INTELLIGENCE

Testing Fundamentals, Prediction, and Correlational Validity

• Purposes of Academic and Psychological Testing:

  • Achievement Measurement: Assesses learned material and knowledge acquisition over an instructional period (e.g., standard classroom exams).

  • Selection and Prediction: Designed to forecast future academic or professional performance (e.g., the SAT and GRE, used to select candidates based on predicted college success).

  • Predictive Efficiency: Longitudinal metrics such as grade point average (GPA) across $4$ or $5$ years of high school or college provide strong predictive power, but require years to collect. Standardized tests achieve comparable predictive utility in an administration time of roughly $1$ hour.

  • Combined Predictive Power: Combining objective test scores with multi-year grade records yields the strongest prediction of future student success.

• Real-World Correlations with IQ:

  • Income: IQ test scores show only a modest or small positive correlation with long-term occupational earnings.

  • Military Advancement: IQ demonstrates its strongest occupational correlation with promotions and success in the military, outpacing correlations found in other professional fields.

  • Job Complexity: A standard observation is that the correlation between IQ and job performance increases as job complexity rises (weaker correlations for low-complexity positions; stronger correlations for high-complexity positions). While intuitively plausible, this relationship remains subject to complex statistical debate.

• Psychological Definitions of Intelligence:

  • Core constructs include the capacity to solve novel problems, deduce correct solutions, comprehend and acquire complex information, and adapt flexibly to unfamiliar environments.

  • Mental Quickness: Simple reaction time paradigms (e.g., pressing a button as rapidly as possible when a visual dot appears on a screen) demonstrate an inverse correlation with IQ: individuals with higher IQ scores exhibit shorter (faster) reaction times. A potential confound is participant motivation.

  • Methodological Limitations in Intelligence Research: Studies examining intelligence rely almost exclusively on correlational designs rather than controlled experimental manipulations. The historical literature prioritizes group-level variances across populations rather than within-person cognitive architecture.

Origins, History, and Misuse of Intelligence Testing

• The Binet-Simon Origins:

  • Alfred Binet: French physician who developed the first formal intelligence test to establish an objective mechanism for identifying schoolchildren requiring supplemental educational assistance.

  • Original Subtests: Measured varied faculties including copying drawings, repeating digit sequences, recognizing currency, and explaining logical scenarios.

  • Constructive Purpose: Formulated with benign intent to elevate struggling students and raise overall societal intellectual achievement through targeted remediation.

  • Mental Age versus Chronological Age: Established age-normed baselines across developmental cohorts ($5$-year-olds, $6$-year-olds, $7$-year-olds). An individual child's raw score yielded a mental age ($\text{MA}$), which was evaluated relative to their chronological age ($\text{CA}$).

• American Adoption, Eugenics, and Forced Sterilization:

  • When imported to the United States, intelligence tests were administered to adult populations rather than solely school-aged children.

  • Test outcomes were co-opted by the eugenics movement to rank racial, ethnic, and social groups hierarchically and justify population control.

  • Buck v. Bell and Compulsory Sterilization: The United States Supreme Court legally upheld involuntary sterilization for individuals receiving low IQ scores to prevent lower-scoring adults from reproducing. Approximately $60{,}000$ individuals were forcibly sterilized under these state-sanctioned programs.

  • Historical Test Bias: Early testing regimes (such as those in $1927$) primarily captured test-taking familiarity and cultural exposure aligned with dominant demographic power structures rather than innate intelligence.

Modern Psychometrics, Standardization, and Structural Validity

• The Wechsler Intelligence Scales:

  • The most pervasive modern assessment system, offering tailored variants for adults (WAIS) and children (WISC).

  • Verbal Subtests:

  • Vocabulary: Providing precise definitions of words.

  • Similarities: Identifying conceptual commonalities between distinct entities (e.g., explaining how an airplane and an automobile are alike).

  • Short-Term Memory: Digit-span forward and backward recall.

