Policy should focus on treating individuals as individuals rather than group members, stopping use of racial group statistics to judge schools and institutions, and assessing whether individuals have fair opportunity to develop capabilities, regardless of race.

The Flynn Effect (upward drift of IQ scores across generations) is qualitatively different from Black-White IQ differences in the US: Dutch psychometricians using complex methodologies found that the Flynn Effect appears to operate on different factors than those driving the Black-White gap, and therefore the Flynn Effect is unlikely to automatically close the racial gap.

Height of average Japanese has increased by over one standard deviation in 40 years due to improved nutrition, and while residual difference from Caucasians persists and is unlikely to diminish further, this demonstrates that large environmental effects on heritable traits are possible.

The Black-White IQ gap closure has stalled because shared environment (basic education, nutrition, family socioeconomic status) has improved for Black children born in the 1980s and 90s compared to the 1960s, yet the gap did not decrease, suggesting shared environment improvements have exhausted their effect on closing the gap.

Spearman's Hypothesis states that the Black-White IQ difference is largest on tests most saturated with g (general intelligence factor), and this has been established empirically by Arthur Jensen and others as a fact, not merely a hypothesis, with implications that g-based differences may have genetic origins since g is highly heritable and correlates with physiological measures like frontal cortex gray matter volume.

Adoption studies show that while environment can have huge effects when children are severely deprived (locked in closets, not spoken to), once a nearly adequate environment is reached, environmental improvements produce minimal additional IQ gains, suggesting diminishing returns on environmental interventions.

Reaction time tests (measuring cognitive processing speed vs. physical motor speed) show consistent Black-White differences in reaction time but Blacks perform better on movement time, suggesting motivation is not a primary factor in test performance differences.

General Social Survey vocabulary test (10 items, unchanged for 20+ years) shows Black-White gap has been declining, but primarily because White performance has declined, not because Black performance has improved, raising questions about what is being measured by test score changes.

Black and North African occupation soldiers in Germany may have differed genetically (North Africans historically distinct from sub-Saharan populations), and inclusion of ~23% North African-fathered children in the Black sample potentially biases Eyferth results, though this confounding is difficult to untangle.

If the Eyferth study was the only example showing environmental causation of the gap, and no similar results appear in adoption studies, other countries, or other time periods, then relying on this single study violates the 'rule of one' in social science where single anomalous findings (like Perry Preschool) are over-cited.

Genetic evidence could be advanced through large-scale genetic marker studies that go beyond binary Black/White classification to measure continuous genetic ancestry, combined with cognitive and physical appearance measures, which would help separate genetic from physical discrimination effects.

Vocabulary acquisition depends on the kind of mind that, upon hearing a word, wonders about its meaning and naturally fits it into sentence context, rather than mere exposure to words, suggesting children from less-educated homes aren't simply hearing fewer words but lack cognitive habits that develop vocabularies.

Both Flynn and Murray demonstrate unusually high intellectual collegiality and good faith in disagreement compared to typical public discussions of race and IQ, with both believing the other is arguing in good faith despite fundamental disagreements about data interpretation and causation.

Vocabulary as a Wechsler subtest was nearly excluded from IQ batteries in the 1930s because psychometricians assumed it was too culturally determined to measure true intelligence, but later evidence showed vocabulary correlates with cognitive ability despite cultural loadedness.

Perry Preschool program is the archetypal example of the 'rule of one' in social science: a small-sample intervention (40 kids control, 40 experimental) showing positive effects, widely cited despite hundreds of other preschool interventions showing no effect, because it supports the desired conclusion.

Income does not drop out of multivariate equations predicting offspring IQ when controlling for other variables, showing an independent relationship between parental income and children's IQ, though the relationship is not huge.

NLSY data (National Longitudinal Survey of Youth) provides the best data for gap closure assessment because: large samples (Black N > 2,500), identical tests over time, most children tested multiple times, sample not limited to school-attending children, and includes both academic achievement and verbal IQ measures plus detailed background information including mother's IQ.

State-administered standardized tests under NCLB face severe penalties if proficiency targets aren't met, creating incentive to inflate proficiency measures and deflate reported gaps, making state-level data unreliable for tracking whether gaps are actually narrowing.

Proper educational policy should enable every child to develop their capabilities to the maximum regardless of race, and the No Child Left Behind Act—by imposing a one-size-fits-all solution that might work for worst inner-city systems onto all schools—was disastrous for American education and should be repealed.

The Black-White gap on standardized tests has substantial sample size limitations: WISC standardizations have 100-300 Black samples (in some cases 135 for children ages 2-6, less than 200 for Stanford-Binet), making strong inferences about group means unreliable because the samples are designed to represent national distributions, not to measure precise group differences.