Intelligence from mother or father: the X chromosome claim against the family studies

0 The short answer

No evidence reviewed here shows that intelligence comes from the mother alone, and the family evidence points to both parents resembling their children to about the same degree. The viral claim needs three steps (intelligence genes are discrete, they sit on the X chromosome, so they come from the mother), and each step fails: a father passes his X chromosome to every daughter, the largest genome-wide study cited here links intelligence to 205 genomic loci, and the mouse imprinting experiments measured where cells ended up in the brain, not intelligence. In Bouchard and McGue's 1981 review of 111 studies, the weighted average IQ correlation was .41 for mother and offspring pairs and .41 for father and offspring pairs. Those are statements about groups of families, and they cannot say what either parent contributed to any one child.

1 Where Did the Claim That Intelligence Comes From the Mother Start?

The claim went viral in September 2016 as a news story, but the trail runs back to one blog post from March 2016 that contained no new research. The fact check by Alex Kasprak, published by Snopes on October 11, 2016 and read for this page on October 6, 2026, states the claim as "a person's intelligence is genetically determined by the mother alone" and rates it Unproven. Snopes reports that the factoid began to circulate after a September 14, 2016 article on a site called Second Nexus was reposted by notable accounts, including George Takei's Facebook page. That article's main source, Snopes says, was a March 2016 post on a blog called Psychology Spot, which cited 14 references. Of those 14, only one was published in the 2010s, in 2012, and it concerned maternal support in the first years of life, not brain genetics.

A second link in the chain was an article that appeared under the same byline on the Cosmopolitan and Good Housekeeping websites. Snopes reports that it opened by suggesting new research had been performed and then cited the same Psychology Spot post as its source, which makes the sourcing circular: the post supported the article and the article repeated the post. The press coverage followed the same pattern. IFLScience (Josh Davis, October 10, 2016, read October 6, 2026) described a story in The Independent that said researchers had found the intelligence genes on the X chromosome, of which a mother carries two and a man one. IFLScience noted that no new study was actually reported on, and that Forbes had already pointed out how little there was to go on.

Snopes reduces the argument behind these stories to three assumptions: that there are specific and discrete genes that universally determine intelligence, that those genes are located on the X chromosome, and that genes on the X must therefore come from the mother. The rest of this page tests each assumption against the primary literature. One limit applies from the start. The Psychology Spot post itself could not be retrieved for this page, so what it said is reported as Snopes describes it, and which mouse papers it cited is not something the fact check lists.

The claim is also not pure invention, which is why it is worth a long answer. Three real bodies of science sit underneath it. The X chromosome does carry many genes in which a damaging variant produces intellectual disability. Genomic imprinting, the silencing of one parent's copy of some genes, is real and does affect the developing mouse brain. And children resemble their parents in measured intelligence. What the stories did was to join those three facts with a conclusion none of them supports. The page on what heritability does and does not mean covers how much of the variation in IQ is genetic; this page is narrower and asks which parent, and through which mechanism, the evidence can speak to.

2 How Does X Chromosome Inheritance Work, and Does It Favor the Mother?

A mother passes an X chromosome to every child, but a father passes his X to every daughter, so the X chromosome does not belong to the mother's side of the family. Snopes quotes the genetic testing company 23andMe on the mechanics: one chromosome pair, the sex chromosomes, is unique, females typically have two X chromosomes and males an X and a Y, and mothers always pass an X to their children. Whether the father passes his X, producing a pair of X chromosomes, or his Y, producing a mixed set, determines the child's sex. A daughter therefore carries one X from each parent. A son carries a single X, and it comes from his mother. Snopes draws the consequence for the viral claim: for a daughter, any gene for intelligence that sat on the X would be just as likely to come from either parent's X.

The mother's own X chromosomes complicate the story further. A mother received one X from her mother and one from her father, so the X she hands on is not simply a "maternal" chromosome. IFLScience notes that chromosomes swap pieces during the formation of sex cells, in the process called crossover, which means that genes do not necessarily stay on any one chromosome and that the X a mother passes on can contain segments from both of her own parents. A statement that a gene on the X "comes from the mother" is true of a son's chromosome in the narrow sense of which parent delivered it, and misleading about whose DNA it ultimately is.

