On average, 61% of adults with below upper secondary attainment in OECD countries do not reach Level 2 in literacy proficiency (where Level 5 is the highest), meaning they are not able to access and understand information in long texts with some distracting information. This compares to 30% of those with upper secondary or post-secondary non-tertiary attainment and only 13% of those with a tertiary qualification. The wide literacy gap across education levels highlights the cumulative impact of formal education on adult skills.
Across OECD countries, tertiary-educated adults (25-64 year-olds) score around 283 points in literacy proficiency on average, ranging from 249 score points in Chile to 314 in Finland, while those with below upper secondary attainment score around 207 points, ranging from 162 score points in Chile to 240 in Sweden. The proficiency gap between these two groups averages 76 points across the OECD.
On average, 25-34 year-olds outperform their 45-54 year-old peers in literacy proficiency by 15 score points across OECD countries. When broken down by educational attainment, however, the proficiency gap by education level matters more than age.
Education at a Glance 2025
Context
As economies and societies undergo rapid technological, demographic and environmental transformations, the demand for strong foundational skills – such as literacy, numeracy and problem solving – has become more pressing than ever. While Education at a Glance traditionally focuses on indicators of educational attainment and participation, understanding the actual skills possessed by adults is essential for assessing the effectiveness of education systems in preparing individuals for lifelong learning, employability and civic engagement. Literacy and numeracy, in particular, are considered foundational skills, in that they are essential for other types of learning: first people learn to read and then they learn through reading. Given that these skills are largely acquired and developed through formal education, measuring proficiency in literacy and numeracy can give governments and policy makers an indication of the effectiveness of their education systems. Literacy and numeracy have become core requirements for navigating increasingly data-driven workplaces and everyday life, from managing personal finances to interpreting public health information.
Survey of Adult Skills (PIAAC); share of 25-64 year-olds scoring at each proficiency level in literacy
For data, see Table 2. For a link to download the data, see Tables and Notes section.
Although closely related to each other, proficiency in literacy and numeracy and educational attainment measure different things. Qualifications earned through formal education do not always reflect the level of an individual’s literacy or numeracy skills – even at the point in life when those qualifications are acquired. Educational qualifications also represent other types of skills that are not reflected in literacy and numeracy proficiency, such as specialised (or practical) knowledge and work-specific skills as well as social and emotional skills and in developing attitudes and motivations that, though crucial, are not directly captured in achievement tests.
Educational attainment has long been widely used as a proxy for skill levels in comparative education analysis. However, findings from the OECD Survey of Adult Skills, a product of the OECD Programme for the International Assessment of Adult Competencies (PIAAC), demonstrate that qualifications do not always align with functional competencies. Many adults with similar levels of formal education exhibit markedly different proficiency levels, depending on the quality of their schooling, the opportunities they’ve had to use and maintain their skills, and broader socio-economic conditions. Conversely, some individuals with relatively low educational attainment may demonstrate strong skill levels, acquired through informal learning, job experience or reskilling later in life.
The PIAAC report Do Adults Have the Skills They Need to Thrive in a Changing World? (OECD, 2024[1]) highlights how large shares of adults in many countries perform below minimum proficiency in numeracy and literacy, limiting their ability to fully participate in labour markets and society. The report also stresses that skill gaps between countries – and within countries between socio-demographic groups – remain significant and persistent. For example, foreign-born adults, older individuals and those with lower levels of formal education are disproportionately represented among low-skilled populations. These findings underline the need to go beyond qualifications and assess skills directly, particularly in the context of rapid labour-market change and growing digitalisation.
This chapter uses data from both Cycle 1 (2012-15) and Cycle 2 (2023) of the Survey of Adult Skills to provide a cross-national and longitudinal perspective on adult proficiency, particularly in literacy. The analysis examines how skill levels are distributed across countries and demographic groups, how they have evolved over time, and how they relate to educational attainment. By comparing results across the two survey cycles, the chapter also assesses countries’ progress in addressing skill gaps over the past decade.
By integrating skill-based indicators into Education at a Glance, this chapter complements traditional attainment-based metrics and enhances understanding of the real capabilities of adult populations. Doing so supports more targeted policy interventions, particularly for countries seeking to improve lifelong learning systems, upskill low-performing groups or align education provision with the evolving needs of the economy. As underscored in the PIAAC international report, fostering a skilled adult population is not only a matter of individual opportunity – it is also critical to building more productive, equitable and resilient societies.
Other findings
On average across OECD countries, 60% of tertiary-educated adults score at or above Level 3 in literacy proficiency, meaning they can interpret, evaluate, and integrate information across complex or lengthy texts, reaching over 80% in Finland and Japan. This falls to only 12% on average among those with below upper secondary attainment, with Chile recording the smallest share.
On average across OECD countries, 25-64 year-old women outperform their male peers in literacy proficiency. However, gender gaps in literacy are small on average, and mask larger cross-country variations. They are widest among low-educated adults and vary in significance and direction across countries.
Literacy scores have declined slightly over the past decade on average across OECD countries, with the largest falls among those with the lowest attainment. Average scores fell by 19 score points for adults with below upper secondary education, 12 score points for those with upper secondary or post-secondary non-tertiary attainment and 9 score points for those with tertiary attainment.
Among foreign-born adults of foreign-born parents, those who speak the language of the host country at home scored on average 247 points in literacy proficiency (a 19 point gap compared to native-born adults) while those who do not scored 229 points (a 38 point gap). Countries such as Canada, Ireland, Lithuania, New Zealand and the Slovak Republic have comparatively small gaps between native-born and foreign-born adults, particularly among those who speak the host language at home. On average, the literacy proficiency gap by immigration background is narrower among adults with a tertiary qualification (41 points) than among those with upper secondary or post-secondary non-tertiary education (47 points) and below upper secondary education (49 points).
This chapter introduces the concept of information-processing skills and presents key indicators from the PIAAC international report (OECD, 2024[1]). In the Survey of Adult Skills, information-processing skills refer to the cognitive abilities required to understand, evaluate, use and engage with written and numerical information in daily life and work contexts. These skills, comprising literacy, numeracy and problem solving, are essential for individuals to effectively navigate and adapt to the demands of the modern information-based economy.
