You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
Of all the questions psychology asks about children, none is more charged than where does intelligence come from? The answer matters not because psychologists enjoy ranking children — they do not — but because the belief that intelligence is fixed and inherited has, historically, been used to justify some of the discipline's worst abuses, from culturally biased immigration testing to eugenic streaming. Getting the science right is therefore both an intellectual and an ethical duty. This lesson covers the first topic of the OCR child option: intelligence, situated in the biological area. Following the applied-option format, it has three strands. The Background sets out what psychologists mean by intelligence and the biological and genetic factors thought to shape it. The Key research is Van Leeuwen, Van den Berg and Boomsma's (2008) A twin-family study of general IQ, taught in full core-study depth — its aim, the powerful twin-family design, the sample, procedure, results and conclusions, and a balanced evaluation. The Application is a method of assessing intelligence, examined critically. Throughout, the nature–nurture debate is the animating question: Van Leeuwen's design is, at bottom, a machine for estimating how much of the variation in children's intelligence is genetic.
| This lesson covers | OCR H567 Component 03, Section B topic | AO focus |
|---|---|---|
| What intelligence is; biological and genetic factors (background) | Child psychology — Intelligence (Biological) | AO1; AO2 |
| Van Leeuwen et al. (2008): aim, twin-family design, sample, procedure | Key research — a twin-family study of general IQ | AO1; AO2 |
| Results (heritability of IQ) and conclusions | Key research — findings and conclusions | AO1 |
| Evaluation: design strengths, assumptions, ethics, generalisability | Key research — evaluation; issues and debates | AO3 |
| Application: a method of assessing intelligence | Child psychology — application | AO2; AO3 |
The specification is referenced descriptively throughout; consult the official OCR H567 specification document for the exact published wording. This lesson develops AO1 (the concept of intelligence and detailed knowledge of the study), AO2 (understanding the twin-family logic and applying it to assessment) and AO3 (evaluating the study and the assessment method, and engaging the nature–nurture debate). Full citation: Van Leeuwen, M., Van den Berg, S. M. & Boomsma, D. I. (2008) A twin-family study of general IQ, Learning and Individual Differences, 18, 76–88.
Intelligence is deceptively hard to define. A workable statement is that intelligence is a general capacity for reasoning, problem-solving, learning from experience and adapting to new situations. Psychometricians noticed early that people who do well on one kind of mental test tend to do well on others — a positive manifold of correlations across tasks — and Charles Spearman proposed that this reflects a single underlying general intelligence, which he called g. On this view, specific abilities (verbal, spatial, numerical) each combine g with a task-specific component. Rival theories fragment intelligence into many parts: Howard Gardner's multiple intelligences (linguistic, musical, bodily-kinaesthetic and more) and Robert Sternberg's triarchic theory (analytical, creative, practical) both argue that a single number cannot capture the richness of human ability. For the OCR topic, the key point is that Van Leeuwen studies general IQ — a psychometric estimate of g — and so inherits both g's strength (it predicts a great deal) and its controversy (it may flatten what intelligence really is).
Definition — general intelligence (g). A single underlying factor proposed to explain why performance across diverse mental tests is positively correlated. An IQ score is an attempt to estimate a person's standing on g relative to their age group.
It is worth pausing on how an IQ score is actually constructed, because the number is less mysterious than it looks and its properties matter for evaluation. A test is given to a large, representative standardisation sample spanning the relevant ages, and the distribution of raw scores at each age is used to define the norm. An individual child's raw score is then expressed relative to that age norm and rescaled so that the population mean is 100 and the spread (standard deviation) is fixed — conventionally 15 points. This is why an IQ of 100 is exactly average by construction, and why roughly two-thirds of people score between 85 and 115: the scale is defined that way. Two consequences follow. First, IQ is a relative measure — it tells you where a child stands compared with age-mates, not an absolute quantity of "thinking stuff". Second, because the scale is anchored to a standardisation sample, a test can drift out of date as populations change, which is one reason the Flynn effect (discussed later) forces periodic re-norming. Understanding that IQ is a normed, relative, statistically constructed estimate of standing on g — rather than a direct readout of brain power — is the conceptual groundwork for reading Van Leeuwen fairly.
The biological area seeks the roots of intelligence in genes and the brain. Several strands of evidence suggest a biological contribution. Family studies show intelligence runs in families, though families share environments as well as genes. Twin studies — the design Van Leeuwen uses — exploit the fact that monozygotic (identical, MZ) twins share essentially 100% of their genes while dizygotic (fraternal, DZ) twins share on average 50%, like ordinary siblings; if MZ twins are markedly more similar in IQ than DZ twins, and their shared environments are comparable, the extra similarity points to genes. Adoption studies compare children's IQ with that of their biological versus adoptive parents. At the neural level, modest correlations have been reported between measures such as overall brain volume and IQ, and between the efficiency of neural processing and IQ, though these are correlational and far from deterministic.
