copy the linklink copied! 3. The socio-demographic distribution of key information-processing skills

This chapter examines differences in skills proficiency between different groups of individuals, defined by age, gender, socio-economic status, and educational attainment. The chapter considers literacy and numeracy proficiency as well as problem solving in technology-rich environments. It outlines the overall picture for all countries and economies participating in the Survey of Adult Skills, with a particular focus on the differences found among the countries that have participated in this latest round of data gathering.

A note regarding Israel

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

This chapter analyses the levels of proficiency of different subgroups of the population taking part in the Survey of Adult Skills (PIAAC), defined in terms of educational attainment, age, gender and socio-economic background. This information will be especially useful for policy makers wanting to design better and more informed policies. For instance, they could target policies better towards adults with low information-processing skills, in order to reduce disparities and improve human capital. The results could also be used to better understand the strengths and weaknesses of current and past policies, as to a large extent the current level of skills of adults in different age groups, or with different levels of educational attainment, reflect the outcomes of policies that were in place when those adults were attending education.

Figure 3.1 presents an overview of proficiency differences associated with belonging to these different population subgroups. The bars show raw (unadjusted) differences, while the dots represent estimated (adjusted) differences that take into account the role of other background characteristics.1

The main findings discussed in this chapter are:

The differences in proficiency between tertiary-educated adults and those who have not attained an upper secondary education are very large in all countries and economies. Among the countries that participated in Round 3, the differences are especially pronounced in Hungary, Peru and the United States, while they are smaller than the average in Ecuador, Kazakhstan and Mexico.
The literacy and numeracy proficiency of tertiary-educated adults in Hungary is among the highest of all countries and economies that have participated in the Survey of Adult Skills. In Mexico, adults who have not completed tertiary education – both those with upper secondary education and those without – have higher proficiency levels than their similarly educated peers in other Latin American countries taking part in PIAAC.
In Ecuador and Peru, early school leavers (i.e. young adults who are not in education and who have not attained upper secondary education) have very low levels of literacy proficiency. This is of particular concern, given the relatively large size of this group (19% of all 16-24 year-olds in Ecuador and 13% in Peru).
Older adults typically have lower proficiency in literacy and numeracy. This is partly due to ageing, and partly to the fact that younger cohorts are often more highly educated. Indeed, differences between age groups tend to be larger in countries that have more recently expanded access to education, such as Korea and Singapore, and smaller in countries where this process has taken place in the more distant past, such as Germany and the United States.
The age-proficiency profile observed in Ecuador, Mexico and Peru is consistent with this interpretation. In these countries, upper secondary attainment rates have increased only very recently and consequently proficiency tends to decline linearly with age, being highest among 16-24 year-olds. In most other countries proficiency peaks among those aged 25-34.
In Kazakhstan, the increase in tertiary completion rates has not translated into a corresponding increase in the proficiency of the adult population. Proficiency among older adults – aged 55-65 – is almost identical to those of younger adults aged 25-34, in spite of the fact that a much larger share of adults in the latter group have attained a tertiary qualification. This also explains Kazakhstan’s relatively small gap between tertiary-educated adults and those who have not attained an upper-secondary qualification.
Gender gaps in literacy proficiency are small, but they are wider in numeracy, a domain in which men tend to outperform women. Notable exceptions are Hungary and Kazakhstan, which have no numeracy gender gap. The numeracy proficiency of women in Hungary is particularly strong compared with other countries and economies that participated in the Survey of Adult Skills.
Because proficiency is related to educational attainment, and in many countries and economies women had tended to attain lower levels of education than men in the past, gender gaps are more pronounced among older cohorts. This is particularly evident in Ecuador, Mexico and Peru, where gender gaps in numeracy are much smaller among younger adults than they are among older adults.
Adults with more highly educated parents tend to have higher proficiency. Gaps related to family background are particularly pronounced in Hungary, Peru and the United States. Most of these differences are accounted for by individual characteristics, as people with highly educated parents also tend to attain higher levels of education themselves. This is especially true in Mexico, where adjusting for individual characteristics strongly reduces the differences related to family background, but less true in Ecuador and Kazakhstan, where the adjustment has a smaller effect on the size of the gap.

