No single composite
The obvious move is to average PISA, spending, well-being and adult skills into one “Education Quality Score” and rank countries. We deliberately don’t. Poland’s whole story is that its standing diverges by life stage — top-tier at grade 4, above average at 15, near the bottom as adults. A blended score would mathematically erase exactly that divergence, the one non-obvious finding in the data.
Per-dimension normalization
Each indicator is scored against its own cross-country distribution for that assessment cycle, and Poland’s standing is free to move pillar to pillar. The honest magnitude is an SD-unit effect size on a single common comparator set:
z = (Poland − mean of countries) / SD of countries
computed on the common 27-country rich set — the countries that sat PIAAC 2023 and PISA 2022 and grade-4 TIMSS/PIRLS — so all three life-stage pillars share one population. This fixes two flaws of an earlier percentile version: it removes composition-inflation (grade-4 had been ranked against a broader, poorer field) and it uses an interval scale rather than a magnitude-discarding percentile. z is never summed across pillars into one number.
The lifecourse spine
| Pillar | Instruments | Poland (SD units) |
|---|---|---|
| A · Grade 4 | PIRLS 2021, TIMSS 2023 | +0.63 |
| B · Age 15 | PISA 2022 | +0.19 |
| C · Adults 16–65 | PIAAC 2023 | −1.23 |
Context pillars — resources & teachers, equity, lifelong learning, well-being, the excellence tail — explain the shape but are not folded into the spine. Innovation outcomes (unicorns, capital) sit in a separate layer outside the education score, because they are dominated by non-education confounds and folding them in would manufacture an “education-causes-unicorns” claim the data does not support.
Principles
- No citation → no claim. Every value carries source + cycle + URL; unknowns are explicit, never guessed.
- Level and direction. Above-average-and-falling is a different story from below-average-and-rising.
- Cross-sectional honesty. Grade 4, age 15 and adults are different cohorts in different years — not one generation tracked over time.
- Surface data-quality flags. Poland’s PIAAC 2023 result carries an official OECD caution; it travels with the number.