Is the Flynn effect on g?: A meta-analysis

Black/White differences in mean IQ have been clearly shown to strongly correlate with g loadings, so large group differences on subtests of high cognitive complexity and small group differences on subtests of low cognitive complexity. IQ scores have been increasing over the last half century, a phenomenon known as the Flynn effect. Flynn effect gains are predominantly driven by environmental factors. Might these factors also be responsible for group differences in intelligence? The empirical studies on whether the pattern of Flynn effect gains is the same as the pattern of group differences yield conflicting findings. A psychometric meta-analysis on all studies with seven or more subtests reporting correlations between g loadings and standardized score gains was carried out, based on 5 papers, yielding 11 data points (total N = 16,663). It yielded a true correlation of − .38, and none of the variance between the studies could be attributed to moderators. It appears that the Flynn effect and group differences have different causes. Suggestions for future research are discussed.     

Establishing an association between the Flynn effect and ability differentiation

The relationship between the Flynn effect and ability differentiation is investigated in a reanalysis of published data on Estonian student cohorts tested in 1933/36, 1997/98 and 2006 on the National Intelligence Test (Must, te Nijenhuis, Must, & van Vianen, 2009). To determine whether there was a relationship we computed the vector correlation between the Flynn effects (d) and the change in the g loading (Δg) between measurement occasions for each of the 10 NIT subtests and for each of the seven cohort comparisons, giving a total N of 70 effect sizes. The association between d and Δg was robustly negative (indicating that the Flynn effects were negatively associated with changes in the g loading of subtests) for all cohort comparisons, with values of r ranging from −.100 to −.461 (N = 10). When all effect sizes were analyzed together, the vector correlation was found to be −.281 (p ? .05, N = 70). This indicates a significant association between the Flynn effect and ability differentiation. Possible causes of this association are discussed.     

The Flynn effect in South Africa

This is a study of secular score gains in South Africa. The findings are based on representative samples from datasets utilized in norm studies of popular mainstream intelligence batteries such as the WAIS as well as widely used test batteries which were locally developed and normed in South Africa. Flynn effects were computed in three ways. First, studies where two different groups take the same test, with several years in between, using representative or comparable samples were used. Second, studies where the same group takes two different test batteries at a specific time were used. Third, the score differences between English- and Afrikaans-speaking Whites in South Africa in the 20th century were compared. The Flynn effect in White groups in South Africa is somewhat smaller than the Flynn effect in Western, industrialized countries (total N = 6534), and the Flynn effect in Indian groups is substantially smaller (total N = 682). Non-verbal IQ scores surpassed increases in verbal IQ scores. The findings from English- and Afrikaans-speaking Whites evidence a leveling out of differences in score gains over the 20th century (total N = 79,310). A meta-regression analysis showed no clear support for the moderators a) method used for computing the Flynn effect gain, b) type of test battery, c) time span, d) quality of the sample, and e) average age of sample.     

The secular rise in IQs in the Netherlands: Is the Flynn effect on g?

IQ scores have been increasing over the last half century, a phenomenon known as the Flynn effect. In this study, we focused on the question to what extent these secular gains are on the g factor. Two IQ batteries: the Interest-School achievement-Intelligence Test (ISI) and the Groningen Final Examination Primary Education (GALO) yielded small and modest negative correlations between standardized gains and g loadings. As these studies employ large samples this suggests that the combined literature now shows a modest negative relationship between d (the secular change in test score) and g.     

White–black differences in g and non-g effects for the SAT and ACT

This research examined g and non-g effects for the SAT and ACT for whites and blacks. SAT scores, ACT scores, and college GPAs were obtained from the National Longitudinal Survey of Youth. g was estimated using the Armed Services Vocational Aptitude Battery. Results indicated that (a) the g loadings of SAT and ACT composite scores were lower for whites than blacks, (b) group differences in the g loadings were related to the math subtests of the SAT and ACT, and (c) non-g variance accounted for surprisingly large percentages of SAT–GPA and ACT–GPA relations (range = 37–67%). The findings are discussed in terms of Spearman’s Law of Diminishing Returns.     

