5. Conclusion: case analyses of inequality impact of reforms
The aim of the report was to describe the trends of inequalities in education achievement in Europe based on re-analyses of large-scale educational assessment studies. We focussed on analysing differences in educational achievement between students with different parental education, gender, immigrant status and urbanisation.
The findings show that inequality by parental education remained relatively strong compared to the other factors in case of both mathematics and reading. This aspect of inequality proved to be persistent across countries and widening of inequalities was especially seen during the period between 2011 and 2023 in case of Grade 8 mathematics results. On a similar note, gender inequality was generally small in mathematics, with boys showing higher performance, and over time a modest increase in boys’ advantage has been observed during the period between 2003 and 2023. When exploring reading performance gender inequalities were also generally small, but in this case girls tended to show higher achievement. Changes over time were generally moderate, but generally boys disadvantage tends to decline.
Inequalities in educational achievement in mathematics between migrant and non-migrant students are visible; however, there is evidence of moderate convergence in several countries at the primary level and the secondary level. Inequalities in reading achievement remain persistent and often sizeable, especially at age 15. There are more countries showing narrowing gaps during the past two decades in Grade 4, while the majority in Grade 8 recorded a widening gap between migrants and non-migrants. Regarding urban–rural disparities our findings suggest that inequality in mathematics achievement is moderate and stable in primary education, while somewhat larger in secondary education. In reading urban−rural inequality is generally small.
In addition to describing trends in inequalities in educational achievement, this report also aimed at providing an analysis of the effect of certain educational policies on educational inequality using data from the international large-scale assessment studies. More specifically, we have selected early childhood education and care (ECEC) policies and examined the results in terms of inequalities in educational achievement. These studies were motivated by the widespread assumption that expanding access to early childhood education (ECE) can help equalise learning opportunities and outcomes across socio-economic groups. Over the past 25 years, many European countries have implemented preschool expansion policies, often with dual aims: improving the compatibility of work and family life, and enhancing school readiness, particularly for children at risk of lower academic performance. These policies have resulted in rising preschool enrolment rates in most countries, including increased participation by disadvantaged children and younger age groups.
The analysis in chapter 4 examined to what extent are early childhood (ISCED 0) access regulation reforms, described in chapter 3 of the report, associated with changes in educational inequality, using multilevel models on PIRLS and TIMSS data. This analysis failed to find an equalising effect of early childhood education reforms, as it showed that the reforms were correlated with an improvement of results for everyone, and the effect of reforms was not particularly strong among children from low-SES families.
Nevertheless, other studies by members of the project team also looked at the effect of early childhood education on educational inequality, by exploiting variation in the participation rates in ECEC and inequality in access to ECEC rather than looking at the effect of specific policy reforms. In the following paragraphs, we provide a short summary of these studies and their results.
5.1 ECEC expansion and socioeconomic achievement gaps
The core question posed by the first study (Steinman et al. 2024) was whether the expansion of ECEC services has led to a decrease in educational inequalities—measured specifically as the gap in test scores between high- and low-socioeconomic status (SES) students. To explore this, Steinman et al. (2024) use data from PISA (2003–2022), focusing on mean differences in reading and mathematics scores between students in the top and bottom 10% of the Economic, Social, and Cultural Status (ESCS) index. According to the findings of the study high-SES students consistently outperform their low-SES peers across countries and years, with some variation in the size of the gap but no clear trend toward convergence.
To examine the relationship between preschool enrolment and later inequalities, the authors employ a country-fixed effects regression approach. This method allows to compare each country with itself over time, thereby controlling for time-invariant national characteristics—such as structural economic differences or long-standing education policies—that could confound cross-country comparisons. Time-varying control variables were also included in the model to account for demographic or economic changes that might influence achievement gaps. The main independent variable is earlier preschool enrolment, typically lagged by about 10 years to align with the age when students are assessed in PISA.
The findings show that in simple cross-sectional comparisons, countries with higher preschool enrolment rates tend to show greater SES-based inequalities in achievement. However, this pattern likely reflects broader inter-country differences rather than causal effects. Once country-fixed effects are introduced, a small negative association emerges between preschool enrolment and mathematics inequality, suggesting that increases in preschool enrolment may be followed by slight reductions in mathematics achievement gaps. For reading, no significant effect is found. When further controls are added to account for time-varying predictors, the effects in both subjects become statistically insignificant. In other words, the authors find no robust evidence that higher preschool enrolment rates systematically reduce SES-based achievement gaps at age 15.
