual variances) of a measurement instrument remain identical across distinct demographic groups or longitudinal timepoints.
15. Summary & Key Takeaways
The age differentiation hypothesis represents a foundational model in cognitive development and psychometrics, positing that human mental ability begins as a relatively unified general capacity (g) in early childhood before fractionating into autonomous, specialized cognitive domains throughout development. While early twentieth-century psychometricians embraced the hypothesis as an intuitive synthesis of Spearmanian and Thurstonian models of intelligence, modern empirical investigations using multigroup confirmatory factor analysis and local structural equation modeling reveal that differentiation occurs predominantly in early childhood, stabilizing by late childhood and early adolescence.
Ultimately, cognitive differentiation reflects the complex interplay between neurobiological maturation—such as synaptic pruning and functional network segregation—and cultural experiences, particularly formal schooling and specialized skill acquisition. Appreciating this developmental divergence is essential for the design, administration, and interpretation of intellectual assessments, guiding clinical diagnoses, and shaping educational practices across the lifespan.
References
- Deary, I. J., Egan, V., Gibson, G. J., Austin, E. J., Brand, C. R., & Kellaghan, T. (1996). Intelligence and the differentiation hypothesis. Intelligence, 23(1), 45–59. https://doi.org/10.1016/S0160-2896(96)90005-3
- Garrett, H. E. (1946). A developmental theory of intelligence. American Psychologist, 1(9), 372–378. https://doi.org/10.1037/h0058385
- Juan-Espinosa, M., Cuevas, F. J., Escorial, S., & García, L. F. (2000). The differentiation hypothesis and the role of the general factor (g) in intelligence: An empirical study with the WAIS-III. Psicothema, 12(4), 580–586.
- Reinert, G. (1970). Comparative factor analytic studies of intelligence throughout the human life-span. In L. R. Goulet & P. B. Baltes (Eds.), Life-Span Developmental Psychology: Research and Theory (pp. 453–484). Academic Press.
- Tucker-Drob, E. M. (2009). Differentiation of cognitive abilities across the life span. Developmental Psychology, 45(4), 1097–1118. https://doi.org/10.1037/a0015864