Figure 3.1. Socio-demographic differences in literacy proficiency

copy the linklink copied! Differences in skills proficiency related to educational attainment

Standardised large-scale assessments are popular because they provide comparable cross-country measures of the skills of respondents. This is their main advantage over more traditional (and easier to collect) indicators such as the number of completed years of schooling. Indeed, the Survey of Adult Skills shows large differences in literacy and numeracy proficiency among adults with the same level of education, especially among people who have completed their schooling in different countries.

This is not to downplay the importance of information on educational attainment. As formal schooling is the main (although not the only) vehicle for educating and building the skills of a country’s population, joint analysis of these two sources of information provides essential insights for policy makers who wish to assess the state of their educational systems.

Care should be taken, however, not to interpret differences as the causal effect of education on skills. Even after accounting for a range of observable characteristics, it is likely that some unobservable trait (such as innate ability) influences both proficiency in the PIAAC assessment and educational attainment. The direction of causality would then partly run from skills to education, rather than from education to skills.

Cross-country comparisons should also always be interpreted with some care. As educational systems vary widely, both across countries and over time, the correct interpretation of the relationship between education and skills often requires supplementary information about the history of educational policies in different countries. This issue is made even more complicated by the fact that different countries and economies participated in PIAAC in different years. For a given age group, respondents in the countries participating in Round 3 would have been born and attended education about five years later than adults in countries participating in the first round.

Finally, in order to account for the fact that many of the youngest participants in PIAAC are still in education, the analysis here is mainly restricted to adults aged 25-65 years.

Proficiency in literacy and numeracy among low- and high-educated adults

In all countries and economies, more highly educated adults perform better in the PIAAC assessment (see Figure 3.2). Among the OECD countries and economies that participated in PIAAC, the average difference between tertiary-educated adults and adults with below upper secondary education is 61 score points in literacy and 70 score points in numeracy. Among the countries that participated in Round 3, the differences in both literacy and numeracy proficiency are larger than the OECD average in Hungary, Peru and the United States, and smaller in Ecuador, Kazakhstan and Mexico.

In Hungary, tertiary-educated adults on average scored about 4 points higher in literacy than the average for that level of education across participating OECD countries, and about 18 points higher in numeracy. Hungarians with an upper secondary education also scored higher than the OECD average in numeracy, while the average numeracy score for those without an upper secondary qualification is not statistically different from the OECD average. Hungary has one of the highest shares of tertiary-educated adults scoring at Level 4/5 in numeracy (33%, compared to 23% across the OECD; Sweden has the highest share, at 36%).

Tertiary-educated adults in the United States have similar proficiency in literacy to their Hungarian counterparts, but they scored lower in numeracy, at 284 points, which is below the OECD average of 291 points.

The very small gap between tertiary-educated adults and adults with below upper secondary education observed in Kazakhstan (in both literacy and numeracy) is due to two reasons. First, tertiary-educated adults scored more than 30 points below the OECD average, in both domains. Second, adults without an upper secondary qualification scored above the average, by 6 points in literacy and by 16 points in numeracy.

In Ecuador, Mexico and Peru, performance in literacy and numeracy is consistently below the corresponding OECD average for all levels of educational attainment. Proficiency is especially low among adults without an upper secondary qualification in Peru: their average score was 157 in literacy and 127 in numeracy. This is well below the level registered by similarly educated adults in other Latin American countries that participated in PIAAC, such as Chile (177 score points in literacy and 154 score points in numeracy), Ecuador (174 and 160 score points) and Mexico (201 and 189 score points). Reflecting their low average score, 67% of adults in Peru without an upper secondary qualification scored below Level 1 in literacy and 78% in numeracy, by far the largest share among all countries participating in PIAAC, while Ecuador came second with 50% below Level 1 in literacy and 61% in numeracy.

Figure 3.2. Differences in literacy proficiency, by educational attainment

On the other hand, tertiary-educated adults in Peru have greater literacy and numeracy proficiency than tertiary-educated adults in Ecuador although in both countries they still performed significantly worse than their peers in Chile and Mexico. Adults in Mexico with upper secondary education or below have the highest proficiency in literacy and numeracy among Latin America countries taking part in PIAAC, while the scores for tertiary-educated adults in Mexico are in line with those of tertiary-educated adults in Chile.