The real biology of the X concerns variability, not origin. Snopes quotes the 2009 paper by Wendy Johnson, Alison Carothers and Ian Deary in Perspectives on Psychological Science, which proposed that a harmful variant on one X in a heterozygous female is likely to be offset toward the population mean by the other X in other cells of her body, while a male has no such buffering because his single X is expressed in all cells. That argument was offered as one possible reason that males show greater variability in general intelligence than females, and the same authors, as Snopes quotes them, did not conclude that all genes for intelligence lie on the X. Their 2009 paper describes a model in which the population distribution of general intelligence is a mixture of two normal distributions and uses it to estimate the proportion of variance potentially due to X genes. It is a hypothesis about how the spread of scores is shaped, and it says nothing about a mother handing intelligence to her children.

Scientific interest in X-linked intellectual traits is older than the 2016 posts. A 1972 paper in the American Journal of Mental Deficiency, Lehrke's "Theory of X-linkage of major intellectual traits", appeared in volume 76, pages 611 to 619, and the same volume carries a reply by the author to commentary from Anastasi and from Nance and Engel. The title and the existence of a published exchange show that the idea was proposed and argued over in the research literature half a century ago. This page did not read the paper's argument, so it does not characterize it, and it has not verified that the 2016 posts drew on it. The point is simple, though: a hypothesis that was debated in 1972 is not "new research" in 2016, and a hypothesis is not a finding.

3 What Did the Mouse Imprinting Experiments Show?

The mouse experiment that fits the story's wording best showed that cells carrying only a father's genome and cells carrying only a mother's genome settle in different parts of the mouse brain, which is the opposite of a story in which only the mother's genes build the thinking brain. Genomic imprinting is the process by which certain genes are expressed from either the maternal or the paternal copy, according to the opening of the abstract of Keverne, Fundele, Narasimha, Barton and Surani (1996), published in Developmental Brain Research (volume 92, issue 1, pages 91 to 100). The team made chimeric mouse embryos, that is, embryos built from a mixture of normal cells and cells with artificial genomes, and followed where the artificial cells ended up. Cells with a duplicated paternal genome, called androgenetic, contributed substantially to hypothalamic structures and not to the cortex. Cells with a duplicated maternal genome contributed substantially to the cortex, striatum and hippocampus but not to the hypothalamic structures. Growth of the brain was enhanced by the maternal-genome cells and retarded by the paternal-genome cells.

Reading that result as support for "intelligence comes from the mother" skips four steps. First, the animals were mice. Second, the cells were engineered to carry one parent's genome only, a condition that does not occur in a normal embryo, where every cell has both. Third, what was measured was the fate of cells and the growth of brain regions, and no behavior or ability was tested for intelligence. Fourth, the authors proposed that imprinting may have facilitated a rapid expansion of the cortex over evolutionary time, which is a hypothesis about brain evolution, not a measurement of how well anyone thinks. The paper also shows both parental genomes doing necessary work, the paternal in the hypothalamus and the maternal in the cortex. It supports a model in which brain development needs both parents' genomes.

The nearest human evidence comes from Turner syndrome, in which a female has a single X chromosome. Skuse and colleagues (1997), in Nature (volume 387, issue 6634, pages 705 to 708), studied 80 females with Turner syndrome and a single X, 55 in whom the X was maternally derived and 25 in whom it was of paternal origin. The abstract states that intelligence is usually normal in Turner syndrome while social adjustment problems are common, and that the group with the paternal X was significantly better adjusted, with superior verbal and higher-order executive function skills. The authors read this as a genetic locus for social cognition that is imprinted and not expressed from the maternally derived X. Three cautions follow. The sample is a clinical group with a single X, so it cannot be applied to people in general. The outcome was social adjustment and specific skills, not IQ. And the direction of the finding, a locus that the maternal X does not express, is the reverse of the viral claim. The authors also speculated that such a locus could explain why males, whose only X is maternal, are more vulnerable to developmental disorders of language and social cognition, which is a proposal and not a result about intelligence.

