The population-level pattern
Smits and Monden analysed about 2.47 million births across 75 low- and middle-income countries. Average total twinning was 13.1 per 1,000 deliveries, with lower rates in much of Asia and rates above 18 per 1,000 in parts of Central Africa [1]. More recent genetic work summarizes spontaneous dizygotic rates of roughly 2 per 1,000 in East Asian populations and 20 per 1,000 in West African populations, while monozygotic twinning remains near 4 per 1,000 [2].
These are population observations, not individual genetic tests. Geography, ancestry, age, parity, nutrition, treatment access and data collection can overlap.
How model v2 represents the context
The calculator applies the following deliberately compressed weights to its dizygotic component:
- West African context: ×1.80
- European context: ×1.00
- Hispanic/Latina context: ×0.70
- South Asian context: ×0.60
- East Asian context: ×0.50
- Mixed/other: ×1.00
The old West African ×3.0 weight was reduced. The categories are coarse, self-reported and unable to represent mixed ancestry or within-group variation. Neutral is preferable when no category is a reasonable fit.
What the result cannot say
The field does not establish ethnicity as a biological cause, measure fertility, or predict an individual genotype. It should never guide clinical care. TwinCalc retains it only because the population gradient is large and repeatedly observed, while labelling its individual predictive value as low.