About the estimates
The figures on this map come from Sebastian Jensen and Emil Kirkegaard's 2024 paper National IQs and Socioeconomic Development. The authors pooled several sources of cognitive-test performance (earlier national IQ compilations by Lynn & Vanhanen, Rindermann and Becker; recent PISA, TIMSS and PIRLS results; the Basic Skills Dataset; and World Bank test-score data) and combined them into a single estimate per country on the familiar IQ scale (mean 100, SD 15).
Two features distinguish this dataset from earlier compilations such as Lynn & Vanhanen (2002, 2012) and Lynn & Becker (2019). First, it uses no geographic imputation: a country receives a score only if test data from that country exist. Second, combining sources reduced the average estimated standard error of national IQs from 5.41 to 2.58 points. The standard error for each country is shown in the tooltip; where it reads 5.41, the estimate rests on a single source and should be treated with more caution.
In the paper, national IQs correlate at r = .88 with a 47-indicator index of socioeconomic development. The authors also review measurement-invariance studies and conclude that cross-national bias is uncommon outside of verbal tests.
Where the 2024 dataset has no estimate, the map falls back to Lynn & Vanhanen's IQ and the Wealth of Nations (2002) when that book has a figure. These areas are striped and labelled in the tooltip, and they are left out of the ranking and all averages. Only one country is affected, the Federated States of Micronesia (84). Lynn & Vanhanen imputed that value from neighbouring countries rather than measuring it. The remaining grey areas are territories and micro-states that neither source covers.
Reading the map responsibly
National averages describe populations, not individuals: the variation within any country is far larger than the gaps between most countries. Estimates for low-income countries often rest on smaller or less representative samples, and test scores reflect schooling, health, nutrition and familiarity with testing as well as underlying ability. The use of national IQ datasets remains contested in the academic literature; the critiques linked below are worth reading alongside the data.
Layers
Religious-majority layers use Pew Research Center's 2020 estimates of religious composition. A country is counted as Muslim-majority or Christian-majority when more than 50% of its population belongs to that group. For eleven micro-states not covered by Pew (e.g. Liechtenstein, Palau, Nauru), majority affiliation follows the CIA World Factbook. Group means in the layer panel are simple unweighted averages across countries.
The High modern slavery layer uses Walk Free's Global Slavery Index 2023, which estimates the share of each country's population living in modern slavery (forced labour and forced marriage) in 2021. A country is included when the estimate is at least 10 per 1,000 people, i.e. 1% of the population. That covers 23 countries, led by North Korea (104.6), Eritrea (90.3) and Mauritania (32.0). Every country's figure appears in its tooltip. The index covers 160 countries; small states and territories without an estimate never fall into this layer.