4 What Does X-Linked Intellectual Disability Show, and What Does It Not?

X-linked intellectual disability shows that rare, large-effect variants on the X chromosome can cause disability, mostly in males, and that is a different question from why people differ across the ordinary range of scores. The strongest argument for an X chromosome role in intelligence comes from this literature, and it deserves a fair statement. Zechner and colleagues (2001), in Trends in Genetics (volume 17, issue 12, pages 697 to 701), reported that the incidence of mental disability is 30 percent higher in males than in females. They examined entries in the OMIM database, a catalog of genes and genetic conditions, for mental disability and several other common defects, and found that, compared with the autosomes (the chromosomes other than the sex chromosomes), the X chromosome contains a significantly higher number of genes that, when mutated, cause mental impairment. They proposed that these genes are involved in the development of cognitive abilities and so exert a large X chromosome effect on general intelligence. The operative word is proposed: the paper offers an inference from disability genes to the normal range, and it is the origin of the argument, not a measurement of it.

The disability side of the evidence is solid and still being refined. Piton, Redin and Mandel (2013), in The American Journal of Human Genetics (volume 93, issue 2, pages 368 to 383), note an unbalanced sex ratio in intellectual disability, 1.3 to 1.4 to 1, and the identification of large affected families showing X-linked segregation, and they report that mutations causing monogenic X-linked intellectual disability have been described in over 100 genes. They then used data on 10,563 X chromosomes from the National Heart, Lung, and Blood Institute Exome Sequencing Project to reassess 106 of those genes, questioned the involvement of ten of them, and flagged 15 more for replication. Snopes, for its part, reports that around 20 percent of the genes scientists associate with what it calls mental retardation are located on the X chromosome. The spread of those numbers is the point: even the list of X genes that cause disability is being pruned as larger reference data arrive.

Three limits keep this evidence from supporting the viral claim. First, a variant that breaks a gene and produces disability is a different kind of event from the common variation that separates one adult score from another, and the effects of the two kinds need not be related. Whether variation in X genes contributes to the normal range is a testable question, and the 2009 paper by Johnson, Carothers and Deary framed it that way: they modeled the proportion of variance in general intelligence that X genes could account for and, in the passage Snopes quotes, proposed targeting the X chromosome with chromosome-wide association studies because no candidate gene proposed up to then was on the X. Second, the X-linked segregation seen in large affected families describes specific diagnosed conditions and has no bearing on how common variants shape the typical spread of IQ scores in a population, which the page on the shape of the IQ distribution describes. Third, nothing in this literature says the mother's X chromosome is superior to the father's. A daughter receives an X from each parent, and Zechner's proposal, if it is right, would apply to both.

Families dealing with a diagnosed condition should speak with a clinical geneticist, and this page offers no medical or genetic advice. For readers who want the descriptive background on what scores in the low range do and do not mean, the page on what a low IQ is covers it without implying anything about parents.

5 What Do Family Studies Report for Mothers and for Fathers?

The largest early synthesis of family IQ studies found the same resemblance to mothers as to fathers, which is consistent with a trait carried on many chromosomes and shared in the home and is not what a mother-only trait would produce. The synthesis is Bouchard and McGue's 1981 review, "Familial Studies of Intelligence: A Review", in Science (volume 212, issue 4498, pages 1055 to 1059). It summarized 111 studies, 59 of them reported in the 17 years after the previous summary by Erlenmeyer-Kimling and Jarvik in 1963, and it yielded 526 familial correlations based on 113,942 pairings. The weighted average correlation between an individual parent and an individual offspring reared together was .42, based on 32 correlations. The authors then split the data by the sex of the parent, and the review's figures give the values in Table 1, as printed.

Table 1. Familial IQ correlations from Bouchard and McGue (1981), Figures 2 and 3, as the review prints them. Each value is a weighted average across studies.

The authors state the conclusion in words as well. The correlations are heterogeneous, they note, and the heterogeneity cannot be attributed to a sex effect, because opposite-sex and same-sex pairings give equivalent averages, or to a maternal effect, because the average correlation of mother and offspring was the same as that of father and offspring. The abstract adds that the heterogeneity is not moderated by the sex of the familial pairing or by the type of intelligence test used.

