New Discoveries After TIMSS 2015 and PISA 2015: Math, Science, Readings & the Historic Impacts for Global Math Education and National Economies.

This document presents the original discoveries made by Dongchan Lee in January 2017, following the release of the TIMSS 2015 and PISA 2015 results. At the time, these findings were conservative estimates based on the best available international assessment data. Nearly a decade later, the post-pandemic PISA 2022 and TIMSS 2023 results have validated, strengthened, and in many cases surpassed the urgency of these original conclusions. The core discoveries are organized into five interconnected analytical frameworks: (1) THE MATH-GDP EXPONENTIAL NEXUS. Using normalized composite math scores from both TIMSS and PISA across 84 countries (with 12 systematic outliers removed), an exponential regression yields y = 439.09e^(0.0088x) with R² = 0.7314. This relationship demonstrates that math skills specifically — not merely average cognitive skills or years of schooling — are the dominant driver of GDP per capita variation across nations. When plotted on a logarithmic GDP scale, the "math chasm" between top performers (Singapore, Liechtenstein, Hong Kong, Korea, Japan, Switzerland) and bottom performers (Kyrgyzstan, Philippines, Morocco, Indonesia) becomes stark. The MMU1 intervention target of ~1.35 standard deviations would close more than half of this global math range. (2) THE MATH-READING GAP AS ECONOMIC PREDICTOR. All developed English-speaking countries and most Latin American countries exhibit stronger reading scores than math scores by large margins. The difference between math and reading scores explains income growth better than mean school years (R² ~ 0.25 in Hanushek & Woessmann's framework) when 3-6 outliers are excluded. Critically, this relationship strengthens with time lag: contemporaneous correlation is weak (R² ~ 0.085), but at 6-year lag it reaches R² = 0.446, suggesting the math-reading differential predicts future GDP rather than merely correlating with current income. As years progress, the relative strength of math over reading impacts GDP per capita with ~50-75% of the overall impact magnitude of average math scores alone. (3) THE STAGNATION DIAGNOSIS. PISA 2000-2015 and TIMSS 1995-2015 data reveal quasi-flat or declining trajectories across most developed nations. The largest math education collapse in 2015 affected 100% of Asian Tigers and most English-speaking nations. Years required to grow national math averages by 1 standard deviation range from 14 years (best case) to 188 years (worst case), with most countries requiring 42-120 years — far beyond any political or generational planning horizon. The technology-based education expansion during this period demonstrably failed to reverse stagnation. (4) PERCENTILE TRAJECTORIES AND MATH POVERTY. Detailed trajectory analysis for the United States, Australia, United Kingdom, Canada, and New Zealand shows declining scores across all percentiles (5th, 25th, 50th, 75th, 95th) from 2000-2015. The "red arrow" math chasm between top and bottom performers widened consistently. For the USA: average scores fell from 493 (2000) to 470 (2015); 5th percentile from 327 to 323; 25th percentile (Math Poverty threshold) from 427 to 408. Similar patterns hold across all English-speaking developed nations, with Latin American nations showing even steeper math poverty concentration. (5) MMU1 PROOF-OF-CONCEPT. Pilot studies conducted in Guatemala (El Alba private school, grades 3-5; ITEC of UVG, grade 1) during 2013-2016 demonstrated school-average gains of ~1.35 standard deviations — equivalent to raising performance from Guatemala F to the average of California or New York. This represents 70-160 years of normal national system improvement compressed into weeks of intervention. The MMU1 framework targets the math-poorest 25-50% of students first, using tablet-based, internet-delivered instruction powered by solar energy, complementing (not replacing) existing school teachers. POST-PANDEMIC VALIDATION (2022-2024): The PISA 2022 results (published December 2023) and TIMSS 2023 results (published December 2024) confirm that the "stagnation" identified in 2017 has become a full-blown decline. The OECD average math score fell by an unprecedented 15 points (~¾ year of learning), with 1 in 4 OECD 15-year-olds now low performers in all three subjects. School closures explained only ~11% of variation, while 65% of students report digital device distraction during math lessons — validating the 2017 conclusion that technology cannot solve systemic math education stagnation. No country has recovered to pre-2018 levels as of 2024. This document serves as the foundational stepping-stone for the 2026 MMU1/USL working paper series, which extends these discoveries with updated regressions, extended time-lag analyses, post-pandemic trajectory decomposition, GDP cost estimation, and a unified implementation framework for ending math poverty nationally and globally within 2-5 years.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.