In the Netherlands the anti-Flynn effect is a Jensen effect

In this study, 63 observations of secular IQ changes (both Flynn and anti-Flynn effects) are collected from three demographically diverse studies of the Dutch population for the period 1975–2005 (representing the 1950–1990 birth cohorts), along with data on g loadings and subtest reliabilities. The method of correlated vectors is used to explore the association between Flynn and anti-Flynn effect magnitudes, both independently and together, and the g loadings of subtests. Despite a positive vector correlation the Flynn effects are not associated with the Jensen effect (r = .307, ns, N = 36), however the anti-Flynn effects are (r = .406, P = .05, N = 27). Combined, the vector correlation becomes negative but non-significant (r = −.111, ns, N = 63). Declines due to the anti-Flynn effect are estimated at −4.515 points per decade, whereas gains due to the Flynn effect are estimated at 2.175 points per decade. The N-weighted net of these is a loss of −1.350 points per decade, suggesting an overall tendency towards decreasing IQ in the Netherlands with respect to these cohorts. The Jensen effect on the anti-Flynn effect suggests that it may be related to bio-demographic changes within the Netherlands which have reduced ‘genetic-g’, despite the presence of large, parallel gains on subtests that may be relatively more sensitive to cultural-environmental improvements.     

The dependability of the general factor of intelligence: Why small, single-factor models do not adequately represent g

Floyd, Shands, Rafael, Bergeron and McGrew (2009) used generalizability theory to test the reliability of general-factor loadings and to compare three different sources of error in them: the test battery size, the test battery composition, the factor-extraction technique, and their interactions. They found that their general-factor loadings were moderately to strongly dependable. We replicated the methods of Floyd et al. (2009) in a different sample of tests, from the Minnesota Study of Twins Reared Apart (MISTRA). Our first hypothesis was that, given the greater diversity of the tests in MISTRA, the general-factor loadings would be less dependable than in Floyd et al. (2009). Our second hypothesis, contrary to the positions of Floyd et al. (2009) and Jensen and Weng (1994), was that the general factors from the small, randomly-formed test batteries would differ substantively from the general factor from a well-specified hierarchical model of all available tests. Subtests from MISTRA were randomly selected to form independent and overlapping batteries of 2, 4 and 8 tests in size, and the general-factor loadings of eight probe tests were obtained in each battery by principal components analysis, principal factor analysis and maximum likelihood estimation. Results initially indicated that the general-factor loadings were unexpectedly more dependable than in Floyd et al. (2009); however, further analysis revealed that this was due to the greater diversity of our probe tests. After adjustment for this difference in diversity, and consideration of the representativeness of our probe tests versus those of Floyd et al. (2009), our first hypothesis of lower dependability was confirmed in the overlapping batteries, but not the independent ones. To test the second hypothesis, we correlated g factor scores from the random test batteries with g factor scores from the VPR model; we also calculated special coefficients of congruence on the same relation. Consistent with our second hypothesis, the general factors from small non-hierarchical models were found to not be reliable enough for the purposes of theoretical research. We discuss appropriate standards for the construction and factor analysis of intelligence test batteries.     

Spearman’s Law of Diminishing Returns and national ability

This research examined Spearman’s Law of Diminishing Returns (SLODR) using national ability as the unit of analysis. National ability was estimated using international standardized tests such as the Programme for International Student Assessment (PISA), Trends in International Mathematics and Science Study (TIMSS), and Progress in International Reading Literacy Study (PIRLS). Factor analysis estimated the national G loadings of tests for high and low ability nations. Consistent with SLODR, the G loadings of tests were lower for higher ability nations. The pattern was confirmed after correcting for school attendance and age biases. Because a test’s g loading is directly related to its predictive validity (correlation with outcomes), our results imply that the predictive validity of tests may be lower for higher ability nations.     