The sex-of-child rows allow a simple check of the X chromosome idea, and the arithmetic of inheritance is all that is needed to run it. A son cannot receive his father's X, so if differences in intelligence were carried on the X, a father's score and his son's score would share no X-based resemblance, while a father's score and his daughter's score would share a great deal. The review reports .38 for father and son and .39 for father and daughter. On the mother's side, a son's only X is his mother's, which under an X-linked account should make mother and son resemble each other at least as much as mother and daughter do, and the review reports .39 for mother and son against .43 for mother and daughter. These are differences of a few hundredths among heterogeneous correlations, and no row singles out the mother. This is our reading of the printed averages, not a claim made by Bouchard and McGue, and it does not prove which mechanism produced the resemblance. What it does show is that the data do not display the signature a mother-only X-linked trait would leave.

Two cautions belong beside the table. The review was published in 1981 and relies on the tests, samples and statistics of earlier decades, and its authors report that roughly half of the intergenerational correlations came from studies in which parents were given different tests from their offspring. And this page found no recent large sample reporting separate mother and father IQ correlations with their offspring's IQ that it could verify, so it does not substitute one; a newer source that does would be a welcome addition. The review is the source this page found that reports mothers and fathers in separate rows.

6 A Claim-by-Claim Check: What Each Piece of the Story Rests On

Taken one at a time, each piece of the viral story is false as stated, true of something other than intelligence, or a hypothesis that has not been demonstrated. The table sets the claims in the order the stories used them and gives, for each, what the opened source reports and how much weight it can carry. The strength column is our assessment of the documents named in the third column and is not a rating taken from any of them.

Table 2. The viral claim, step by step, against the primary sources. SNP means single nucleotide polymorphism, a one letter difference in DNA.

The table also shows where the honest uncertainty lies. Johnson and colleagues called in 2009 for chromosome-wide studies of the X, which shows it had not been searched in the way the autosomes were, and that is why the second row says untested at scale instead of refuted. This page did not find a large X-specific association result it could verify, and it does not claim one exists or does not exist. The claim that mothers pass on intelligence fails on the family data and on the mechanics of inheritance. The narrower claim that some X genes contribute to variation in intelligence remains an open scientific question, and the family correlations do not rule it out.

7 Why Can a Parent and Child Correlation Reflect the Home and Not Only Genes?

Parents give their children both a set of genes and a household, so a correlation between a parent's score and a child's score cannot, taken alone, say which one carried the resemblance. Both ingredients are visible in the 1981 review. Parents who did not rear their children still resembled them: the weighted average for single parent and offspring reared apart was .22, against .42 for those reared together. And adoptive parents, who share no genes with an adopted child, still correlated with that child, with a weighted average of .19 for adopting parent and offspring and .24 for adopting midparent and offspring. A purely genetic account would put the adoptive figure near zero, and a purely environmental account would put the reared-apart figure near zero, and the data show something in each. The review's averages cannot apportion the two, and adoptive placements are not random, so the adoptive figure may include selective placement as well as the effect of the home. That last caution is our reading.

The balance between the two also changes with age. Plomin and Deary's 2015 review in Molecular Psychiatry (volume 20, issue 1, pages 98 to 108) reports that the heritability of intelligence increases from about 20 percent in infancy to perhaps 80 percent in later adulthood. Heritability is a population statistic about the share of variation associated with genetic differences in a given sample, and the page on how much of IQ is genetic explains why it does not describe how much of one person's score is inherited. The relevance here is narrower: a parent and child correlation measured in early childhood is partly a measure of a shared household, and the same correlation measured in adulthood is partly something else.

A mother's score is also likely to travel with other features of a family, such as her education and income, and a child's score may be shaped by each of them. The page on breastfeeding and IQ treats the mother's own score as a confounder, which is the same trap in a different setting: whenever a mother's characteristic is correlated with a child outcome, the mother's genes, her behavior and her circumstances are tangled together. The page on birth order and IQ covers another family factor. None of these pages proves that the home matters more than genes or the reverse. They show that a bare parent and child correlation is the least informative design on the question, which is why researchers use adoption, sibling and genomic designs, taken up next.