Accounting for differences in other socio-demographic characteristics like age, gender and socio-economic background (using parents’ educational attainment as a proxy) reduces education-related differences in proficiency in all countries and economies, although not by a large amount. The reduction is more pronounced in Hungary, Peru and the United States than in most other participating countries.

Proficiency in problem solving in technology-rich environments among low- and high-educated adults

Differences in proficiency related to educational attainment are even larger in the domain of problem solving in technology-rich environments (Figure 3.3). In most countries and economies, a large share of low-educated adults (those without upper secondary education) lacked even the very basic proficiency in the use of information and communications technology (ICT) needed to sit the problem-solving assessment, such as being able to operate a mouse. As a result, 41% of low-educated respondents across OECD countries did not receive a score in this domain. The share not receiving scores was even larger in middle-income countries, approaching 70% in Ecuador and Mexico and surpassing 85% in Peru. This share was only slightly above the OECD average in Hungary, at 47%, in line with the average in Kazakhstan, and much smaller in the United States, at 30%.

Low-educated adults who undertook the problem-solving assessment performed rather poorly. As a result, the overall share of low-educated adults scoring at Level 2 or 3 is very small, averaging 7% across OECD countries. The share is negligible in a large number of countries, including most Round 3 countries with the exception of Hungary and the United States (where about 3% of low-educated adults scored at Level 2) and Kazakhstan (where 6% of low-educated adults scored at Level 2).

Adults with upper secondary education were much more likely to possess the basic ICT skills needed to participate in the assessment: the share who failed the ICT core or had no ICT experience was below 20% in most Round 3 countries. It was substantially higher only in Kazakhstan (at 28%) and in Peru (at 40%). Performance on the assessment, however, was still relatively poor. Among Round 3 countries, only Hungary and the United States approached (but did not reach) the OECD average of 20% of adults with upper secondary education scoring at Level 2 or 3.

Basic ICT skills are nearly universal among tertiary-educated adults and across OECD countries only 4% of adults in this group were not able to participate in the problem-solving assessment. The shares were slightly higher in Ecuador and Mexico (7%), Kazakhstan (9%) and Peru (12%). The differences between countries and economies were larger when it came to demonstrating proficiency in the assessment itself: while on average 48% of tertiary-educated adults across OECD countries scored at Level 2 or 3 in the problem-solving assessment, this share was much lower in many Round 3 countries, including Ecuador (12%), Peru (14%), Kazakhstan (24%) and Mexico (26%). These performances were also below those of other middle-income countries that have participated in previous rounds of PIAAC, such as Chile (where 30% of tertiary-educated adults scored at Level 2 or 3 in problem solving). In contrast, 52% of tertiary-educated adults scored at Level 2 or 3 in problem solving in Hungary and 49% in the United States.

Skills and education among younger adults

The previous analysis has focused on respondents aged 25 and over because younger adults might be still in education or have not yet made important educational choices. As a result, they are hardly comparable to older adults in terms of their highest completed level of education, and deserve a separate analysis.

For the purposes of this analysis, young adults are classified into three groups, defined in terms of the key transition points in the “typical” pathways throughout the education system. The first group is composed of so-called “early school leavers”, i.e. young adults who left formal education without achieving an upper secondary qualification. The second is composed of those who completed upper secondary education, but decided not to enrol in tertiary education. The third group is composed of young adults who are enrolled in tertiary education or who have already completed a tertiary qualification. In the case of this latter group, the analysis is restricted to respondents aged 20 to 24 years, because country differences in the typical age at which students graduate from upper secondary school would generate large (and artificial) differences across countries in the share of 16-19 year-olds who are enrolled in tertiary education.

Figure 3.3. Problem-solving proficiency, by educational attainment

Figure 3.4. Differences in literacy proficiency by educational attainment, young adults aged 16-24

Figure 3.4 shows a large gap in literacy proficiency between early school leavers and other young adults, on average equal to 41 score points across OECD countries. This proficiency gap is close to the OECD average in all the countries participating in Round 3, with two exceptions: Hungary, where it is 58 score points (the second highest among all countries and economies participating in PIAAC, after Chile at 59 points), and Kazakhstan, where it is at 15 score points (the second smallest, after Lithuania at 14 points). The large gap in Hungary is due to the below-average performance of early school leavers (220 points, against an OECD average of 236); on the other hand, the performance of young adults still in education or who have attained an upper secondary qualification is in line with the OECD average, at 278 score points. In contrast, the small gap in Kazakhstan is due to the below-average performance of young adults still in education, whose score of 250 is well below the OECD average (278 points); the performance of early school leavers in Kazakhstan is in line with the OECD average, at 235 score points.