The first section provides an overview of the distribution of proficiency levels in literacy, numeracy and adaptive problem-solving (see section on Definitions) across participating countries and subnational entities. Subsequent sections break down these results by educational attainment, age group, gender and migrant background. For brevity, the chapter focuses primarily on literacy, as the patterns observed are similar across the other skill domains. Additional analyses of PIAAC data are presented in other chapters in Part A.
Data from Cycle 2 of the Survey of Adult Skills reaffirm the strong link between adults’ proficiency in literacy, numeracy and adaptive problem-solving and a wide range of individual and societal outcomes. These foundational and transversal skills not only support participation in the labour market, but also enable individuals to engage meaningfully in civic life, navigate complex information environments, and manage their health, finances and daily tasks more effectively. High levels of proficiency in these domains are associated with greater trust in others and institutions, greater political efficacy and better self-reported health. Conversely, limited skills constrain life opportunities and reinforce social disadvantage across generations (OECD, 2024[1]).
Across countries and demographic groups, greater proficiency in literacy, numeracy and adaptive problem-solving is consistently associated with more favourable outcomes, regardless of formal educational attainment. Adults with strong skills but lower qualifications often outperform their more educated but less proficient peers in many aspects of life, underscoring the distinct and critical role of actual competencies (OECD, 2024[1]). This pattern highlights the importance of focusing not just on access to education, but also on the quality and effectiveness of learning throughout life. It also points to the potential for skills assessments to complement qualifications for recognising individual capabilities.
Comparisons between Cycle 1 (2012-15) and Cycle 2 (2023) of the Survey of Adult Skills show that gaps in proficiency have persisted or even widened in several countries, particularly across socio-economic and generational lines. Although some countries and economies have made progress in raising overall performance or narrowing disparities, others show signs of stagnation or increasing inequality. These findings reflect the combined influence of education and training systems, labour-market structures and access to lifelong learning. They also underscore the urgency of ensuring that all individuals – regardless of age, background or circumstance – have opportunities to develop and maintain the skills needed to adapt, participate and thrive in a rapidly changing world.
Overview of skills proficiency among adults
The review of proficiency scores and levels shows similar results with regard to literacy, numeracy and adaptive problem-solving. Although this section displays the results for all three types of skills, the remainder of the chapter focuses on literacy. The analysis of labour-market and economic outcomes in Chapters A3 and A4 will be based on numeracy proficiency levels instead. Box 1 summarises how the analytical power of the three skill domains differ, and their recommended use in analysis.
PIAAC Cycle 2 assessed the population aged 16 to 65 across three cognitive domains: literacy, numeracy and adaptive problem solving (although this chapter focuses on 25-64 year-olds). These domains are designed to capture distinct but interrelated skill sets that support individuals’ personal, civic and economic engagement. All three domains display similar cross-national patterns in terms of average scores and subgroup distributions: although the domains are psychometrically distinct, they are empirically correlated, particularly at the population level.
The scores for the three domains appear similar but are measured using conceptually distinct units. As a result, it is not meaningful to compare proficiency across domains. For instance, if someone scores 240 points in literacy and 260 points in numeracy, it would be incorrect to conclude that they are “better at numeracy than at literacy” or that they possess greater numeracy skills than literacy ones (OECD, 2025[2]).
The domain best suited for analysis depends on the research question, as the choice of domain can meaningfully influence the interpretation of results. Both literacy and numeracy provide effective tools for comparing skills across age, gender and educational attainment but have different strengths. Adaptive problem solving introduces a forward-looking dimension reflecting digital literacy and cognitive flexibility, but its recent introduction and more limited coverage means it is best used as a complementary domain. The PIAAC international report and its data analysis manual support combining domains where possible to provide richer, more nuanced interpretations (OECD, 2024[1]; OECD, 2025[2]).
Literacy is a foundational skill for assessing adult capability and civic engagement. It also tends to exhibit more stable distributions than numeracy, making it particularly effective for monitoring equity and making comparisons across demographic groups. Literacy is broadly communicable to non-specialist audiences, and for these reasons it is the main focus of this chapter.
Numeracy stands out as the preferred domain for analysing economic and labour-market outcomes such as employment prospects and relative earnings, particularly when used alongside educational attainment (OECD, 2014[3]). Numeracy proficiency is strongly linked to employment, earnings and job quality across countries and population groups, even after adjusting for educational attainment (OECD, 2024[1]). Numeracy is often used in regression models focused on employment-related outcomes, because of its closer alignment with the types of quantitative reasoning tasks encountered in the modern workplace (OECD, 2025[2]). Average proficiency by education level reveals consistent gradients in both literacy and numeracy, but numeracy often exposes wider disparities, particularly among highly educated and low-educated adults, and more pronounced skill gradients, particularly by occupation or field of study. Its greater variance and steeper proficiency gradients also makes it more suitable for in-depth stratified or occupational analyses (OECD, 2024[1]). It is therefore the main focus of the PIAAC analysis in Chapters A3 and A4.
Adaptive problem solving, introduced in PIAAC Cycle 2, adds breadth to the assessment framework by measuring individuals' ability to solve complex, unfamiliar tasks in dynamic, technology-rich environments. Although conceptually distinct from both literacy and numeracy, adaptive problem-solving correlates moderately with both and serves as a useful indicator of cognitive flexibility (OECD, 2024[1]). However, its use in demographic analysis has been limited, partly because it has fewer assessment items and narrower country coverage. Adaptive problem solving should therefore be used as a supplementary indicator rather than a primary variable in disaggregated cross-country or subgroup comparisons (OECD, 2025[2]).
Literacy
Literacy proficiency allows individuals to accessing, understanding, evaluating and reflecting on written texts in order to achieve one’s goals, to develop one’s knowledge and potential, and to participate in society (see section on Definitions). Data from the Survey of Adult Skills (OECD, 2024[1]) provide an updated profile of literacy skills among adults aged 25 to 64, revealing how well countries are preparing their populations for these challenges.
According to the 2023 data, the average literacy score across OECD countries is 259 points, ranging from 214 score points in Chile to 297 score points in Finland (Table 1). On average, 42% of adults performed at or above Level 3 for literacy (30% at Level 3 and 12% at or above Level 4, see Box 2 for short descriptions of proficiency levels), meaning they can understand and respond appropriately to dense or lengthy texts (see section on Definitions). Meanwhile, about 27% of adults perform at or below Level 1, indicating difficulty with basic written information, and 31% at Level 2 indicating those who can, for example, integrate of information from multiple sources (see Box 2). These figures highlight the persistence of significant literacy gaps even in high-income countries (Figure 1 and Table 2).