  • General Information and Comprehension: Responding to broad civic and social queries (e.g., explaining why societies collect taxes).

  • Performance Subtests:

  • Picture Completion: Identifying missing critical details within an illustration.

  • Digit Symbol Coding: Rapid substitution tasks mapping symbols to corresponding numbers under time constraints.

  • Picture Arrangement: Ordering scrambled sequential cartoon panels to generate a coherent narrative.

  • Modern psychometrics over the past $20$ years has prioritized eliminating culturally biased questions.

• IQ Scoring and Normal Distribution Norming:

  • Ratio IQ Formulation: The classical formula defined intelligence quotient as:     $\text{IQ} = \frac{\text{MA}}{\text{CA}} \times 100$

  • Example: A $5$-year-old child ($\text{CA} = 5$) performing at the cognitive benchmark of an average $10$-year-old ($\text{MA} = 10$) yields:     $\text{IQ} = \frac{10}{5} \times 100 = 200$

  • Deviation IQ Norming: Modern scoring applies standard score transformations ($z$-scores) that force raw score distributions into a Gaussian (normal) bell curve.

  • Defined Parameters: The mean is fixed by convention at $\mu = 100$ with a standard deviation of $\sigma = 15$.

  • Statistical Dispersion: Approximately two-thirds of the population ($68.2\%$\text{ of individuals}) fall within one standard deviation of the mean, between $\text{IQ} = 85$ and $\text{IQ} = 115$.

  • Alternative Testing Models: Non-Gaussian scales exist, such as the Wonderlic Personnel Test utilized at the National Football League (NFL) Scouting Combine, which yields a raw score range from $0$ to $40$.

• Psychometric Properties: Reliability versus Validity:

  • Reliability: The temporal consistency and repeatability of a measurement instrument. Measuring physical height represents a highly reliable index, but possesses zero validity as an index of intelligence. Flipping a coin to determine an IQ score is neither reliable nor valid. Modern IQ tests demonstrate high test-retest reliability.

  • Validity: The degree to which an instrument assesses the theoretical construct it purports to measure.

  • Predictive Correlates: IQ scores correlate positively with high school and collegiate GPA, adult occupational status, earned income, marital stability, legal compliance (avoiding incarceration), and physiological longevity.

  • Explained Variance ($R^2$): IQ accounts for roughly $25\%$ of the total variance across these life milestones. The remaining $75\%$ of variance is driven by non-IQ factors, including personality dimensions, socio-cultural conditions, formal education, and generational family wealth.

  • Socioeconomic Benchmarks per Standard Deviation:

  • Mean ($\text{IQ} = 100$): Corresponds on average to completing high school and undertaking $1$ to $2$ years of community college coursework.

  • $+1\sigma$ ($\text{IQ} = 115$): Typified by university graduation and entry into white-collar professions.

  • $-1\sigma$ ($\text{IQ} = 85$): Typified by high school non-completion and employment in unskilled or skilled manual labor.

The Flynn Effect and Reasoning Modalities

• The Flynn Effect Phenomenon:

  • Longitudinal examination of raw (un-normed) test scores across the globe revealed a consistent, secular increase in measured IQ scores since approximately the year $1920$.

  • Identified by political scientist James Flynn; the label was coined by Richard Herrnstein and Charles Murray in their publication The Bell Curve ($1994$).

  • Critique of The Bell Curve: Herrnstein and Murray posited that differences in Black and White American IQ distributions reflected fixed biological divergences—a premise contradicted by biological science.

• Deconstructing Secular Gains: Abstract versus Functional Reasoning:

  • The historical rise in IQ is not driven by gains in arithmetic computation, vocabulary breadth, or general factual knowledge, all of which have remained stable.

  • Gains are localized almost entirely within abstract categorical subtests, particularly similarity judgments.

  • Differential Scoring of Similarity Questions:

  • Abstract/Taxonomic Response: Stating that dogs and rabbits are both mammals earns maximum credit on standardized rubrics.