A Jensen effect on dysgenic fertility: An analysis involving the National Longitudinal Survey of Youth

In this study we attempt to determine whether dysgenic fertility is associated with the Jensen effect. This is investigated with respect to a US population representative sample of 8110 individuals from the National Longitudinal Survey of Youth for whom there exists complete data on IQ and fertility. In addition to the general sample, the sample was also broken out by race and sex so as to examine whether or not the Jensen effect manifested amongst different sub-populations. The method of correlated vectors revealed significant Jensen effects in five of the seven samples, and in all cases the effect was in a direction indicating that subtests with higher g-loadings were associated with larger dysgenic fertility gradients. The magnitude of the difference between Spearman’s ρ and Pearson’s r was non-significant in all cases, suggesting that biasing factors were minimally influencing the result. This finding suggests that dysgenesis occurs on the ‘genetic g’ at the heart of the Jensen effect nexus, unlike the Flynn effect, which is ‘hollow’ with respect to g. Finally, the finding is discussed in the context of two converging lines of evidence indicating that genotypic IQ or ‘genetic g’ really has been declining over the last century.     

Jensen Effects and African/Coloured/Indian/White differences on Raven's Standard Progressive Matrices in South Africa

A test is made to determine whether various ethnic group differences on tests of cognitive performance in South Africa are like the Black/White differences in the United States in being positively associated with a tests' g loadings, where g is the general factor of intelligence. A non-parametric re-analysis is made of data from 1056 White, 1063 Indian, 778 mixed-race “Coloured,” and 1093 Black 14 year olds on the Raven's Standard Progressive Matrices Test in South Africa, given without time limits by Owen (1992) [Personality and Individual Differences, 13, 149]. The new analyses showed that the more highly correlated an item was with g, the more it predicted the White/Indian/Coloured/African differences on the test (Spearman's rhos from 0.35 to 0.85; all Ps<0.01). The effects remained regardless of which group g was extracted from. Understanding group differences around the world requires new research on the nature and nurture of g.     

Global behavioral variation: A test of differential-K

Despite Rushton’s path-breaking work into evolutionary forces affecting life history traits, not many attempts at operationalizing the differential-K spectrum at the level of countries or racial groups have been made so far. We report the construction of a “national K” factor from country-level behavioral variables. This K factor is closely related to country-level intelligence (“g”), operationalized by a composite score of IQ and scholastic achievement. We further demonstrate relationships of both g and K with measures of current environment and hypothesized evolutionary antecedents. Whereas K is predicted most powerfully by intelligence, log-transformed GDP (lgGDP) and skin reflectance, g is predicted by skin reflectance, lgGDP, cranial capacity, and a measure of evolutionary novelty.     

Were the Victorians cleverer than us? The decline in general intelligence estimated from a meta-analysis of the slowing of simple reaction time

The Victorian era was marked by an explosion of innovation and genius, per capita rates of which appear to have declined subsequently. The presence of dysgenic fertility for IQ amongst Western nations, starting in the 19th century, suggests that these trends might be related to declining IQ. This is because high-IQ people are more productive and more creative. We tested the hypothesis that the Victorians were cleverer than modern populations, using high-quality instruments, namely measures of simple visual reaction time in a meta-analytic study. Simple reaction time measures correlate substantially with measures of general intelligence (g) and are considered elementary measures of cognition. In this study we used the data on the secular slowing of simple reaction time described in a meta-analysis of 14 age-matched studies from Western countries conducted between 1889 and 2004 to estimate the decline in g that may have resulted from the presence of dysgenic fertility. Using psychometric meta-analysis we computed the true correlation between simple reaction time and g, yielding a decline of − 1.16 IQ points per decade or − 13.35 IQ points since Victorian times. These findings strongly indicate that with respect to g the Victorians were substantially cleverer than modern Western populations.     

