by Lipton Matthews
Research on sex differences in intelligence is longstanding and empirically contested, yet discussing findings that favor men is often treated as taboo rather than science.
The controversy surrounding the nomination of E. J. Antoni to head the US Bureau of Labor Statistics briefly brought public attention to a topic that is usually confined to academic psychology. During discussions with interns at the Heritage Foundation, Antoni reportedly stated that men and women differ in intelligence distributions and that males are more likely to appear at the extremes, including very high IQ levels. The comments triggered criticism and media attention, and the nomination was eventually withdrawn. This was not the first time such ideas had cost someone a prestigious position. Larry Summers was similarly pushed out of Harvard after suggesting that inherent differences in ability between the sexes could account for why men tend to dominate at the top end of STEM performance.
Furthermore, conservative commentator Helen Andrews has recently argued that Summers’s departure marked an early landmark moment in the rise of what is now called wokeness. Yet the question Antoni raised has been studied for decades within psychometrics and cognitive science. One of the most prominent contributors to this debate was Richard Lynn, who developed a developmental theory proposing that average intelligence differences between males and females vary across the life course.
Lynn’s theory emerged partly from research on performance in tests designed to measure abstract reasoning, particularly Raven’s Progressive Matrices. These tests are widely used in psychometrics because they rely on pattern recognition rather than language or acquired knowledge. For this reason, they are often treated as relatively culture-free measures of the general intelligence factor, commonly referred to as “g”. Different versions of the test are used at different ages, including the Colored Progressive Matrices for younger children, the Standard Progressive Matrices for adolescents and adults, and the Advanced Progressive Matrices for high-ability samples. Because these tests load strongly on general intelligence, they have been widely used in attempts to measure potential cognitive differences between demographic groups.
For many years, the dominant view among psychologists was that males and females do not differ significantly in their average intelligence scores. Early standardization studies reported similar means for boys and girls through childhood and early adolescence. Later reviews frequently concluded that any differences appearing in individual samples were inconsistent or too small to interpret confidently. Researchers, therefore, argued that sampling variation or measurement error could easily produce small fluctuations in mean scores across studies.
Lynn challenged this interpretation by proposing that sex differences follow a developmental trajectory. According to this theory, boys and girls mature cognitively at different rates. During early childhood, the sexes perform at roughly the same level on general intelligence measures. In late childhood, girls may show a slight advantage due to earlier maturation. However, male development continues for a longer period during adolescence. By around the age of 16, the average difference reverses and begins to favor males, eventually stabilizing in adulthood at a modest but measurable level.
Empirical attempts to evaluate this hypothesis have relied heavily on meta-analysis. In one synthesis of Raven’s Progressive Matrices studies, researchers compiled data from numerous independent samples across different age groups. Strict inclusion criteria were applied so that only general population samples with adequate numbers of male and female participants were analyzed. Many earlier studies had small sample sizes, sometimes fewer than 100 individuals, which made it difficult to detect subtle differences. Statistical power calculations suggested that several hundred participants were required to reliably detect differences on the order of a few IQ points.
When these datasets were combined, the results broadly aligned with the developmental pattern proposed by Lynn. Among children aged roughly 6–14 years, boys and girls performed similarly on the Standard Progressive Matrices. However, beginning in mid-adolescence, a consistent male advantage emerged. Across adult samples, the difference averaged approximately 0.33 standard deviations, which corresponds to roughly 5 IQ points.
Evidence from other versions of the Raven test produced somewhat different patterns. Studies using the Colored Progressive Matrices for younger children showed a smaller male advantage, averaging around 0.21 standard deviations, or about 3.2 IQ points. Researchers suggested that part of this difference may reflect the cognitive composition of the test. Some items emphasize spatial visualization or pattern completion, domains in which previous research has often observed modest male advantages.
Large-scale meta-analytic work has also attempted to quantify the overall magnitude of sex differences across many cognitive tests. One analysis compiling more than 2,000 effect sizes from over 15 million individuals reported that adult males scored approximately 2.57 IQ points higher on average than females in general ability tests. Two of the three statistical approaches used in that analysis also supported the developmental hypothesis that the male advantage increases with age.
Another approach to evaluating the developmental hypothesis has been the use of latent variable modeling, which attempts to estimate the underlying g factor from multiple cognitive subtests. In this framework, researchers use statistical methods such as multi-group confirmatory factor analysis to test whether the general intelligence factor changes differently across sexes as individuals age. Analyses of longitudinal survey data show mixed but suggestive patterns. In two large datasets following individuals across adolescence and adulthood, estimates of the male advantage increased as participants aged. One model estimated that the difference rose from about 1.21 IQ points in early adolescence to approximately 5.53 points later in development, while another dataset showed an increase from roughly 0.18 to 2.46 points. However, alternative statistical models produced smaller or inconsistent differences, suggesting the importance of test composition and measurement assumptions.