The literacy proficiency of young early school leavers in Ecuador and Peru is the lowest among all participating countries and economies, at 173 and 176 score points respectively, slightly below the average score registered in Chile (184 score points). Early school leavers in Mexico also performed below the OECD average, at 211 score points, while their proficiency in the United States is slightly above the average, at 241 score points. In practice, the relevance of this gap depends on the size of the population of early school leavers. Mexico stands out in this respect, as 36% of its 16-24 year-olds can be classified as early school leavers. Leaving school early is less widespread in Ecuador (19% of 16-24 year-olds), Hungary and Peru (both at 13%), although still above the OECD average of 11%. In the United States the share is lower than average, at 9% (see Table A3.14 in Annex A).

Young adults currently enrolled in tertiary education or with a tertiary qualification have higher levels of proficiency than respondents in the same age group with at most an upper secondary qualification and not enrolled in tertiary education (see Figure 3.5). The average gap in literacy proficiency between these two groups is 35 score points across OECD countries. The gap is larger than the average in all Round 3 countries with two exceptions: the United States (32 score points), and Kazakhstan (15 score points, the lowest among all countries and economies participating in PIAAC). This small gap observed in Kazakhstan is mainly due to the low performance of young adults enrolled in tertiary education or with a tertiary qualifications, who score almost 40 points below the OECD average (257 compared to 295 score points).

The gap is especially pronounced in Hungary (at 45 score points), where young adults enrolled in tertiary education or with a tertiary degree scored above the OECD average, at 300 score points) while those who did not go on to tertiary education scored below the average (255 compared to 260 score points). In Ecuador and in Peru the proficiency of both groups is the lowest among all participating countries and economies. Young adults enrolled in tertiary education averaged 236 score points in Peru, compared to 232 in Ecuador, while the picture is reversed for young adults not enrolled at tertiary level: 194 score points in Peru compared to 195 in Ecuador.

copy the linklink copied! Differences in skills proficiency related to age

One of the primary objectives of the Survey of Adult Skills is to shed light on the mechanisms that drive the evolution of skills over people’s lifetimes. This is not an easy task, as proficiency is simultaneously influenced by a variety of factors that are not easily distinguishable. These factors can be classified in three broad categories: 1) investments in skills, in the form for instance of formal education or adult training; 2) biological processes that drive cognitive functioning; and 3) life experiences, such as employment status or personal interests, that lead to more or less intense practice of skills at work or in daily life. Importantly, these factors are likely to be interdependent: education can affect labour-market opportunities and, therefore, the use of skills at work, and even the biological process of ageing can have different consequences depending on life circumstances and the intensity of skills use.

In cross-sectional data like those collected in the Survey of Adult Skills, observed differences in proficiency by age are inevitably the combined result of age effects (i.e. the consequences of growing older), cohort effects (i.e. the consequences of being born in a particular year, and therefore being exposed to experiences such as a particular type of education), and period effects (i.e. shocks that take place at a given point in time and affect all cohorts in the same way). Controlling for observable differences across individuals born in different years (notably for differences in the level of education) can help to identify age effects, assuming that the quality of education does not change over time. While there is some evidence that this assumption might not hold (Paccagnella, 2016[1]; Paccagnella, 2016[2]), in most countries the observed cross-sectional differences are likely to provide a reasonable approximation of the underlying age effects.

This conclusion, however, is necessarily country specific. Cross-sectional differences are in fact greatly influenced by the timing and speed of a country’s expansion of educational attainment. This is a process that is common to most countries in the world, but that has occurred in different countries at different times. This is especially relevant when analysing data from countries at different stages of economic development (see also Chapter 3 of OECD (2016[3])).