The highest average scores are observed in Finland, Japan, Norway and Sweden, each scoring well above the OECD average. In these countries, over 60% of adults performed at or above Level 3, while less than 20% were at or below Level 1. At the other end of the scale, Chile, Poland and Portugal report the lowest average literacy scores. In these countries, 40% or more of adults perform at or below Level 1, and 22% or less were at or above Level 3 (Table 2). These results reflect differences not just in education systems but also in broader patterns of adult learning, literacy use in daily life and access to lifelong learning opportunities.
Proficiency levels
In each of the three domains assessed, proficiency is considered as a continuum of ability involving the mastery of information-processing tasks of increasing complexity. The results are represented on a 500-point scale. To help interpret the results, the reporting scales have been divided into “proficiency levels” defined by particular score-point ranges. Six proficiency levels are defined for literacy and numeracy (Levels 1 through 5 plus below Level 1) and five for problem solving in technology-rich environments (Levels 1 through 4 plus below Level 1). Each proficiency level is described in terms of the characteristics of the types of tasks that can be successfully completed by adults with proficiency scores in the range of scores that defines a level. Short descriptions of the types of tasks related to each level are provided below, while more details can be found in the Definitions section and in Tables 2.4, 2.5 and 2.6 in the PIAAC international report (OECD, 2024[1]).
Literacy and numeracy:
Below Level 1: Basic tasks involving simple text or numerical information.
Level 1 (scores equal to or higher than 176 points): Tasks requiring the identification of information in short texts or simple mathematical operations.
Level 2 (scores equal to or higher than 226 points): Tasks involving integration of information from multiple sources or application of basic mathematical concepts.
Level 3 (scores equal to or higher than 276 points): Tasks requiring interpretation and evaluation of complex texts or more advanced mathematical reasoning.
Level 4 (scores equal to or higher than 326 points): Tasks involving complex reasoning and problem-solving with intricate texts or mathematical information.
Level 5 (scores equal to or higher than 376 points): Tasks requiring the synthesis and critical evaluation of complex information or advanced mathematical concepts.
Adaptive problem solving:
Below Level 1: Tasks involving simple problem solving in familiar contexts.
Level 1 (scores equal to or higher than 176 points): Tasks requiring the application of basic strategies to solve problems in well-defined situations.
Level 2 (scores equal to or higher than 226 points): Tasks involving problem-solving in less familiar contexts with some complexity.
Level 3 (scores equal to or higher than 276 points): Tasks requiring adaptive reasoning and problem solving in complex and dynamic situations.
Level 4 (scores equal to or higher than 326 points): Tasks requiring complex reasoning and multi-step solutions to one or more goals.
Numeracy
Numeracy proficiency reflects adults’ ability to access, use and reason critically with mathematical content, information and ideas represented in multiple ways in order to engage in and manage the mathematical demands of a range of situations in adult life (see section on Definitions). In an increasingly data-driven society, strong numeracy skills are essential for economic resilience and adaptability.
Across OECD countries and economies, the average numeracy score is 262 points, ranging from 211 score points in Chile to 294 score points in Finland (Table 8, available on line). On average, 44% of adults reach at least Level 3 proficiency (30% at Level 3 and 14% at or above Level 4, see Box 2 for short descriptions of proficiency levels), demonstrating the ability to work with mathematical concepts and reasoning (see section on Definitions). Meanwhile, although 30% reach Level 2 (i.e. those who can integrate concepts from different mathematical procedures, see Box 2), approximately 25% score at or below Level 1, indicating difficulty with basic arithmetic tasks and limited ability to apply numerical reasoning in everyday contexts (Table 9, available on line).
Similarly to literacy, Finland and Japan report the highest numeracy scores, averaging at or above 290 points, and a majority of their adult populations score at or above Level 3. In contrast, adults in Chile, Poland and Portugal record the lowest average scores, below 240 points. In these countries, at least 39% of adults score at or below Level 1, with only a small fraction reaching Level 3 or above (Table 8 and Table 9, available on line). These findings point to persistent gaps in basic numeracy that may limit individuals' participation in training, employment and civic life.
Adaptive problem-solving
In a rapidly evolving digital world, adaptive problem-solving is increasingly essential. This skill refers to the capacity to achieve one’s goals in a dynamic situation, in which a method for reaching a solution is not immediately available. It requires engaging in cognitive and metacognitive processes to define the problem, search for information, and apply a solution in a variety of information environments and contexts (see section on Definitions). It is crucial for navigating both the workplace and daily life, especially as technology becomes more embedded in social, financial and civic activities.
The average score for adaptive problem solving is 249 points across participating OECD countries and economies (Table 15, available on line). On average, 31% of adults reach at least Level 3 of proficiency in this domain (26% at Level 3 and 5% at Level 4), meaning that they can integrate simultaneously several important variables and consider the impact of several problem elements on each other, while 31% are at or below Level 1 (Table 16, available on line), meaning they struggle with tasks that involve basic digital interfaces or routine problem solving (see section on Definitions).
Finland, Japan and Sweden I record the highest average proficiency in this domain – above 270 score points – with at least 52% of adults achieving at or above Level 3. In contrast, Chile reports the lowest average proficiency, at 214 score points, with only 10% of adults reaching at or above Level 3 (Table 15 and Table 16, available on line). These differences suggest that efforts to improve problem solving must go beyond technical skills and address broader issues of access, confidence and opportunities to practise such skills in daily life.
Literacy proficiency by educational attainment
Educational attainment is a major determinant of literacy proficiency in adulthood. Formal education lays the foundation for reading comprehension, vocabulary development and the ability to analyse and synthesise text – skills that remain essential well beyond the classroom. However, the extent to which educational qualifications translate into actual skills can vary across countries, systems and demographic groups. Although higher education levels are associated with higher literacy scores, there are cross-country variations in how pronounced these differences are.
Across participating OECD countries and economies, adults with tertiary education average around 283 score points in literacy, compared to about 207 score points for those with below upper secondary education – a 76 point gap (Figure 2). The share of those scoring at or above Level 3 is 60% among tertiary-educated adults, but only 12% among those with below upper secondary attainment. Conversely, 61% of adults with below upper secondary education are at or below Level 1, compared to just 13% of tertiary-educated adults (Table 2). These figures confirm the steep literacy gradient by education level and underscore the cumulative effect of formal education on adult skill levels.