8 What Do Adoption and Sibling Designs Separate?

Designs that compare people raised in different homes while holding genetic background closer to constant show that the rearing environment moves IQ scores by a few points on average, and they leave the genetic findings standing. The cleanest recent example is Kendler, Turkheimer, Ohlsson, Sundquist and Sundquist (2015), in Proceedings of the National Academy of Sciences (volume 112, issue 15, pages 4612 to 4617). Using a complete national Swedish sample of male-male siblings, the team identified 436 full-sibships in which at least one brother was reared by one or both biological parents and another by adoptive parents. IQ was measured at age 18 to 20 in the military conscription examination, and the education of the rearing parents was rated on a 5-point scale. Controlling for clustering of offspring within biological families, the adopted brothers had an IQ 4.41 points higher (standard error 0.75) than their non-adopted brothers, and each additional unit of rearing parental education was associated with 1.71 units of IQ (standard error 0.44).

The authors replicated the result in 2,341 male-male half-sibships, where adoption was associated with a gain of 3.18 IQ points (standard error 0.34) and each unit of rearing parental education with 1.94 IQ units (standard error 0.18). Their conclusion, in the abstract, is that using full and half-sibling sets matched for genetic background gave replicated evidence that rearing environment affects IQ measured in late adolescence, and that a portion of the adopted siblings' IQ could be explained by the educational level of their adoptive parents. A 4.41 point difference is about three tenths of a standard deviation on a scale with a standard deviation of 15, which is our arithmetic and a useful size to hold in mind, since the page on why the standard deviation is 15 explains how such differences are read.

The design earns its weight because it removes a major rival explanation. Brothers from one set of biological parents share genetic background on average, so a difference between a brother who stayed home and a brother who was adopted away is less likely to reflect the genes of their parents than a comparison between unrelated families would be. This is a group finding about Swedish men tested around the end of adolescence. It does not say what rearing does for a woman, for a person tested at another age, or for any particular family, and the abstract reports parental education without separating the mother's from the father's. It also does not say the home explains most of the variation. The review of family correlations earlier on this page showed both ingredients at work, and Kendler's design shows one of them cleanly: the family a child is raised in changes average scores in a measurable way, which is a different matter from the claim that the mother's genes decide them. For what deliberate effort can and cannot change in a score, the page on what can change an IQ score reviews the training evidence separately.

9 What Do Genome-Wide Studies Add?

Genome-wide studies find that intelligence is linked to a very large number of small genetic differences spread through the genome, which is the reverse of a few intelligence genes on one chromosome. Davies and colleagues (2011), in Molecular Psychiatry (volume 16, issue 10, pages 996 to 1005), analyzed 3,511 unrelated adults with data on 549,692 single nucleotide polymorphisms, which are one letter differences in DNA that are common in the population. They estimated that 40 percent of the variation in crystallized-type intelligence and 51 percent of the variation in fluid-type intelligence between individuals is accounted for by linkage disequilibrium between genotyped common SNP markers and unknown causal variants, and they describe these as lower bounds for narrow-sense heritability. They partitioned genetic variation by chromosome and found that, on average, longer chromosomes explain more variation, and using SNP data alone they predicted about 1 percent of the variance in an independent sample. Their conclusion is that the results are consistent with many genes of small effects. The methods section states that the relationship estimates came from autosomal SNPs, so the study did not test the X chromosome, which is a limit on what it can say about the X. The terms crystallized and fluid refer to the broad abilities Gc and Gf, which the page on fluid versus crystallized intelligence and the page on the CHC model of abilities define.

Savage and colleagues (2018), in Nature Genetics (volume 50, issue 7, pages 912 to 919), pooled 269,867 individuals. They identified 205 associated genomic loci, 190 of them new, and 1,016 genes, and they note that earlier genome-wide meta-analyses had identified 24 loci. The associated genes are strongly expressed in the brain, specifically in striatal medium spiny neurons and hippocampal pyramidal neurons, and gene set analyses implicate pathways related to nervous system development and synaptic structure. The jump from 24 loci to 205 as the sample grew is itself informative: the more people are studied, the more small-effect variants emerge, which is what a highly polygenic trait looks like. For a shorter account of what these findings mean for heritability, see the page on whether IQ is genetic; for what brain structure and function research adds, the page on intelligence and the brain reviews it.