The central biological concept is heritability. Heritability (often written h2) is the proportion of the variation in a trait, within a particular population and environment, that can be attributed to genetic differences among individuals. It is essential — and a favourite exam discriminator — to understand what heritability is not: it is not the proportion of an individual's intelligence that is "due to genes" (that question is meaningless, as height requires both genes and food); it does not mean a trait is fixed or unimproveable; and a heritability estimate is specific to the population and environment sampled, so it can differ between groups and change over time. With those cautions in place, twin and family studies consistently estimate the heritability of adult IQ as substantial — commonly in the region of half the variance, and often higher in adulthood than in childhood. Van Leeuwen's contribution is to test this with an unusually strong design.
| Design | What it exploits | Its limitation |
|---|---|---|
| Family study | IQ resemblance among relatives | Relatives share environment as well as genes |
| Twin study (MZ vs DZ) | MZ share ~100% genes, DZ ~50% | Assumes MZ and DZ shared environments are equally similar |
| Adoption study | Child vs biological vs adoptive parent | Selective placement; restricted range of adoptive homes |
| Twin-family study | Twins and their parents/siblings together | More complex modelling, but tests the twin study's assumptions |
Van Leeuwen and colleagues, working within the long-running Netherlands Twin Register, set out to estimate the heritability of general intelligence in a design that improved on the classic twin study. A standard twin study rests on the "equal environments assumption" — that MZ and DZ twin pairs experience equally similar environments — and on the assumption that there is no assortative mating (that parents do not pair up according to their intelligence) and no special "twin environment" effect. Their aim was to test general intelligence in a twin-family design that included not just twins but their parents and non-twin siblings, so that these assumptions could be checked rather than merely assumed, and to see whether a simple genetic model fitted the data.
The study used a family design incorporating twins, drawn from families registered with the Netherlands Twin Register. The sample comprised a large number of families — on the order of a couple of hundred — each containing a twin pair together with a sibling and both parents, giving several hundred individuals in total spanning two generations. Testing children, siblings and parents allowed the researchers to model the resemblance between many types of relative simultaneously (twin–twin, twin–sibling, parent–child, spouse–spouse). Because the design is a quasi-experiment/correlational family study — genetic relatedness is not manipulated but occurs naturally — it establishes association rather than causation in the strict experimental sense, but the modelling of multiple relationships is what gives it power.
Intelligence was measured with a standard, well-validated psychometric instrument — a Wechsler intelligence scale (age-appropriate versions: a children's scale for the younger participants and the adult scale for parents) — administered under controlled conditions. From the subtests, the researchers derived measures of general cognitive ability, focusing on general IQ (an estimate of g) rather than only specific subtest scores. Zygosity of the twins (whether MZ or DZ) was established by the standard register methods (questionnaire and, where needed, biological confirmation), because the whole design depends on correctly classifying twins as identical or fraternal.
The analytic procedure is the heart of the study. Using structural equation modelling (specifically, model-fitting in a genetic-covariance framework), the researchers partitioned the variance in general IQ into additive genetic (A), shared/common environment (C) and non-shared/unique environment (E) components, and tested how well models with different combinations of these fitted the observed pattern of resemblances among all the relatives. Crucially, including parents and siblings let them test for assortative mating (spouse–spouse correlation), for shared environmental effects, and for whether the genetic contribution behaved additively.
The central finding was that individual differences in general intelligence were substantially heritable. The best-fitting model attributed the large majority of the variance in general IQ to additive genetic factors, with the remainder accounted for mainly by non-shared (unique) environment; the shared (common) environment contributed little or nothing once the full family structure was modelled. In other words, once genes and each individual's own idiosyncratic experiences were accounted for, growing up in the same household explained very little of why family members differed in IQ. The heritability estimate for general intelligence in this design was high — a substantial majority of the variance — consistent with the upper end of the twin-study literature.
Two further results matter for evaluation. First, the researchers found evidence consistent with assortative mating for intelligence (a positive spouse–spouse resemblance): partners tended to be somewhat similar in IQ. This is important because unmodelled assortative mating can inflate DZ (and sibling) resemblance and thereby bias a classic twin study's heritability estimate — so testing for it is exactly the improvement the twin-family design was designed to deliver. Second, a simple additive genetic model fitted well, meaning the data did not demand more exotic genetic mechanisms to be explained.
| Variance component | Contribution to general IQ (this study) | Interpretation |
|---|---|---|
| Additive genetic (A) | Large majority | Genes are the main source of IQ differences here |
| Shared environment (C) | Little to none | The common family environment explained little variance |
| Non-shared environment (E) | Modest remainder | Each individual's unique experiences (and error) |
| Assortative mating | Present (positive spouse correlation) | Partners somewhat alike in IQ; must be modelled to avoid bias |
Van Leeuwen and colleagues concluded that general intelligence is highly heritable and that, in this Dutch sample, the shared family environment plays a negligible role in explaining differences in IQ once genetic relatedness and assortative mating are properly modelled. The finding of assortative mating underlines that classic twin studies which ignore it may misestimate heritability, and thus validates the more elaborate twin-family approach. The authors were careful — as you must be — to frame this as a statement about the sources of variation within this population, not a claim that any individual's intelligence is fixed by their genes or that environment is unimportant to intelligence in an absolute sense. Heritability being high does not preclude powerful environmental effects on the level of everyone's intelligence (the point behind the well-known secular rise in IQ scores over the twentieth century).
The study is a pivotal exhibit in the nature–nurture debate, and it is worth drawing the lesson out precisely. Van Leeuwen tilts the evidence towards nature for differences in intelligence: most of why children in this sample differed in IQ traced to their genes. But the sophisticated reading resists two temptations. The first temptation is to conclude that environment does not matter — a mistake, because heritability concerns variation within a given environment, and a strongly enriching or depriving environment could shift the level for everyone without showing up as "shared environment" variance. The second temptation is to treat "shared environment contributes little" as meaning parenting is irrelevant; more cautiously, it means the differences between these particular Dutch families did not translate into large IQ differences — a finding that may not hold across more environmentally unequal societies, where shared-environment effects on IQ are often larger. The honest conclusion is interactionist: genes account for much of the spread of IQ in a relatively uniform environment, but the environment sets the stage on which that genetic variation plays out.
Subscribe to continue reading
Get full access to this lesson and all 10 lessons in this course.