Overview of the Flynn effect

Following WW2, various researchers found and reported secular gains in IQ, but it was not until additional reports appeared in the 1980s that researchers began to look for the cause or causes. It was quickly apparent that the gains were not limited to any group or nation, but the manifestation of the gains was different depending on time and place. For every discovery, there was a different or opposite result in a different data set. Gains have been large, small, variable, and even negative. Some researchers have found that the gains were on g, while more have found no g loading. Abstract test formats, such as the Raven have often shown the greatest gains, but gains have also appeared in tests of crystallized intelligence. Some data has shown greater gains for the lower half of the intelligence distribution, while others have shown greater gains in the top half, and others have shown equal gains at all levels. Hypotheses for the causes have included environmental factors, genetic effects, reduced fertility, and methodological dependence. Two models are discussed.     

Improved matrix reasoning is limited to training on tasks with a visuospatial component

Recent studies (e.g.,  and ) have provided evidence that scores on tests of fluid intelligence can be improved by having participants complete a four week training program using the dual n-back task. The dual n-back task is a working memory task that presents auditory and visual stimuli simultaneously. The primary goal of our study was to determine whether a visuospatial component is required in the training program for participants to experience gains in tests of fluid intelligence. We had participants complete variations of the dual n-back task or a short-term memory task as training. Participants were assessed with four tests of fluid intelligence and four cognitive tests. We were successful in corroborating Jaeggi et al.'s results, however, improvements in scores were observed on only two out of four tests of fluid intelligence for participants who completed the dual n-back task, the visual n-back task, or a short-term memory task training program. Our results raise the issue of whether the tests measure the construct of fluid intelligence exclusively, or whether they may be sensitive to other factors. The findings are discussed in terms of implications for conceptualizing and assessing fluid intelligence.     

Rushton’s contributions to the study of mental ability

This essay describes Rushton’s contribution to examining the nexus of intelligence, race, and genetics, specifically what I termed “Spearman’s hypothesis”. It states that Black–White differences are “most marked in just those [tests] which are known to be saturated with g”. I (Jensen) had confirmed this hypothesis using large data sets in the 1970s and 1980s and also found that Black–White differences were most marked on the more heritable rather than the more cultural subtests. Rushton confirmed and extended these findings in many highly innovative ways and demonstrated Spearman’s hypothesis applied among samples of Gypsy Roma in Serbia, and East Asian, European, South Asian, Colored and Black samples in South Africa. He has not only documented group differences in brain size, intelligence, life span, family structure, infant mortality, developmental precocity, personality, and temperament, and rates of two egg twinning, and crime among East Asians, Europeans, and Africans, but also provided a life history theory that explains them.     

Expanding the g-nexus: Further evidence regarding the relations among national IQ, religiosity and national health outcomes

The current study seeks to better understand how religiosity and health are positioned within the g-nexus. Specifically, the degree to which differences in average IQ across nations is associated with differences in national religiosity (i.e., belief rate) and national health statistics independent of differences in national wealth is examined. Consistent with expectations, results show that, independent of national wealth and belief rate, IQ has a positive influence on national health as indicated by fertility rate, infant mortality rate, maternal mortality rate, and deaths due to HIV/AIDS, and life expectancy. Additionally, as hypothesized, IQ and belief rate interacted to influence reproductive health (i.e., fertility rate, infant- and maternal mortality). Specifically, high IQ acts as a buffer against the negative effects of belief rate; when IQ is high belief rate has no effect, but when IQ is low belief rate has a strong negative effect. The pattern of findings from this study, combined with previous research, serve to confirm that general cognitive ability (i.e., the g-factor) is an important and central node within a larger nexus of psychological and social variables. Theoretical and epidemiological implications are discussed.     