Two points connect this evidence to the question of mother or father. First, because the genetic signal is spread across a very large number of variants, and because most chromosomes come in pairs that carry the same genes in different variants, as IFLScience puts it, a child receives one copy of each such chromosome from each parent. Under a polygenic account, then, both parents contribute variants to a child's score, and the two parents' contributions are expected to be of similar size. Second, neither abstract reports a parent-of-origin analysis, so these studies speak to which variants are associated with intelligence and not to which parent supplied them. They are evidence against the first two steps of the viral claim, the discrete genes and their location on the X, and they are silent on the third.

10 Do Mothers and Fathers Shape Children's Scores Differently Through Genes They Did Not Pass On?

Parents influence children through the genes they pass on and also through genes they carry but do not pass on, which act through the home, and a large study of that second channel found similar maternal and paternal contributions for educational attainment. Kong and colleagues (2018), in Science (volume 359, issue 6374, pages 424 to 428), call the second channel genetic nurture. Sequence variants in the parental genomes that are not transmitted to a child are often ignored in genetic studies. The team showed that nontransmitted alleles can affect a child through their impacts on the parents and other relatives. Using results from a meta-analysis of educational attainment, they found that the polygenic score computed for the nontransmitted alleles of 21,637 probands with at least one parent genotyped has an estimated effect on the educational attainment of the proband that is 29.9 percent of that of the transmitted polygenic score. The abstract states that paternal and maternal polygenic scores have similar effects on educational attainment, while mothers contribute more than fathers to nutrition and health-related traits.

The result bears on the viral claim in a specific way. A story in which the mother's contribution dominates would predict that maternal scores outweigh paternal scores for a cognitive outcome, and for educational attainment the abstract reports similar effects. Two cautions apply. The outcome was educational attainment, a social and academic outcome that correlates with intelligence but is not an IQ score, and the abstract reports that the genetic nurture effect extends to other traits without stating IQ among them. And the design infers an environmental channel from genotypes, so whether the nontransmitted effect reflects the home or other processes, such as the pairing of parents with similar traits, is a methodological question this page has not re-examined. The estimate is reported here as the 2018 paper's own.

The note about nutrition and health is worth keeping in view without over-reading. The abstract gives no reason for why mothers contribute more than fathers to those traits, and this page does not supply one. What matters for the question asked is the contrast: even in a design built to find parent-specific effects, the educational attainment results did not single out the mother. A reader who wants to see how parents' educational background shows up in average scores can start from the page on average IQ by education level, which covers the group pattern. Expectations are another proposed pathway from family to score, and the page on the Pygmalion effect reviews the evidence for that one.

11 Why Do Children of Very High Scoring Parents Score Closer to the Average?

Because parent and child scores correlate at about .4 to .5 and not at 1.0, families in which the parents score far from the mean produce children who, on average, score less far from it, and that regression toward the mean says nothing about which parent contributes more. The arithmetic is simple and is our own, applied to the published averages. Take the weighted average correlation of .42 between a single parent and an offspring reared together from the 1981 review, and assume for illustration that parents and children are scored on the same scale with a mean of 100 and a standard deviation of 15, and that the relationship is linear. Then children whose parent scored 130, thirty points above the mean, would average about 112.6, which is .42 times 30 added to 100. The review also reports a weighted average of .50 for midparent and offspring, where the midparent score is the average of the two parents. At that correlation, families in which the parents average 130 would have children averaging about 115, and families in which the parents average 70 would have children averaging about 85. These are averages across many families, and they describe the shape of a population relationship, not the score of any one family's children.