  • Functional Response: Stating that dogs are used to hunt rabbits represents an ecologically valid, functional relationship, but receives reduced point credit.

  • Coley's Expert versus Novice Categorization Paradigm:

  • Research by John Coley at Northeastern University examined how individuals generalize novel biological phenomena (e.g., "If a frog contracts Disease X, what other organism has Disease X?").

  • Novices (Undergraduate Students): Defaulted to taxonomic, perceptual similarity, selecting toads.

  • Domain Experts (Ecologists and Biologists): Relied on functional transmission dynamics, identifying predators that ingest frogs or prey that frogs consume.

  • Standard IQ scoring metrics penalize functional expert reasoning while rewarding naive abstract classification.

  • Cross-Cultural Classification in Papua New Guinea:

  • Indigenous tribespeople in Papua New Guinea tasked with sorting domestic objects (such as potatoes and knives) arranged them functionally, pairing a peeling knife alongside a potato.

  • Western university students grouped items taxonomically (all foodstuffs clustered separately from all cutlery).

  • When asked how an unwise person ("an idiot") would organize the stimuli, tribal participants noted that an unwise individual would isolate identical functional types into disconnected, abstract piles.

Classical and Contemporary Models of Intelligence

• Spearman’s Two-Factor Theory ($g$ and $s$):

  • Charles Spearman observed the "positive manifold": diverse, content-independent cognitive subtests (mathematical, spatial, linguistic) correlate positively with one another.

  • Specific Factors ($s_1, s_2, s_3, \dots$): Represent variance unique to a designated subtest (e.g., arithmetic computation or vocabulary recall).

  • General Intelligence ($g$): A latent, higher-order statistical construct derived from the positive intercorrelations among the specific tests. General intelligence is an explanatory statistical construct rather than an independently testable psychological entity.

• Thurstone’s Primary Mental Abilities ($1950$):

  • L. L. Thurstone disputed the unity of Spearman’s $g$, proposing seven distinct, modular mental abilities:

  • Verbal comprehension

  • Verbal fluency

  • Number facility

  • Spatial visualization

  • Associative memory

  • Perceptual speed

  • Inductive reasoning

  • Thurstone's model was largely abandoned due to persistent intercorrelations among these faculties and its failure to account for specialized artistic genius (such as the musical virtuosity of Jimi Hendrix).

• Cattell’s Dichotomy: Fluid versus Crystallized Intelligence:

  • Fluid Intelligence ($G_f$): Core cognitive processing capacity, working memory operations, mental speed, and the ability to solve novel problems in unfamiliar domains without relying on prior instruction. Fluid intelligence peaks in late adolescence or early adulthood, plateaus, and exhibits steady senescent decline in later life.

  • Crystallized Intelligence ($G_c$): The accumulation of factual knowledge, linguistic definitions, and procedural competencies acquired via cultural exposure and schooling. Crystallized intelligence increases across adulthood, only showing precipitous declines under neuropathological states (e.g., Alzheimer's disease).

• Carroll’s Three-Stratum Model ($1995$):

  • Formulated by John Carroll to synthesize prior frameworks into a hierarchical system, integrating broad categories like fluid intelligence ($G_f$) and crystallized intelligence ($G_c$) beneath a single overarching general cognitive stratum.

Neurobiology and Cognitive Substrates

• Raven’s Progressive Matrices:

  • A non-verbal test consisting of abstract geometric grids containing a missing segment that must be deduced from multiple choice alternatives.

  • Purely via statistical evaluation, Raven’s Progressive Matrices demonstrates the highest correlation with all other validated intelligence measures, serving as the benchmark measure of general fluid ability.

• Prefrontal Executive Load and the $n$-Back Paradigm:

  • Complex abstract reasoning tasks activate the frontal lobes, localized within the dorsolateral prefrontal cortex (DLPFC).