These methodological differences illuminate an important challenge in the literature. Estimates of sex differences can vary depending on the statistical model used to extract the g factor. Higher-order factor models and bifactor models sometimes produce different latent means even when analyzing the same data. A meta-analysis of effect sizes drawn from multiple aptitude tests found a small but statistically significant male advantage of approximately 0.19 standard deviations, although the heterogeneity across tests was extremely high, indicating that results vary widely depending on the specific cognitive battery examined.
Moreover, biological correlates of intelligence have also been explored as possible contributors to these differences. Neuroimaging research consistently shows that male brains are, on average, larger than female brains by roughly 10–12 percent, a difference that remains after controlling for height and body size. Meta-analyses of brain imaging studies report a correlation of approximately 0.28 between brain volume and intelligence. Based on this relationship, researchers estimate that differences in average brain size could theoretically correspond to an intelligence gap between roughly 4.6 and 6.7 IQ points. Both verbal and performance IQ appear to correlate with intracranial volume at similar magnitudes, suggesting that the relationship between brain size and intelligence may operate at the level of the general factor.
Researchers emphasize that this relationship is probabilistic rather than deterministic. Larger brain size does not automatically produce higher intelligence, and cognitive performance depends on many neurological characteristics, such as connectivity and neural efficiency. Nevertheless, if brain size contributes causally to general intelligence, then systematic anatomical differences between males and females could influence average cognitive outcomes.
The literature has also addressed concerns that sampling biases might exaggerate male advantages. One hypothesis is that low-ability males may be underrepresented in typical research samples because men are disproportionately represented in populations that are difficult to study, such as prisoners or the homeless. Estimates suggest that prisoners have average IQ scores near 90 and homeless populations around 85. However, even if all such individuals were excluded from scientific samples, statistical adjustments indicate that the resulting bias would increase the estimated male advantage by only about one-tenth of an IQ point. This indicates that sampling bias of this type cannot account for the differences observed in cognitive test results.
Researchers have also noted that some cognitive abilities in which males tend to outperform females may be underrepresented in standard testing batteries. Spatial abilities such as three-dimensional mental rotation often show relatively strong male advantages, yet such tasks are not always included in general intelligence batteries. If these domains were more fully represented, the magnitude of male advantages might be somewhat larger than currently estimated.
Despite these findings, the literature does not present a unanimous consensus. Some latent variable analyses report negligible differences in general intelligence or even small female advantages depending on the statistical model used. In addition, men and women often show opposite advantages in specific abilities. Women tend to perform better in processing speed and certain verbal tasks, while men often score higher on visual-spatial processing. Because intelligence tests combine multiple subtests measuring different abilities, the resulting overall differences can depend heavily on the composition of the test battery.
Invariably, the empirical literature enunciates a complex pattern rather than a single definitive conclusion. Many studies of reasoning tests and meta-analyses of cognitive batteries report a small male advantage that becomes visible during late adolescence and adulthood, consistent with the developmental theory proposed by Lynn. Biological research on brain size and intelligence provides one potential mechanism that could contribute to such differences. At the same time, statistical modeling choices, test composition, and sample characteristics can influence the magnitude and even the direction of estimated effects.
Indeed, Antoni expressed a politically incorrect opinion on sex differences in intelligence. However, as explored earlier, this topic has been examined for decades and replicated in a number of empirical studies using large datasets and meta-analytic methods. Some of this research reports small average differences that favor men, particularly in late adolescence and adulthood. In scientific inquiry, such findings should not be inherently controversial. Research regularly reports results that paint women favorably in areas such as verbal ability, processing speed, or educational attainment, and these findings are typically discussed without difficulty.
The asymmetry arises when evidence suggesting male advantages is treated as illegitimate or beyond discussion. If empirical claims about sex differences are evaluated differently depending on which sex is favored, the result is not scientific neutrality but a form of anti-male bias that discourages open examination of evidence. Equally important is that the government and private sector would save resources if such beliefs became mainstream, because there would be no reason to expend resources on programs to promote women in STEM.
* * *
Lipton Matthews is a researcher and podcaster. His work has been featured in Mises, The Federalist, Chronicles, American Thinker, Epoch Times, and other publications. He is also author of Busting African Delusions: Institutions, Human Capital, and the Path to Progress.
Source: https://amgreatness.com/2026/05/14/the-heresy-of-sex-differences/
No comments:
Post a Comment