The spread around those averages matters as much as the averages. With a correlation of .50, the children within a group of families sharing one midparent score would still vary around their group average with a standard deviation of about 13 points (15 multiplied by the square root of 1 minus .25, again our arithmetic), only slightly less than the 15 points of the whole population. That is the statistical reason a parent's score cannot be turned into an expectation about a particular child, and the page on the shape of the IQ distribution shows how wide the normal range is. The review itself notes that .50 is substantially less than the simple genetic expectation of .707 for midparent and offspring and discusses why, including the fact that roughly half of the intergenerational correlations came from parents and offspring taking different tests. Test norms also age, which the page on the Flynn effect explains.

Parents also resemble each other, and that confounds any mother-versus-father comparison. Bouchard and McGue report a weighted mean assortative mating coefficient of .33, much smaller than the .50 sometimes reported in the literature. Plomin and Deary put the spousal correlation for intelligence at about .40 and contrast it with about .10 for personality and psychopathology and about .20 for height and weight, and they note that, because of positive assortative mating, children with highly intelligent mothers are also likely to have highly intelligent fathers. A mother's score therefore partly stands in for the father's, and a single-parent correlation cannot tell whose genes or whose home produced the resemblance. This is the main reason the sex-of-parent comparison is made within the same review, using many studies, instead of from one family or one dataset.

12 What Does the Evidence Support, Stated Narrowly, and How Should a Score Be Read?

The evidence supports three narrow statements about groups, and it does not support the claim that intelligence comes from the mother. First, in a review of 111 studies, parents and their offspring reared together correlated about .41 for mothers and about .41 for fathers on measured IQ, with adoptive parents correlating about .20 and .18, so both parents resemble their children and both a shared genetic background and a shared home appear in the numbers. Second, genome-wide studies link intelligence to a very large number of genetic variants, 205 loci in the largest meta-analysis cited here, so a child's genetic contribution comes from both parents across the genome and not from one chromosome. Third, the rearing environment matters on average: adopted-away Swedish brothers scored 4.41 IQ points higher than their home-reared brothers, and parental education at the rearing home was associated with higher scores. What the evidence does not support is the claim that the X chromosome is the intelligence chromosome, that imprinting makes the mother's genes decisive for the thinking brain, or that a father's contribution is smaller. It also says nothing about what any reader's own parents contributed, what any reader's score is, or what any child's score will be.

The way scores are interpreted is governed by professional standards that apply to the viral claim as much as to a test report. The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014), read for this page on October 6, 2026, state in Standard 1.0 that each intended score interpretation for a specified use should be clearly articulated and supported by appropriate validity evidence, and in Standard 1.2 that a rationale should be presented for each intended interpretation, together with a summary of the evidence and theory bearing on it. A statement that a score reflects the mother's genes is an interpretation, and a group correlation is not the evidence that supports it for an individual. Standard 6.10 adds that interpretations released to an audience should describe what the test covers, what scores represent, the precision or reliability of the scores and how scores are meant to be used. The APA Guidelines for Psychological Assessment and Evaluation, approved in March 2020 and read on October 6, 2026, state in Guideline 7 that individual performance on a test is one piece of an assessment, to be integrated with other sources of information. The pages on reliability and validity and on how IQ scores differ from percentiles explain why a score is read with its measurement error.

Applied to the topic of this page, those standards point to a simple practice. Treat family correlations as descriptions of populations. Do not convert a parent's score into a statement about a child's, or the reverse. And where a reader wants a measurement of their own, take an instrument that reports its score with a confidence interval and says what it can and cannot support. ACIS reports a Full Scale IQ and six indices on the standard scale, with percentiles and a 95 percent confidence interval, and states its limits plainly: it is online and unsupervised, it is not a clinical or diagnostic instrument, it is not for hiring, school accommodations or admission to high IQ societies, and it is in English only. It cannot say where a family's resemblance comes from, and no test can. The documentation is in the technical manual, and the page on what an IQ test measures and the page on how to choose a test are the places to start. The page on what intelligence is gives the definitions researchers work with.