  • The $n$-Back Task Protocol:

  • Participants monitor a continuous stream of visual stimuli (e.g., spatial coordinates of an illuminated square on a screen).

  • $0$-Back: Detects immediate target conditions or identical sequential items.

  • $1$-Back: Evaluates whether the current stimulus matches the item presented one step prior.

  • $n$-Back ($n \ge 2$): Requires maintaining, continuously updating, and comparing dynamic mental sequences against stimuli presented $n$ trials earlier.

  • Cognitive load escalates exponentially with higher values of $n$, requiring intensive DLPFC activation.

• Neural Efficiency Hypothesis:

  • Functional neuroimaging during reasoning tasks (including the $n$-back and Raven's matrices) reveals that individuals with lower baseline IQ demonstrate greater metabolic activation across the frontal lobes than high-IQ individuals.

  • Smarter brains process equivalent cognitive problems with greater metabolic economy, requiring fewer neural resources to execute identical operations.

Behavioral Genetics and Environmental Determinants

• Genetic Heritability Estimates:

  • Complete mapping of intelligence to specific genetic loci remains elusive.

  • Behavioral genetics relies on kinship and twin designs:

  • Monozygotic (identical) twins reared together display IQ score correlations approaching $1.0$.

  • Monozygotic twins reared apart in disparate households maintain extremely high IQ correlations.

  • Genetically unrelated individuals reared in the same domestic environment exhibit modest IQ correlations.

  • Biological siblings reared in separate environments retain significant positive correlations.

  • Across behavioral genetic paradigms, broad heritability ($h^2$) of general intelligence is estimated at approximately $50\%$, leaving the remaining $50\%$ attributable to environmental factors.

• Environmental Influences and Confounding Variables:

  • Breastfeeding Duration:

  • A longitudinal cohort study tracing $3{,}000$ individuals from infancy to young adulthood identified a $5$ to $7$ point IQ advantage for children breastfed for at least $6$ months.

  • Confound: No mechanistic property of human milk has been established to drive this cognitive difference. Maternal ability to sustain prolonged breastfeeding correlates with household financial stability, parental leave access, and socio-economic resources.

  • Birth Weight:

  • Higher infant birth weight positively correlates with subsequent cognitive performance, reflecting superior prenatal nutrition, maternal healthcare access, and optimal gestational environments.

  • Birth Order Dynamics:

  • Firstborn children display an average IQ of approximately $103$, with mean IQ scores declining progressively with later birth ranks.

  • Theoretical Mechanism: The intellectual environment of a home with two adult parents ($30$ years of age) and one newborn possesses a higher intellectual mean than a home where additional infant siblings dilute the intellectual interaction per child.

  • Caveats: Sample variations introduce wide statistical error margins in studies of later-born children whose older siblings died prematurely.

  • Early Linguistic Exposure:

  • Longitudinal household audio monitoring reveals that total volume of verbal engagement delivered to infants correlates with vocabulary and academic achievement.

  • Total words spoken scales with socioeconomic status (SES): professional households deliver vastly higher word volumes than working-class or social-welfare households, reflecting discretionary parental time and economic bandwidth.

• Working Memory Intervention Claims:

  • Experimental trials investigating adaptive daily training on working memory tasks (e.g., dual $n$-back) demonstrated post-test gains on transfer fluid intelligence measures relative to non-trained controls.

  • These findings sparked the commercial development of mobile "brain-training" applications.

  • Subsequent broad-scale empirical replications failed to consistently substantiate these transfer effects, leaving working memory training as a method to permanently increase generalized fluid intelligence controversial.

Scientific Racism, Stereotype Threat, and Socioeconomic Disparities

• Scientific Racism and Socio-Historical Classifications:

  • Historical attempts to biologize race argued that group disparities in average IQ between Black and White Americans represented genetic variations, rather than consequences of structural economic disparities.

  • The United States Immigration Act of $1924$ established restrictive quotas targeting Southern Italian and Eastern European Jewish populations based on biased intelligence metrics.