Finland, Japan, Norway and Sweden show particularly strong outcomes for tertiary-educated adults, with average scores of 300 points or above and more than 70% reaching Level 3 and above in literacy (Table 1 and Table 2). The literacy gap between low- and high-educated adults is relatively narrow in countries like Croatia and the Slovak Republic, suggesting more equitable education and lifelong learning systems. In contrast, countries such as Chile, Switzerland and the United States have much lower literacy scores among adults with below upper secondary education and over 70% of low-educated adults at or below Level 1. Austria, France and Germany also show wide gaps, highlighting systemic challenges in foundational education and limited opportunities for skill development among low-qualified adults (Figure 2 and Table 2).
Survey of Adult Skills (PIAAC); 25-64 year-olds; in score points
For data, see Table 1. For a link to download the data, see Tables and Notes section.
Literacy skills by age group
The Survey of Adult Skills found that in nearly all OECD countries and economies the highest average proficiency is observed among younger age groups (16-24 year-olds, and especially 25-34 year-olds), while older age groups tend to score the lowest (OECD, 2024[1]). This is true across different levels of educational attainment, with young adults (25-34 year-olds) scoring on average 15 points higher than older adults (45-54 year-olds). High literacy scores for young adults reflect a combination of age-related cognitive factors and cohort effects, particularly linked to the expansion of education and training opportunities (Kautz et al., 2014[4]). Although the overall pattern is consistent across countries, the size of the generational gap and the impact of educational attainment on literacy proficiency vary considerably.
In all participating OECD countries and economies, tertiary-educated 25-34 year-olds outperform their older peers in literacy proficiency, averaging 291 score points compared to 281 score points among tertiary-educated 45-54 year-olds. Finland, Estonia and Japan lead in literacy scores among tertiary-educated young adults, averaging above 310 score points (Figure 3 and Table 3).
Tertiary-educated young adults also average 73 score points more than their peers with below upper secondary attainment. Similar gaps are observed among 45-54 year-olds, with a 75 point difference between those with tertiary and below upper secondary attainment (Figure 3 and Table 3).
The proficiency gap by education level is substantial across countries and seems to matter more than age factors. The average difference between tertiary-educated and low-educated adults (73-75 points) far exceeds the average 15 point gap observed between younger and older adults. In Germany and Switzerland, the gap is over 99 score points between young adults with tertiary education and those with below upper secondary education, showing pronounced inequalities in proficiency outcomes linked to educational background. In contrast, countries like Italy, Lithuania and Poland demonstrate relatively smaller gaps by attainment level, suggesting more equitable educational systems or less stratification in literacy outcomes (Figure 3 and Table 3).
Survey of Adult Skills (PIAAC); in score points
For data, see Table 3. For a link to download the data, see Tables and Notes section.
Gender differences in literacy skills
Gender differences in proficiency are generally small across OECD countries (OECD, 2024[1]). The Survey of Adult Skills found that, on average, women score 3 points higher than men in literacy proficiency. Among tertiary-educated adults, men score 2.2 points higher than women in literacy proficiency, with minimal differences across countries (Figure 4). In contrast, among those with lower educational attainment (below upper secondary), women score 1.1 points higher on average, but the gender gaps vary significantly in size and direction across countries and economies (Figure 4).
Among adults with below upper secondary attainment, men outperform their female counterparts by 17 score points in the Flemish Region (Belgium) and by 14 score points in Israel. Conversely, women outperform men in Czechia (16 points) and New Zealand (25 points). Among those with tertiary attainment, gender gaps are generally smaller. The largest differences are in Chile and Japan, where highly educated men outperform women by more than 9 points (Figure 4).
Educational attainment remains a stronger predictor of literacy proficiency than gender alone. Compared to the average gender gap of just 3 points, on average across OECD countries and economies, the gap between tertiary-educated women and those with below upper secondary attainment is 75 score points. For men, the corresponding gap is 78 points. The largest education-related proficiency gaps are observed in Germany, Switzerland and the United States, where differences exceed 106 points for men and 97 points for women. This underscores the critical role of educational level in shaping literacy outcomes (Figure 4).
Survey of Adult Skills (PIAAC); 25-64 year-olds; in score points
For data, see Table 1. For a link to download the data, see Tables and Notes section.
Comparatively, differences in numeracy and adaptive problem-solving proficiency are also relatively small, with men displaying higher average scores. On average, men score 10 points higher than women in numeracy (see Table 8, available on line) and 3 points higher in adaptive problem solving (see Table 15, available online). In numeracy, gender gaps are more pronounced among tertiary-educated adults. Women are under-represented in STEM and less likely to work in numeracy-intensive jobs, partly explaining numeracy gender performance (OECD, 2024[1]). These gaps mirror patterns seen in PISA and highlight the lasting impact of educational and occupational trajectories on adult skill development (OECD, 2024[1]).
Moreover, between Cycle 1 and Cycle 2 of the Survey of Adult Skills, literacy proficiency has declined more strongly among men than women (OECD, 2024[1]). As a result, gender gaps in literacy have narrowed in many countries, with women now outperforming men on average (OECD, 2024[1]). This shift is linked to the transversal nature of literacy proficiency (OECD, 2019[5]). Unlike numeracy, which is more strongly tied to fields and occupations more commonly pursued by men, literacy is less dependent on occupational context and thus less sensitive to gendered choices in education and employment (Borgonovi, Choi and Paccagnella, 2021[6]; OECD, 2024[1]). In Chile, Germany and the Flemish Region (Belgium), the gender gap in numeracy narrowed due to improvements in women’s proficiency, whereas in most other countries, the narrowing resulted from significant declines in men’s numeracy scores (Table 17 on line).
Trends in adult literacy proficiency between PIAAC cycles
Monitoring changes in adult literacy proficiency over time helps assess the long-term effectiveness of education and training systems, the impact of demographic and labour-market shifts, and whether adult learning opportunities are keeping pace with societal needs. Comparisons between Cycle 1 (2012-15) and Cycle 2 (2023) provide insight into whether literacy levels are improving, stagnating or declining (see Box 3 for the methodological considerations involved in comparing the results of the two cycles). Average proficiency has remained relatively similar between cycles although most countries experienced slight changes, with some seeing gains due to expanded educational attainment in younger cohorts, and others declines due to population ageing or limited adult learning engagement.