Steinmann et al. (2024) discuss several possible reasons for these findings. First, limited statistical power may constrain the ability to detect small effects, as each country has a maximum of only seven measurement points over time. Second, the gross enrolment ratio used as the main indicator of preschool participation may not capture relevant differences in enrolment by SES. Additionally, the long-term academic benefits of preschool attendance may vary by context and quality, limiting their equalising potential.
5.2 ECEC participation inequalities and primary school outcomes
In their second study (Steinmann et al. 2026) the authors use PIRLS data on reading achievement in grade 4 to study the effect of preschool enrolment. Grade 4 achievement data allow a different perspective on the issue as the achievement test is temporally closer to preschool, making fade-out effects less pronounced. In addition, in the second analysis the authors take explicitly into account that increasing preschool enrolment does not necessarily lead to a more equitable access to these services among children from lower socioeconomic background.
Although disadvantaged children often benefit particularly strongly from participation in high-quality ECEC, it remains well established that socioeconomically disadvantaged children participate less frequently in universal ECEC programs than their more privileged peers in most countries (e.g., Kulic et al., 2019; OECD, 2025; Van Lancker, 2013). This participation gap represents a critical policy challenge, as it may limit the equalising potential of early education systems. Therefore, the second study by the authors directly tests whether increasing ECEC participation rates can reduce socioeconomic inequalities in educational achievement—particularly when overall expansion is accompanied by a narrowing of participation gaps between socioeconomic groups. The authors analysed data from the Progress in International Reading Literacy Study (PIRLS) study that evaluates reading achievement among fourth-grade students between 2001 and 2021. Socioeconomic status (SES) was measured using parental education information and students were classified as high SES if at least one parent or guardian had completed ISCED Level 5 or higher. This operationalization allowed to capture substantial differences in family educational background while maintaining comparability across countries and time.
Trend analysis showed that although participation in ECEC services expanded overall, the relative difference between socioeconomic groups remained largely stable. The main research question aimed to trace trends in socioeconomic inequalities in reading achievement back to trends in ECE participation rates and to inequalities in ECE participation. To test this mediation hypothesis, the authors estimated separate mediation path models using country-fixed effects regression analyses. The authors estimated the individual paths of the mediation model using separate country-fixed effects regressions. As hypothesized, they found a significant negative association between overall ECE participation rates and socioeconomic inequalities in ECE participation. In other words, as overall participation increases within a country, the participation gap between high- and low-SES children tends to decrease. They also observed a significant positive association between socioeconomic inequalities in ECE participation and inequalities in reading achievement. This suggests that larger participation gaps are associated with larger achievement gaps. However, overall ECE participation rates do not directly and significantly predict inequalities in reading achievement. Importantly, these findings remain robust after including the control variables.
Results obtained with the estimation of the full mediation model supported the hypothesised indirect pathway. Specifically, increasing ECE participation rates were indirectly associated with a small but statistically significant decrease in socioeconomic inequalities in reading achievement, but only when increases in participation were accompanied by reductions in participation gaps between SES groups. The indirect effect is negative and statistically significant, indicating that expansion may contribute to greater equity in achievement if it reduces unequal access (Steinmann et al. 2026).
5.3 ECEC and children with an immigrant background
Another study, by Strømme et al. (2025) analysed whether preschool attendance was an equalising factor for children with immigrant background which could reduce the reading scores gap for these children in 4th grade. Based on data from PIRLS from 2021 they defined immigrant background as speaking other language than test at home. The study showed that in all 24 European education systems participating in PIRLS 2021, children with both parents speaking a foreign language had lower percentages of preschool attendance than those with both parents speaking local language (language of test) and sequentially, these children had lower achievement in reading skills as well.
The analysis found that preschool attendance had in general a positive effect on reading skills for all children, but particularly for children with an immigrant background, helping to narrow the achievement gap between native and immigrant students in primary education.