Figure 3.6 shows the relationship between age and proficiency for the Round 3 countries as well as for the average of all countries and economies that have participated in the Survey of Adult Skills. There is a clear negative correlation between age and proficiency, which is present in all countries, with the partial exception of Kazakhstan and the United States. A possible explanation for this is that the expansion of educational attainment took place much earlier in the United States than in the other Round 3 countries. Indeed, 55-65 year-olds in the United States are more likely than in many other countries to have a tertiary degree, and the gap in educational attainment between 25-34 year-olds and these older adults is very small (see Figure 2.2 in Chapter 2).

Figure 3.5. Differences in literacy proficiency by educational attainment, young adults aged 20-24

Figure 3.6. The relationship between skills proficiency and age

This explanation does not hold in the case of Kazakhstan, however. Tertiary education rates for 55-65 year-olds in Kazakhstan are about half the rates observed among 25-34 year-olds (27% compared to 50%). The share of adults who have not attained an upper secondary qualification is similar in the two age groups (14% for older adults and 11% for 25-34 year-olds), meaning that, over time, an increasing share of adults have progressed from a secondary to a tertiary qualification. This upgrade in educational attainment does not appear to have translated in a corresponding increase in the skills of the adult population, possibly because of a decline in the quality of education. This hypothesis is plausible in light of the fact that, in contrast with older adults, 25-34 year olds performed significantly below the OECD average for the same age group.

In most countries and economies, the relationship between age and proficiency is an inverted U-shaped curve, with a peak between the mid-twenties and the early-thirties. In contrast, in Ecuador, Mexico and Peru, proficiency declines more or less linearly with age. This is probably due to the fact that upper secondary attainment rates in these countries have increased only very recently. On average across OECD countries, only 15% of 25-34 year-olds have not completed their upper secondary education, compared to 50% in Mexico, 36% in Ecuador and 26% in Peru. When looking at adults under 25, the share of respondents who have completed upper secondary education is actually close to the OECD average of 59% in Ecuador (57%) and higher in Peru (76%), and is not very distant in Mexico (43%). This is a good example of how the relationship between age and proficiency in cross-sectional data is influenced by cohort effects, such as different cohorts experiencing the effects of different education policies.

Proficiency in literacy and numeracy among older and younger adults

Figure 3.7 presents the average literacy score of adults in different age bands and the average difference between 55-65 year-olds and 25-34 year-olds. In most countries, these two groups have the lowest and highest average scores respectively in literacy. In a few countries adults aged under 25 have the highest average proficiency, but the differences with respect to 25-34 year-olds are very small (with the partial exception of Peru, where the youngest group of adults scored 10 points more on averages).

Literacy proficiency among older adults is lowest in Peru (175 score points) and Ecuador (181 score points). In both these countries, about half the adult population aged 55-65 scored below Level 1, and about one-third at Level 1. Proficiency among older adults in Mexico is only slightly higher: they averaged 197 score points although one-third still scored below Level 1, and one-third at Level 1. Average scores are much higher in Hungary (246 score points), Kazakhstan (249 score points, in line with the OECD average) and the United States (264 score points).

Ecuador and Peru recorded the lowest average proficiencies also for adults aged 25-34: 202 score points in Ecuador and 203 score points in Peru. Two-thirds of these adults in those countries scored at or below Level 1 in literacy. Average proficiency is significantly higher in Mexico (230 score points) and Kazakhstan (249 score points), and on a par with the OECD average in Hungary (276 score points) and the United States (279 score points).

There is much less variation across countries in the size of the differences between the older and younger age groups. In most countries, including Hungary, Mexico and Peru, the gap in literacy proficiency between the two groups is around 30 score points. The gap is smaller in Ecuador, at 21 score points, in the United States, at only 15 score points and in Kazakhstan (where the two groups recorded the same average proficiency).