On average across participating OECD countries and economies with data for all years, average literacy proficiency among 25-64 year-olds fell by 9 score points since PIAAC Cycle 1, to 260 points (Table 6, available on line). However some countries, such as Finland, recorded meaningful increases in average literacy scores – by 10 points – suggesting sustained improvements in education quality or participation in adult learning. In contrast, Lithuania and Poland saw declines of over 25 score points. These declines may reflect ageing populations or changes in the composition of the adult population. Meanwhile, countries like Estonia, Norway and Flemish Region (Belgium) reported relatively stable scores over time.
This pattern of slight overall decrease, coupled with diverging national trajectories, points to the importance of national policy contexts in shaping adult skill levels. Countries investing in lifelong learning and inclusive education systems appear better positioned to sustain or improve literacy skills over time (OECD, 2023[7]). The declines in scores over time are disproportionally attributable to changes among adults with lower attainment. The average score declined by 19 score points for adults with below upper secondary education, by 12 score points for those with upper secondary or post-secondary non-tertiary attainment and by 9 score-points for tertiary-educated adults (Figure 5 and Table 6, available on line).
Comparing differences in literacy proficiency between Cycle 1 and Cycle 2 among adults in the same age group reveals a general downward trend. On average across OECD countries and economies with data for both cycles, adults aged 25-34 in 2023 scored 9 points lower than their counterparts in 2012-15, while 45-54 year-olds saw a 7 point decline. These declines are evident across most OECD countries, although the scale varies by country. Among the younger age group, the largest falls are observed in Lithuania (26 points), New Zealand (23 points), Poland (34 points) and the Slovak Republic (26 points) (Table 7, available on line).
Survey of Adult Skills (PIAAC); 25-64 year-olds; in score points
1. Year of reference: 2015 instead of 2012.
For data, see Table 6, available on line. For a link to download the data, see Tables and Notes section.
The decline in literacy proficiency is most pronounced among those with below upper secondary education, where proficiency scores fell by an average of 17 points for younger adults and 20 points for older adults. Austria, Japan and New Zealand stand out in this category: in Austria and Japan, the decline reaches 41 points for older adults, while in New Zealand, both age groups experienced declines of over 40 points. For those with upper secondary education, older adults showed on average a 11 point decline in literacy proficiency, compared to 9 points for the younger cohort. Tertiary-educated adults also experienced declines, although to a lesser extent (8 points for younger adults and 7 points for older adults on average) (Table 7, available on line).
Comparing 25-34 year-olds in 2012-15 with 35-44 year-olds in 2023 and so on provides further insight into ageing effects on skills on cohorts over time. Average literary proficiency among 35-44 year-olds in 2023 was 15 score points lower than that for 25-34 year-olds in 2012-15, while 35-44 year-olds in 2012-15 had an average literacy proficiency 17-score points higher than 45-54 year-olds in 2023. For older adults – comparing those aged 45-54 in 2012-15 to those aged 55-64 in 2023 – the decline reaches 21 points (Table 7, available on line). These findings confirm the general pattern of skill attrition with age, which may reflect both natural cognitive decline and reduced opportunities for skill use (OECD, 2019[5]).
However, the change in proficiency differs by educational attainment. Among adults with below upper secondary education, the younger cohort (aged 25-34 in 2012-15) experienced an average decline in literacy proficiency of 26 score points, which was slightly larger than the 25-point decline experienced by the older cohort (aged 45-54 in 2012-15). There are a few exceptions to the general trend. In Estonia, the Flemish Region of Belgium and Japan, the drop among older low-educated adults was over 20 points larger than among their younger peers. At higher attainment levels, although both cohorts saw proficiency declines, these were more pronounced among older adults. For those with upper secondary education, older adults showed on average a 21-point decline in literacy proficiency, compared to 16 points for the younger cohort. Among tertiary-educated adults, the decline was 20 points for older adults compared to 12 points for younger ones (Table 7, available on line). These smaller declines suggest that higher educational attainment can buffer, but not entirely offset, skill loss over time.
Overall, the decline in literacy proficiency appears to be driven by a combination of factors affecting both younger and older adults. Larger declines among younger adults with low educational attainment may point to weaknesses in initial education, while skill loss among older adults suggests broader societal challenges, such as limited opportunities for skill use. These dynamics also vary by country: for instance, in Estonia and Japan the older cohort showed a greater decline in skills than the younger one across all education levels, whereas in Sweden and the Netherlands, the steepest declines are observed among younger adults. These differences highlight the need for country-level analysis to identify individual patterns that may be obscured by the overall averages.
The 2023 Survey of Adult Skills (PIAAC) introduced several modifications to the design of the survey instruments (both the background questionnaire and the cognitive assessment) and the data collection. These changes sought either to adapt the instruments to societal and contextual developments that occurred between the two cycles, or to correct and improve measurements (OECD, 2024[8]). Some caution is therefore advised when comparing results across the two cycles.
A more detailed discussion of the differences between the first and second cycle of PIAAC can be found in the Reader’s Companion (OECD, 2024[8]) as well as in the Data Analysis Manual (OECD, 2025[2]) and in the PIAAC Technical Report (OECD[9]).
Linking error
When comparing average proficiency scores between the first and second cycles of PIAAC, a linking error must be taken into account due to changes in the set of assessment items. Although about one-third of the items used in Cycle 2 were trend items used in Cycle 1, the same score does not have exactly the same meaning in both cycles. This difference is modelled as a random variable known as the linking error (3.27 for literacy, 2.95 for numeracy), which should be added to the standard error of any trend statistic expressed as a proficiency score (OECD, 2025[2]). It should be noted that the linking error does not apply to changes in score-point differences between subgroups (e.g. gender or age gaps in proficiency scores), as the associated uncertainty cancels out.
Doorstep interviews
In Cycle 2, respondents who could not complete the background questionnaire due to language barriers were administered a short, self-completed doorstep interview available in over 40 languages. This collected key demographic and background data (e.g. age, gender, education, employment status and migration history) to generate plausible values, allowing these individuals to contribute to population-level estimates. In Cycle 1, similar respondents were classified as literacy-related non-respondents and excluded from proficiency estimates. For cross-cycle comparisons, adults who only completed the doorstep interview have been excluded, as they would not have received a score in Cycle 1.
