Learn

Intelligence

What intelligence actually is from a psychometric standpoint — ability versus achievement, the permeability gradient inside an IQ battery, and the eight subtests of the MaCA mapped to CHC theory.

1. Psychometrics Is a Method, Not a Category

The word “intelligence” generates more confusion than almost any other construct in psychology, primarily because it conflates psychometrics (a methodology for measuring individual differences) with a theory of mind. Psychometrics is a toolkit: factor analysis, item response theory, confirmatory factor models. Intelligence is what emerges when that toolkit is applied to cognitive task performance data. The confusion between method and construct leads to the mistaken assumption that “IQ” names something that either exists as a neurobiological entity or doesn’t exist at all.

2. Construct Validity, and Why Intelligence Is the Paradigm Case

Cronbach and Meehl’s 1955 construct validity framework was written partly to address intelligence, because intelligence had accumulated more predictive validity evidence than any other psychological construct. The criterion validity of IQ — its correlations with educational attainment, occupational complexity, health outcomes, and longevity — is not in dispute. What remains contested is the explanatory account: what does g measure at the neurobiological level?

3. Ability vs. Achievement — The Real Distinction

The ability/achievement distinction is not a clean dichotomy but a gradient. Achievement tests measure what has been learned in specific instructional contexts. Ability tests attempt to measure capacity relatively free of specific prior instruction. In practice, all cognitive test performance reflects both learning history and current processing capacity — the distinction is one of degree of schooling-permeability.

4. The Continuum, Not the Dichotomy

Vocabulary MCQ sits at the high-permeability end: it measures crystallized knowledge accumulated over years of verbal exposure. Fluid reasoning (progressive matrices) sits at the low-permeability end: novel figural pattern completion is far less dependent on specific prior instruction. But even fluid reasoning tasks reflect some practice effects, some cultural assumptions, and some prior exposure to test-taking conventions. The permeability gradient is real; a clean ability/achievement boundary is not.

5. The Permeability Gradient Inside an Ability Battery

Subtest CHC Ability Stratum What It Measures
Progressive Matrices Gf Fluid reasoning Inductive pattern completion on novel figural material
Mental Rotation Gv Visuospatial processing Generating, maintaining, and rotating an internal spatial representation
Color Stroop Gs / executive attention Inhibitory control Suppressing a prepotent reading response in favor of color naming
TOMM Round 1 Performance validity Effort / engagement Recognition memory designed to flag non-credible performance
TOMM Round 2 Glr Delayed visual recognition Consolidation of incidentally encoded visual material
General Knowledge MCQ Gc Crystallized semantic memory Acquired declarative knowledge about the world
Vocabulary MCQ Gc Crystallized verbal knowledge Lexical-semantic knowledge — most reliable single marker of long-term verbal ability
Verbal Analogies Gf / Gc bridge Verbal reasoning Reasoning over verbal relations — stored knowledge + on-the-fly mapping

6. Why the Composite Beats the Scatter — A Meehlian Point

In Clinical versus Statistical Prediction (1954), Meehl reviewed roughly twenty studies and found that mechanical, formula-based combinations of test scores either equalled or beat the inferential judgments of trained clinicians. The composite IQ score, despite its apparent crudeness, predicts life outcomes better than attempts to interpret the scatter of subtest scores. The pattern — verbal higher than performance, memory lower than reasoning — has less predictive validity than the overall composite in most applied contexts, though scatter can have clinical significance in neuropsychological evaluation.