  • Biological taxonomy demonstrates that human racial constructs do not constitute distinct biological categories. In The Histories (written circa $400\text{ BCE}$), Herodotus categorized human racial groupings in ways that do not map onto contemporary racial categories.

  • The observed Black-White IQ gap in the United States emerges as early as age $4$, tracking median household income, systemic resource disparities, and environmental access.

• Claude Steele’s Stereotype Threat Research:

  • Investigated performance gaps between high-achieving Black and White undergraduate students at Stanford University, controlling for baseline academic achievement.

  • Experimental Manipulation:

  • Neutral Condition: When an identical difficult verbal test was administered without descriptive framing, Black and White Stanford students performed equally.

  • Stereotype Threat Condition: When the test was framed as a diagnostic measure of underlying intelligence, Black participants underperformed relative to their White peers.

  • Mechanistic Definition: Stereotype threat is a disruptive psychological state occurring when an individual fears that their personal performance may inadvertently validate a pre-existing negative cultural stereotype regarding their social group.

  • Generalizability Across Demographics:

  • Women taking standardized mathematics examinations underperform when gender stereotypes are made salient, despite outperforming men across standardized metrics nationally.

  • Older adults underperform on memory evaluations when age-related cognitive decline is primed.

  • Gay men display diminished physical performance when athletic stereotypes are cued.

  • White individuals show performance decrements in motor timing and dance rhythms when negative coordination stereotypes are evoked.

  • The magnitude of performance impairment is mediated by how deeply the targeted stereotype threatens the individual's core social identity.

Cognitive Mindsets and Longitudinal Interventions

• Carol Dweck’s Mindset Framework:

  • Evaluated subjective beliefs regarding the malleability of intelligence among seventh and eighth-grade students.

  • Implicit Theories of Intelligence:

  • Entity Theory (Fixed Mindset): The belief that intelligence is a static, genetically predetermined trait incapable of meaningful alteration.

  • Incremental Theory (Growth Mindset): The belief that cognitive faculties can be developed through effort, strategic problem-solving, and deliberate practice.

  • Academic Outcomes: Longitudinal tracking of mathematics coursework demonstrated that students endorsing a fixed mindset exhibited stagnant academic trajectories, whereas those adopting a growth mindset achieved progressive gains in their mathematical grades and standardized test scores.

• The HighScope Perry Preschool Project ($1962$):

  • A randomized controlled trial conducted in Ypsilanti, Michigan, tracking $123$ low-SES Black children over a longitudinal span of $40$ years.

  • Study Design:

  • Participants were randomly assigned to either an intensive intervention group (a high-quality active learning preschool curriculum coupled with weekly home visits for $1$ year) or a control group (standard regional conditions, receiving no formal preschooling).

  • Cognitive Trajectory:

  • The intervention produced immediate, significant surges in childhood IQ scores during the preschool period.

  • By age $8$, the absolute IQ differential between the program and control cohorts had converged, leading early skeptics to erroneously claim the intervention was ineffective.

  • Adult Life Outcomes at Age $40$:

  • Criminality: The preschool intervention cohort exhibited a substantial reduction in chronic criminality, with only $36\%$ having five or more arrests by age $40$, compared to $55\%$ in the control group.

  • Economic Independence: Program participants demonstrated significantly higher median annual earnings, higher rates of employment, and higher degrees of financial self-sufficiency.

  • Educational Attainment: Significantly elevated rates of high school graduation, superior scores on standardized basic literacy and achievement measures at age $14$, and higher rates of homework completion at age $15$.

  • The long-term data demonstrate that high-quality early childhood intervention produces transformative life improvements that persist even after initial gains on standardized IQ instruments regress toward the mean.

Questions & Discussion

• Classroom Inquiry:

  • Prompt: "Any questions?"

  • Response: No student questions were raised. The remaining instructional plan comprises an unstructured review session on Monday to answer open-floor student questions, followed by the administration of Exam 2 on Wednesday.