Science & Methodology

Scientific practice, replication, and epistemology.

The Heritability of Intelligence: What Twin and Adoption Studies Show

nonacademicresearch.org Editorial · May 10, 2026 · submitted by nonacademicresearch.org Editorial · nar:iirsia2gp6cwmapy52

Twin and adoption studies consistently find that genetic factors explain a substantial portion of variation in measured cognitive ability — heritability estimates of 50–80% in adults. This finding is frequently misunderstood: heritability describes variation within a population, not the proportion of intelligence determined by genes versus environment, and does not imply that group differences in test scores are genetic in origin. Environmental interventions can and do substantially improve cognitive outcomes. The heritability of intelligence is real, important, and consistently misapplied in public debates about inequality and education policy.

Do Economic Forecasts Work? The Evidence on Prediction Accuracy

nonacademicresearch.org Editorial · May 9, 2026 · submitted by nonacademicresearch.org Editorial · nar:2xd6jplbm6iib7jyx3

Professional economic forecasters — central banks, international institutions, private forecasters — demonstrate limited accuracy beyond short time horizons, particularly for recessions. Forecasters systematically fail to predict downturns in real time, show herding behavior, and underperform simple benchmark models at horizons beyond one or two quarters. This is not a failure of effort or intelligence; it reflects fundamental limits on predictability in complex adaptive systems.

The Replication Crisis: What Failed to Replicate and Why

nonacademicresearch.org Editorial · May 9, 2026 · submitted by nonacademicresearch.org Editorial · nar:qcetlew5tjfj0fyhgl

Large-scale replication projects in psychology found that approximately 36–62% of published findings failed to replicate at conventional significance levels, with average effect sizes roughly half those originally reported. This is not a crisis unique to psychology — similar patterns have been documented across medicine, economics, and cancer biology. The causes are structural: publication bias, small samples, undisclosed analytical flexibility, and the statistical consequences of testing many hypotheses.