From Intelligence:

International differences in math and science tilts: The stability, geography, and predictive power of tilt for economic criteria

Volume 92, May–June 2022, 101646

DavidBecker, Thomas R.CoyleTyler, L.Minnigh, Heiner Rindermann

Highlights
• Math/science tilt was based on PISA results from 2000 to 2018 in 86 countries.

• Math/science tilt was computed by contrasting math and science results with reading results.

• Math/science tilt patterns were stable over time, cohorts, testing waves, and methods.

• East Asian countries showed math/science tilt; Western countries showed verbal tilt.

• Math (not science) tilts predicted economic growth (β = .25 to .45 vs .08 to .18).

Abstract
Tilt represents an ability pattern and is based on differences in two dimensions (e.g., math and reading), yielding strength in one dimension (math) and weakness in the other (reading). Whereas prior research examined tilt relations with economic productivity at the individual or country levels, the current study is the first to examine the stability, geography, and predictive power of math, verbal, and science tilts for economic growth at the country level. Tilt was based on math, science, and reading results in seven PISA (Programme for International Student Assessment) waves from 2000 to 2018 (Nmax = 86 countries). Tilt was computed by contrasting math or science scores with reading scores, yielding math tilt (math > reading) and science tilt (science > reading). Tilt patterns were stable across the seven PISA waves (mean stabilities math-reading r = .74 and science-reading r = .53). Tilt showed a geographic gradient, with math/science tilt in East Asia and verbal tilt in Europe and the Americas (r = .38 to .42, for tilt and geographic coordinates). Tilt did not mediate relations between country level ability and economic growth. However, math tilt (math > reading) directly and positively predicted higher economic growth (β = .25 to .45), whereas science tilt (science > reading) was negligibly related to economic growth (β = .08 to .18). The results were insensitive to the year of data collection and the parametric procedure used. Future research should consider factors that influence tilt such as differential investment of educational resources.

Sounds plausible, but I’m guessing there isn’t too much within continent variance: just East Asians being good at math.

[Comment at Unz.com]