Source Priority
EduScores gives priority to primary PISA material published by the OECD.
The preferred source order is:
- official PISA databases and downloadable datasets;
- statistical annex tables published with PISA results;
- official PISA reports and technical documentation;
- OECD data tools and country or economy notes;
- other official education authority material when additional context is necessary.
Secondary summaries are not used as substitutes when the underlying official data is available.
Data Verification
A figure is checked against the source appropriate to the metric being reported.
Mean scores, proficiency shares, subgroup results, participation information and trend values are not assumed to come from the same table or comparison population.
Where different OECD publications use different country sets, matched comparison groups or trend definitions, the relevant official definition takes priority over a simple subtraction of headline values.
This is particularly important for comparisons between assessment cycles. An OECD trend series may use a consistent group of countries that differs from the countries included in a headline average for an individual year.
Published and Derived Values
EduScores distinguishes between values reported directly by a source and values calculated from source data.
Derived values can include:
- score-point differences between two cycles;
- gaps between a country result and an OECD benchmark;
- percentage-point changes;
- subgroup differences;
- ranges and distribution spreads;
- matched-country comparisons;
- other descriptive calculations based on published values.
Derived figures are presented as calculations based on the underlying data and are not described as separately published OECD statistics.
Rounding can cause a calculated difference to vary slightly from a change reported directly by the OECD. When an official trend figure is available, that figure takes precedence.
Rankings
Rankings are generally based on reported mean scores for the selected subject, year and participating set.
PISA results include sampling and measurement uncertainty. A one-point or several-point difference does not necessarily establish a statistically significant difference between two systems.
A displayed rank should therefore be read as an ordering of reported estimates rather than a precise league table.
Where the OECD provides confidence ranges for rankings or statistical significance tests, those measures provide more context than the displayed numerical position alone.
Statistical Significance
Statements about increases, declines or differences are described as statistically significant only when the underlying OECD material supports that interpretation.
A numerical change obtained by subtracting two rounded means is not automatically described as a confirmed increase or decline.
Where significance information is unavailable, the wording remains descriptive.
OECD Averages
The term OECD average can refer to different comparison groups depending on the publication, metric and assessment cycle.
Trend analysis may use a consistent OECD comparison set rather than the complete membership represented in a single-cycle headline average.
Pages identify the relevant benchmark when this distinction affects interpretation.
Proficiency Levels
PISA proficiency levels describe performance ranges on an assessment scale.
The share of students below Level 2, at or above Level 2, or at Levels 5 and 6 provides information that cannot be obtained from a mean score alone.
Mean scores and proficiency distributions are treated as separate measures. Two systems with similar averages can still have different shares of low-performing and high-performing students.
Student Groups and Background Indicators
Results involving gender, socio-economic background, immigrant status, language, digital behaviour or other student characteristics are reported according to the definitions used in the relevant PISA documentation.
These comparisons are descriptive unless the evidence supports a stronger conclusion.
An association between a student characteristic and a PISA outcome does not by itself establish that the characteristic caused the observed difference.
Country and Economy Names
EduScores generally uses the naming conventions found in the relevant PISA release.
Some PISA participants are economies, territories, regions or specific jurisdictions rather than sovereign countries. Their inclusion reflects the reporting structure of PISA and is not a statement about political or legal status.
Missing and Non-Comparable Data
A blank observation, missing symbol or unavailable result is not treated as zero.
Data may be unavailable because a system did not participate, a metric was not collected, reporting standards were not met, a result was excluded from a comparison or the relevant series is not comparable across cycles.
Notes are included where necessary to prevent missing data from being interpreted as a performance result.
Updates
Pages may be revised when the OECD releases corrected data, additional technical documentation, updated databases or later PISA publications.
A newer official source takes precedence over an earlier secondary reference or an older version of the same dataset.