Adjusted differences
This chapter does not present adjusted differences between population subgroups. However, the Survey of Adult Skills (PIAAC) provides an analysis of both adjusted and unadjusted differences. Unadjusted differences show the observed gaps in proficiency between groups, while adjusted differences account for differences in socio-demographic characteristics such as age, gender or immigrant background that are independently associated with proficiency. These adjusted estimates, derived from linear regression models, offer insights into the extent to which the observed gaps reflect underlying group differences rather than the effect of the characteristic being studied (e.g. educational attainment). Readers interested in these analyses are referred to the PIAAC international report for detailed results (OECD, 2024[1]) and the technical manual for more details (OECD, 2025[2]).
Skills gaps by migration and language background
Migration and language background can significantly influence adult skill levels, particularly literacy proficiency. Adults born abroad or speaking a different language at home may face additional challenges in developing and maintaining literacy skills in the language of their host country, especially if they had limited access to high-quality initial education or encounter language barriers in everyday life. The Survey of Adult Skills sheds light on these disparities by distinguishing between groups based on both migration status and the language spoken at home (OECD, 2024[1]). Gaps persist between native-born and foreign-born adults, and are even wider between those who do or do not speak the host country language at home.
On average across participating OECD countries and economies, native-born adults with native-born parents averaged 266 score points in literacy, while foreign-born adults with foreign-born parents score averaged 247 points if they speak the host country language at home (a 19 point gap) and 229 points if they do not (a 38 point gap). These differences suggest that both language exposure and migration-related barriers (such as interrupted schooling or credential recognition) might contribute or are related to lower proficiency. Countries such as Canada, Ireland, Lithuania, New Zealand and the Slovak Republic show comparatively small gaps between native-born and foreign-born adults, particularly among those who speak the host language at home. In these countries, inclusive education policies and access to adult learning opportunities may help narrow disparities. In contrast, France, Germany and the Flemish Region (Belgium) report much larger gaps, especially for foreign-born adults who do not use the host language at home – often exceeding 70 points (Table 5). However, caution is needed when interpreting these figures, as in some cases estimates may be based on relatively small samples.
The composition of immigrant populations significantly influences literacy outcomes across countries; although inclusive education policies and access to adult learning opportunities are vital, educational backgrounds and language proficiency among migrants also play crucial roles. For instance, Canada's immigration system employs a points-based model that prioritises highly educated individuals with strong language skills. This approach has resulted in a substantial proportion of immigrants possessing university degrees (Parisa Mahboubi, 2024[10]). In contrast, countries like Germany and the Republic of Türkiye have experienced significant influxes of refugees, particularly from Syria, with varied educational backgrounds and who may not speak the host country's language at home. For example, studies show that Syrian refugees in Türkiye have educational distributions similar to those in pre-war Syria, with many having limited formal education (OECD, 2024[11]; Güray Kirdar, Koç and Dayıoğlu, 2023[12]).
On average across OECD countries, the gap in literacy proficiency associated with educational attainment is wider than that associated with migration background. The gap associated with migration background narrows slightly with educational attainment. Among those with tertiary education, the gap between native-born adults with native-born parents and foreign-born adults who do not speak the language of the host country at home is 41 score points, widening to 47 score points among those with upper secondary or post-secondary non-tertiary attainment and 49 points for those with below upper secondary education (Figure 6 and Table 5).
Survey of Adult Skills (PIAAC); 25-64 year-olds; in score points
For data, see Table 5. For a link to download the data, see Tables and Notes section.
Age groups: Adults refer to 25-64 year-olds. Younger adults refer to 25-34 year-olds. Older adults refer to 45-54 year-olds.
Educational attainment refers to the highest level of education successfully completed by an individual. See the Reader’s Guide at the beginning of this publication for a presentation of all ISCED 2011 levels.
Literacy
Literacy in PIAAC Cycle 2 is defined as the ability to access, understand, evaluate and reflect on written texts to achieve one's goals, develop knowledge and potential, and participate in society. This encompasses both traditional print-based texts and digital texts, acknowledging the growing importance of navigating and interpreting information in digital environments. Tasks may involve multiple sources and formats, including continuous (e.g. sentences, paragraphs), non-continuous (e.g. charts, tables) and mixed texts, reflecting a range of genres and contexts.
Numeracy
Numeracy in PIAAC Cycle 2 is accessing, using and reasoning critically with mathematical content, information and ideas represented in multiple ways in order to engage in and manage the mathematical demands of a range of situations in adult life. The assessment covers engagement with mathematical information in both traditional and digital environments, including tasks that require understanding and applying mathematical concepts in real-life contexts. An additional assessment of numeracy components focuses on skills essential for achieving automaticity and fluency in managing mathematical and numerical information.
Adaptive problem solving
Replacing the previous domain of problem solving in technology-rich environments, Cycle 2 introduced adaptive problem solving, defined as the capacity to achieve one's goals in dynamic situations where a solution method is not immediately apparent. This requires engaging in cognitive and metacognitive processes to define the problem, search for information and apply a solution across various information environments and contexts. This broader construct reflects the evolving nature of the problem-solving skills needed in today's complex and digital world.
Proficiency levels
In PIAAC Cycle 2, proficiency in each domain is measured on a continuous scale and categorised into levels to aid interpretation. While the specific score ranges and descriptions for each level are detailed in the official assessment framework, the general structure is as follows (see Tables 2.4, 2.5 and 2.6 in the PIAAC international report for more detailed descriptions (OECD, 2024[1])).
Literacy Proficiency Levels
Below Level 1: Tasks at this level require the respondent to read brief texts on familiar topics and locate a single piece of explicitly stated information. The text structure is simple and the information that needs to be located is identical in form to what is in the question or directive. Only basic vocabulary knowledge is required. There is no need to understand the structure of sentences or make inferences. Texts are short, and there is little competing information.
Level 1: Tasks at this level require the respondent to read relatively short digital or print texts to locate a single piece of explicitly stated information. There is little, if any, competing information. The information in the question or directive is identical to or synonymous with the information in the text. Some tasks may require the respondent to enter personal information onto a document, such as a form. Only basic vocabulary knowledge is required, and the tasks rely mainly on simple matching or locating strategies.