In most countries and economies, about half of the gap can be accounted for by differences in observable characteristics, notably in educational attainment, as younger cohorts are normally more educated than older cohorts. This effect is even stronger in Ecuador and Peru, where the adjusted differences are as small as 8 and 7 points, respectively. Both countries have experienced a significant expansion of basic education recently, more pronounced than in OECD countries. While 58% of older adults in Peru, and 68% in Ecuador, have not completed upper secondary education, these shares drop to 26% and 36%, respectively, among 25-34 year-olds. On average across OECD countries the differences between the two groups are smaller (20 percentage points), because educational expansion occurred earlier. In Germany, for instance, the shares of adults without upper secondary education and with a tertiary education are almost the same in the two age groups, and as a result adjusted differences are nearly identical to the unadjusted ones. Similarly, in the United States, where the difference in the share of tertiary-educated adults between the two age groups is only 7 percentage points (compared to a 16 percentage-points difference across OECD countries), adjusting for education has a minor effect on the size of the gap (which decreases from 15 to 9 points). As discussed above, it is hard to identify how much of the remaining (adjusted) gap captures an age effect and how much it captures cohort or period effects, which might be due to a number of reasons, such as changes in the quality of education over time. This seems to be a plausible explanation in the case ok Kazakhstan, where older adults, despite being less educated than the younger cohorts, have the same level of literacy proficiency. As a consequence, after accounting for the different level of educational attainment, the estimated gap turns out to be negative, meaning that older adults would be more proficient than 25-34 year-olds, had they attained the same level of education.

Figure 3.7. Age differences in literacy proficiency

Proficiency in problem solving in technology-rich environments among older and younger adults

As the assessment of problem solving required test takers to have some basic familiarity with digital devices and applications, it might be expected that age differences would be more pronounced in this domain. Indeed, Figure 3.8 shows particularly low levels of proficiency in problem solving in technology-rich environments among older adults. In Ecuador, Mexico and Peru, the vast majority of older adults (55-65 year-olds) did not undertake the problem-solving assessment because they lacked the necessary ICT experience or failed a very basic test of ICT skills. Virtually no older adults demonstrated problem-solving proficiency at Level 2 or 3 in these three countries as a result. Proficiency in problem solving in technology-rich environments was on a par with the OECD average for older adults in Hungary, and well above the average in the United States.

When looking at younger adults, it is more interesting to look at their actual level of proficiency, rather than at the difference between them and older adults. Digital technologies might not yet be as widespread in middle-income countries as they are in more developed ones, but there is little doubt that a fast adoption of digital technologies will be essential if countries are to reap the benefits of globalisation, technological change and economic integration (OECD, 2019[5]).

From this perspective, it is concerning to see very low levels of proficiency in using ICT among younger adults in a number of countries. Mexico and Peru have the highest shares of 25-34 year-olds failing the core ICT test or having no ICT experience (28% in Mexico and 33% in Peru), similar to what is observed in Turkey (27%). Ecuador does slightly better, at 22%, but is still well behind Chile (13%). The share of this age group scoring at Level 2 or 3 is only 7% in Ecuador, 9% in Peru and 13% in Mexico.

Lack of basic ICT skills or experience was less common in Kazakhstan, where only 10% of 25-34 year-olds failed the ICT core or had no ICT experience. However, proficiency in problem solving in technology-rich environments remained well below the OECD average, with only 18% of adults in this age group scoring at Level 2 or 3. In both Hungary and the United States the share of these younger adults at Level 2 or 3 is close to 40% (in line with the OECD average), and the share who failed the ICT core or lacked ICT experience is below average, at 5% in Hungary and 4% in the United States.

When looking at the youngest adults (16-24 year-olds), however, there are some signs of improvement over time (i.e. comparing the older to younger cohorts) in the lowest performing countries. While the share of 16-24 year-olds scoring at Level 2 or 3 remains very low (only in Mexico does it reach 18%), they are less likely than 25-34 year-olds to opt out of the computer based assessment, fail the ICT core or to lack ICT experience. No clear sign of progress is apparent in Kazakhstan, where the performance of 16-24 year-olds was very close to that of 25-34 year-olds.

copy the linklink copied! Differences in skills proficiency related to gender

Proficiency in literacy and numeracy among men and women

Gender differences in literacy and numeracy skills are typically small, and countries that participated in Round 3 of PIAAC are no exception (Figure 3.9). Peru is the only Round 3 country where the gender gap in numeracy is above the OECD average; Peru’s gap of 16 score points might appear small when compared with differences related to educational attainment or age, but is not negligible in size, as this represent almost one-third of the international standard deviation. Peru is also the only Round 3 country (and one of the few countries among all those that participated in PIAAC) where women have a statistically significant (although small) advantage in literacy.