13 Sources Behind This Page

Every figure on this page is traceable to one of the sources below, each linked in the body where the figure is first used. The 1981 review by Bouchard and McGue was read from an open copy of the original article, and the correlations in Table 1 come from the printed Figures 2 and 3. The fact check by Snopes and the article by IFLScience were opened on October 6, 2026. The 1972 paper by Lehrke is cited from its verified catalog record and was not read, so the page says only what its title and the published exchange show. The post on the Psychology Spot blog could not be retrieved, and what it said is reported as Snopes describes it. Arithmetic marked as ours (the expected child averages, the standard deviation of 13, and the three tenths of a standard deviation) is not attributed to any author.

• American Educational Research Association, American Psychological Association and National Council on Measurement in Education. Standards for Educational and Psychological Testing. AERA, 2014, Standards 1.0, 1.2 and 6.10, read October 6, 2026.
• American Psychological Association, APA Task Force on Psychological Assessment and Evaluation Guidelines. APA Guidelines for Psychological Assessment and Evaluation. Approved by the APA Council of Representatives, March 2020, Guideline 7, read October 6, 2026.
• Bouchard T J and McGue M. Familial studies of intelligence: a review. Science, 1981, volume 212, issue 4498, pages 1055 to 1059.
• Davies G, Tenesa A, Payton A, Yang J, Harris S E, Liewald D, Ke X, Le Hellard S and 24 others. Genome-wide association studies establish that human intelligence is highly heritable and polygenic. Molecular Psychiatry, 2011, volume 16, issue 10, pages 996 to 1005.
• Davis J. Nope, your intelligence probably is not solely inherited from your mother (title shortened here to avoid the contraction in the original headline). IFLScience, October 10, 2016, read October 6, 2026.
• Johnson W, Carothers A and Deary I J. A role for the X chromosome in sex differences in variability in general intelligence? Perspectives on Psychological Science, 2009, volume 4, issue 6, pages 598 to 611.
• Kasprak A. Intelligence is inherited only from your mother? Snopes, October 11, 2016, rating Unproven, read October 6, 2026.
• Kendler K S, Turkheimer E, Ohlsson H, Sundquist J and Sundquist K. Family environment and the malleability of cognitive ability: a Swedish national home-reared and adopted-away cosibling control study. Proceedings of the National Academy of Sciences, 2015, volume 112, issue 15, pages 4612 to 4617.
• Keverne E B, Fundele R, Narasimha M, Barton S C and Surani M A. Genomic imprinting and the differential roles of parental genomes in brain development. Developmental Brain Research, 1996, volume 92, issue 1, pages 91 to 100.
• Kong A, Thorleifsson G, Frigge M, Vilhjalmsson B, Young A, Thorgeirsson T, Benonisdottir S, Oddsson A and 7 others. The nature of nurture: effects of parental genotypes. Science, 2018, volume 359, issue 6374, pages 424 to 428.
• Lehrke R. Theory of X-linkage of major intellectual traits. American Journal of Mental Deficiency, 1972, volume 76, pages 611 to 619.
• Piton A, Redin C and Mandel J L. XLID-causing mutations and associated genes challenged in light of data from large-scale human exome sequencing. American Journal of Human Genetics, 2013, volume 93, issue 2, pages 368 to 383.
• Plomin R and Deary I J. Genetics and intelligence differences: five special findings. Molecular Psychiatry, 2015, volume 20, issue 1, pages 98 to 108.
• Savage J E, Jansen P R, Stringer S, Watanabe K, Bryois J, de Leeuw C A, Nagel M, Awasthi S and 109 others. Genome-wide association meta-analysis in 269,867 individuals identifies new genetic and functional links to intelligence. Nature Genetics, 2018, volume 50, issue 7, pages 912 to 919.
• Skuse D H, James R S, Bishop D V, Coppin B, Dalton P, Aamodt-Leeper G, Bacarese-Hamilton M, Creswell C, McGurk R and Jacobs P A. Evidence from Turner's syndrome of an imprinted X-linked locus affecting cognitive function. Nature, 1997, volume 387, issue 6634, pages 705 to 708.
• Zechner U, Wilda M, Kehrer-Sawatzki H, Vogel W, Fundele R and Hameister H. A high density of X-linked genes for general cognitive ability: a run-away process shaping human evolution? Trends in Genetics, 2001, volume 17, issue 12, pages 697 to 701.

14 Frequently Asked Questions