Level 2: At this level, tasks require respondents to make matches between the text and information, including some that may require low-level inferences. Some competing information may be present. The texts may be continuous, non-continuous, or mixed, and the task may require integration of two or more pieces of information. The information may need to be compared or contrasted. There is some use of digital tools and navigation across pages may be necessary.
Level 3: Tasks at this level require the respondent to integrate several pieces of information and to recognize the relationship between different parts of a text, or to evaluate their relevance. There may be a need to perform multi-step operations, compare and contrast or reason about the information. Texts are often dense or lengthy, and multiple distractors are present. Navigation across a variety of digital environments or layout features may be required.
Level 4: Tasks at this level require the respondent to perform multiple-step operations to integrate, interpret, or synthesize information from complex or lengthy texts. The texts may be unfamiliar in topic, with complex structures. Competing information is often present, and a high level of inferencing is necessary. The tasks may require critical evaluation of information and distinguishing relevant from irrelevant content.
Level 5: At this highest level, tasks require the respondent to search for and integrate information across multiple, dense texts; construct syntheses; make high-level inferences or use specialised background knowledge. Texts are complex and lengthy and may contain dense or ambiguous information. Tasks demand a high level of abstraction, logic, and reasoning, and often require evaluating the reliability of different sources or resolving conflicting pieces of information.
Numeracy Proficiency Levels
Below Level 1: Tasks at this level require the respondent to carry out simple processes such as counting, sorting, performing basic arithmetic operations with whole numbers or understanding simple percentages, such as 50%. Tasks are based on concrete, familiar contexts where the mathematical content is highly accessible. No interpretation of text is required. Instructions and numerical information are straightforward and require minimal inference or problem structuring.
Level 1: Tasks at this level involve basic mathematical content such as quantities and money, time, or simple measurements. Respondents may be required to perform simple one-step operations such as arithmetic with whole numbers or percentages in concrete contexts. The mathematical information is explicitly presented, and the tasks require little or no text interpretation or complex reasoning.
Level 2: Tasks at this level require the application of two or more steps and may involve calculation with decimals, percentages, and fractions. Respondents may need to interpret simple data representations such as tables or graphs and understand proportional relationships. The contexts are more varied and may be less familiar. Tasks may involve some reasoning and choosing appropriate arithmetic operations.
Level 3: Tasks at this level require respondents to understand and work with mathematical information that may be embedded in less familiar contexts. They often require several steps and involve problem-solving, proportional reasoning, or working with simple algebraic formulas. Respondents must interpret and evaluate data from various sources and use appropriate strategies to identify relevant mathematical processes.
Level 4: Tasks at this level involve understanding a broad range of mathematical information, including formal and abstract mathematical representations. Respondents may be required to integrate multiple sources of data, make inferences based on quantitative evidence, or solve problems in unfamiliar contexts. Tasks demand reasoning, analysis, and the ability to select and apply appropriate strategies flexibly.
Level 5: At this level, tasks require the respondent to conceptualise, evaluate and apply mathematical or statistical information in complex and abstract settings. Problems may be highly unfamiliar and require the use of sophisticated reasoning strategies and advanced quantitative tools. Tasks often involve modelling, structuring, and critically assessing real-world problems using formal mathematics.
Adaptive Problem-Solving
Below Level 1: Tasks at this level require carrying out simple, routine procedures in highly familiar contexts. The problem to be solved is immediately apparent and involves no unexpected developments or need for goal setting. Success can be achieved through straightforward recognition and recall, with minimal need to adjust responses or monitor progress. The environment is predictable, and only a single action or step is typically necessary.
Level 1: Tasks require executing a short sequence of steps in response to a clearly defined and concrete problem. Situations are familiar, interfaces are standard, and the necessary knowledge is commonly held. There may be some need to identify relevant options or perform simple adjustments, but no significant reasoning or re-planning is needed. Respondents succeed by applying familiar strategies or known procedures with limited need for adaptation.
Level 2: Tasks involve resolving problems in somewhat unfamiliar situations. Respondents need to interpret the problem context, plan steps, and monitor progress. Situations may involve partial or evolving information, moderate amounts of irrelevant or distracting information, and the need to choose among alternatives. Success depends on adaptive reasoning–identifying what is relevant, discarding what is not, and adjusting strategies mid-process.
Level 3: Tasks involve multiple steps, ambiguous goals, or constraints that emerge during the task. Respondents must evaluate options, deal with unexpected information or outcomes, and show flexibility in problem-solving. Success requires coordinated use of planning, reasoning, and monitoring, as well as learning from feedback. Respondents are required to manage cognitive complexity and demonstrate self-regulation in dynamic situations.
Level 4: Tasks involve solving complex problems in unfamiliar and evolving situations with competing goals and multiple constraints. Respondents must independently set objectives, plan multi-step strategies, evaluate the relevance and reliability of information, and adjust their approach based on feedback or new conditions. High-level reasoning, abstraction, and critical thinking are essential. These tasks reflect sophisticated real-life problem-solving skills under uncertainty and require sustained cognitive effort and decision-making autonomy.
The Survey of Adult Skills, part of the OECD Programme for the International Assessment of Adult Competencies (PIAAC), evaluated the skills of adults aged 16 to 65 in three key areas: literacy, numeracy and adaptive problem solving. These competencies are essential for navigating social contexts, succeeding in the labour market, engaging in education and training, and participating fully in civic life.
In addition to assessing skills, the survey gathered detailed background information on respondents, including their education, employment history and various outcomes such as health status. It also collected data on how frequently adults engage in literacy and numeracy tasks, their use of digital technologies at work and in daily life, and the importance of transversal skills such as collaboration and time management in their jobs. Respondents were also asked whether their qualifications and skills matched their job requirements and if they had autonomy in key aspects of their work.
The assessment was primarily conducted via computer, although respondents with little or no computer experience were offered a paper-based version. The test was administered in the official language(s) of each country or, in some cases, a widely spoken minority language.