Associations between cortical thickness and general intelligence in children,adolescents and young adults

Neuroimaging research indicates that human intellectual ability is related to brain structure including the thickness of the cerebral cortex. Most studies indicate that general intelligence is positively associated with cortical thickness in areas of association cortex distributed throughout both brain hemispheres. In this study, we performed a cortical thickness mapping analysis on data from 182 healthy typically developing males and females ages 9 to 24 years to identify correlates of general intelligence (g) scores. To determine if these correlates also mediate associations of specific cognitive abilities with cortical thickness, we regressed specific cognitive test scores on g scores and analyzed the residuals with respect to cortical thickness. The effect of age on the association between cortical thickness and intelligence was examined. We found a widely distributed pattern of positive associations between cortical thickness and g scores, as derived from the first unrotated principal factor of a factor analysis of Wechsler Abbreviated Scale of Intelligence (WASI) subtest scores. After WASI specific cognitive subtest scores were regressed on g factor scores, the residual score variances did not correlate significantly with cortical thickness in the full sample with age covaried. When participants were grouped at the age median, significant positive associations of cortical thickness were obtained in the older group for g-residualized scores on Block Design (a measure of visual-motor integrative processing) while significant negative associations of cortical thickness were observed in the younger group for g-residualized Vocabulary scores. These results regarding correlates of general intelligence are concordant with the existing literature, while the findings from younger versus older subgroups have implications for future research on brain structural correlates of specific cognitive abilities, as well as the cognitive domain specificity of behavioral performance correlates of normative gray matter thinning during adolescence.     

IQ, skin color, crime, HIV/AIDS, and income in 50 U.S. states

In 50 U.S. states, we found a positive manifold across 11 measures including IQ, skin color, birth rate, infant mortality, life expectancy, HIV/AIDS, violent crime, and state income with the first principal component accounting for 33% of the variance (median factor loading = .34). The correlation with a composite of total violent crime was higher with skin color (r = .55), a more biologically influenced variable than with GDP (r = −.17), a more culturally influenced variable. These results corroborate and extend those found at the international level using INTERPOL crime statistics and at the county, provincial, and state levels within countries using local statistics. We interpret the cross-cultural consistency from an evolutionary life history perspective in which hierarchically organized traits culminate in a single, heritable, super-factor. Traits need to be genetically organized to meet the trials of life—survival, growth, and reproduction. We discuss brain size and the g nexus as central to understand individual and group differences and we highlight melanin and skin color as a potentially important new life history variable.     

Equivalency and reliability of vectors of g-loadings across different methods of estimation and sample sizes

We examine the variability in g-loadings with particular focus on the differential stability (i.e., rank order inconsistencies) of the vector of loadings due to sample size and method of estimation. Our results, based on 1600 factor-analyses, show that different methods of estimation require different sample sizes to achieve the same level of stability in loadings. However, all methods examined yielded reliable vectors when sample sizes of N = 400 were used. Some methods yielded reliable vectors with sample sizes of N = 200. Such results provide guidelines for choosing an appropriate sample size when g-loadings are of interest.     

An integrated perspective on the relation between response speed and intelligence

Research in the field of mental chronometry and individual differences has revealed several robust regularities (Jensen, 2006). These include right-skewed response time (RT) distributions, the worst performance rule, correlations with general intelligence (g) that are more pronounced for RT standard deviations (RTSD) than they are for RT means (RTm), an almost perfect linear relation between individual differences in RTSD and RTm, linear Brinley plots, and stronger correlations between g and inspection time (IT) than between g and RTm. Here we show how all these regularities are manifestations of a single underlying relationship, when viewed through the lens of Ratcliff’s diffusion model ( [Ratcliff, 1978] and [Ratcliff et al., 2008] ). The single underlying relationship is between individual differences in general intelligence and individual differences in “drift rate”, which is just the speed of information processing in Ratcliff’s model. We also test and confirm a strong prediction of the diffusion model, namely that the worst performance rule generalizes to phenomena outside of the field of intelligence. Our approach provides an integrative perspective on intelligence findings.