Twenty-seven countries and economies participated in both cycles of the survey. While only one round of the second cycle has been conducted so far (in 2022/23), the first cycle was carried out in three rounds: Round 1 in 2011/12, Round 2 in 2014/15 and Round 3 in 2017. As different countries participated in the three rounds, the amount of time that has elapsed between the two data collections is not the same for all countries and economies. The majority of them (21 out of 27) participated in Round 1 of the first cycle, 11 years before the second cycle. For this reason, this chapter often refers to changes that occurred “over the past decade”, for ease of exposition. Five countries participated in Round 2 of Cycle 1, eight years before the second cycle. Hungary participated in Round 3 of Cycle 1, only six years before the second cycle (the United States also participated in Round 3). Because of these differences, the size of the change in proficiency between the cycles is not comparable across the participants in the different rounds of the first cycle. All the figures in this chapter, therefore, group countries and economies according to when they participated in the first cycle, and no results are given for the average across OECD countries.
The international report of the 2023 Survey of Adult Skills analyses skills more comprehensively also focusing on a larger age range covering the population aged 16-65, while this Chapter mainly refers to adults aged 25 to 64. In addition, Cycle 2 of the Survey of Adult Skills includes data from doorstep interviews to evaluate language barriers in on order to administer the questionnaires: data from doorstep interviews are often included in Education at a Glance tables, except when this was not possible due to methodology (for example, in tables comparing Cycle 1 and Cycle 2). Notes under each table specify whether data from the doorstep interviews are included.
More information on sampling and methodology is available in the Technical Report of the Survey of Adult Skills (OECD[9]).
For further details, refer to the OECD Handbook for Internationally Comparative Education Statistics (OECD, 2018) and the Education at a Glance 2025 Sources, Methodologies and Technical Notes (https://doi.org/10.1787/fcfaf2d1-en).
Data on proficiency levels and mean scores are based on the Survey of Adult Skills (PIAAC) (2012-15 and 2023). PIAAC is the OECD Programme for the International Assessment of Adult Competencies.
References
[6] Borgonovi, F., A. Choi and M. Paccagnella (2021), “The evolution of gender gaps in numeracy and literacy between childhood and young adulthood”, Economics of Education Review, Vol. 82, p. 102119, https://doi.org/10.1016/j.econedurev.2021.102119.
[12] Güray Kirdar, M., İ. Koç and M. Dayıoğlu (2023), “School integration of Syrian refugee children in Turkey”, Labour Economics, Vol. 85, https://doi.org/10.1016/j.labeco.2023.102448 (accessed on 2 June 2025).
[4] Kautz, T. et al. (2014), “Fostering and Measuring Skills: Improving Cognitive and Non-cognitive Skills to Promote Lifetime Success”, OECD Education Working Papers, No. 110, OECD Publishing, Paris, https://doi.org/10.1787/5jxsr7vr78f7-en.
[2] OECD (2025), Survey of Adult Skills 2023 Data Analysis Manual, OECD Publishing, Paris, https://doi.org/10.1787/25a87a9d-en.
[9] OECD (2025), Survey of Adult Skills 2023 Technical Report, OECD Publishing, Paris, https://doi.org/10.1787/80d9f692-en.
[1] OECD (2024), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/b263dc5d-en.
[11] OECD (2024), International Migration Outlook 2024, OECD Publishing, Paris, https://doi.org/10.1787/50b0353e-en.
[8] OECD (2024), Survey of Adult Skills – Reader’s Companion: 2023, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/3639d1e2-en.
[7] OECD (2023), OECD Skills Outlook 2023: Skills for a Resilient Green and Digital Transition, OECD Publishing, Paris, https://doi.org/10.1787/27452f29-en.
[5] OECD (2019), Skills Matter: Additional Results from the Survey of Adult Skills, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/1f029d8f-en.
[3] OECD (2014), Education at a Glance 2014: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/eag-2014-en.
[10] Parisa Mahboubi (2024), Quality Over Quantity: How Canada’s Immigration System Can Catch Up With Its Competitors, C.D. Howe Institute, Toronto, https://cdhowe.org/publication/quality-over-quantity-how-canadas-immigration-system-can-catch-its (accessed on 2 June 2025).
To download the data for the figures and tables in this chapter, click StatLink above.
Data cut-off for the print publication 13 June 2025.
Table 1. Adults’ mean literacy proficiency, by educational attainment level and gender (2023)
Note: Includes adults who were only administered the doorstep interview due to a language barrier. Columns showing data for all levels of education is available for consultation on line.
Table 2. Distribution of adults by literacy proficiency levels, by educational attainment and gender (2023)
Note: Includes adults who were only administered the doorstep interview due to a language barrier. Columns showing data for men and women, and for all levels of education are available for consultation on line.
Table 3. Adults' mean literacy proficiency, by educational attainment and age group (2023)
Note: Includes adults who were only administered the doorstep interview due to a language barrier. Columns showing data for 35-44 and 55-64 year-olds, and for all levels of education are available for consultation on line.
Table 4. Distribution of adults by literacy proficiency levels, by educational attainment and age group (2023)
Note: Includes adults who were only administered the doorstep interview due to a language barrier. Columns showing data for 35-44, 45-54 and 55-64 year-olds, and for all levels of education are available for consultation on line.
Table 5. Adults' mean literacy proficiency, by educational attainment, immigrant background and language spoken at home (2023)
Note: Includes adults who were only administered the doorstep interview due to a language barrier. Columns showing data for native-born of foreign-born parents and for all levels of education are available for consultation on line.
a – category not applicable; b – break in series; c – there are too few observations to provide reliable estimates; d – contains data from another column; m – missing data; r – values are below a certain reliability threshold and should be interpreted with caution x – contained in another column (indicated in brackets). For further control codes, see the Reader’s Guide.
For further methodological information, see Education at a Glance 2025: Sources, Methodologies and Technical Notes ( https://doi.org/10.1787/fcfaf2d1-en).
Survey of Adult Skills (PIAAC); 25-64 year-olds; in score points
Note: For notes on this table and a link to download the data, see Tables and Notes section above.
Survey of Adult Skills (PIAAC); 25-64 year-olds; in per cent
Note: For notes on this table and a link to download the data, see Tables and Notes section above.
Survey of Adult Skills (PIAAC); in score points
Note: For notes on this table and a link to download the data, see Tables and Notes section above.
Survey of Adult Skills (PIAAC); in per cent
Note: For notes on this table and a link to download the data, see Tables and Notes section above.
Survey of Adult Skills (PIAAC); 25-64 year-olds; in score points
Note: For notes on this table and a link to download the data, see Tables and Notes section above.