Hungary and Kazakhstan are among the few countries where there is no gender difference in numeracy proficiency. Hungarian women scored 271 points on average, 15 points above the OECD average (while Hungarian men scored 274 points, only 6 points above the average), although they still lag slightly behind top-performing countries like Japan (282 score points), Finland (277 score points) and the Slovak Republic (275 score points). In Kazakhstan both men and women score below the OECD average, but the gap is much narrower for women, at only 9 score points

Gender differences are typically more pronounced among older adults (Figure 3.10). This is mainly due to two reasons. The first is that the educational attainment of women has progressively caught up with that of men. The second is that women and men still tend to make different occupational choices, or have different labour-market outcomes for a given level of education, which might affect the extent to which they have the opportunity to practise and maintain their level of proficiency. This is particularly evident in Peru, where gender gaps in numeracy are as high as 19 score points (among the highest across all PIAAC participating countries and economies) for adults aged 25 and over, but are as low as 5 score points (below the OECD average) among adults aged 24 and under. This is probably due to the fact that only among these younger adults do women have roughly similar levels of educational attainment to men. In all age groups over 25, women are over-represented among adults without an upper secondary education and under-represented among upper secondary educated adults. A similar pattern (although on a smaller scale) is evident for gender gaps in literacy as well.

Figure 3.8. Problem-solving proficiency among younger and older adults

In Ecuador, gender gaps in numeracy are also wider than the OECD average for adults aged 45 and over, but are much smaller for adults below that age. A similar pattern is observed in Mexico, although the gender gap among older adults is in line with the OECD average. In Hungary, on the other hand, gender gaps are essentially identical (and extremely small) across the different age groups, in both literacy and numeracy. The numeracy proficiency among women aged 45 and over in Hungary is almost 20 points above the OECD average, and one of the highest overall.

Figure 3.9. Gender differences in literacy and numeracy proficiency

Figure 3.10. Gender gap in literacy and numeracy, by age

Proficiency in problem solving in technology-rich environments among men and women

Gender differences are also not particularly pronounced in the domain of problem solving. Men have a slight advantage. On average across OECD countries, 32% of men score at Level 2 or 3, compared to 28% of women, although a similar share of men and women have no computer experience or have failed the ICT core test (Figure 3.11).

In Ecuador and Peru gender gaps are less pronounced at the top of the distribution: the percentage of men scoring at the two highest levels is 6% in Ecuador and 7% in Peru, while the corresponding shares for women are 4% in Ecuador and 6% in Peru. At the bottom end of the distribution, a larger share of women have no computer experience or fail the ICT core test: 35% in Ecuador and 47% in Peru, compared to 31% and 41% for men.

In Hungary and the United States men are over-represented at both the top and the bottom of the proficiency distribution. The share of men at Level 2 and 3 is 2 percentage points higher than the share of women in Hungary and 3 percentage points in the United States, while at the lower end (respondents with no computer experience or failing the core assessment), the share of men is 1 percentage point higher in Hungary and 2 percentage points higher in the United States.

In Kazakhstan no gender differences are observed in this domain, as the share of adults scoring at the different levels is nearly identical for both men and women.

Figure 3.11. Problem-solving proficiency among men and women

Figure 3.12. Differences in literacy proficiency, by parents’ educational attainment

copy the linklink copied! Differences in skills proficiency related to socio-economic background

It is well known that the family background in which people grow up is a crucial factor affecting their outcomes such as educational attainment or earnings. Given that people cannot choose their family background, reducing its impact on outcomes is often considered an important policy objective, in order to increase equality of opportunity.

Proficiency in literacy and numeracy among adults with high- and low-educated parents

The best proxy for socio-economic background available in the Survey of Adult Skills is the level of educational attainment attained by the parents of the participants in the Survey. Using this indicator, the results confirm the importance of socio-economic background on adults’ outcomes: across the three rounds of the study, adults with at least one tertiary-educated parent score on average 41 points more than adults from families in which neither parent attained upper secondary education (Figure 3.12).

The countries that participated in Round 3 are no exception. In Ecuador and Mexico the differences are very close to the OECD average, at 41 score points; they are higher in the United States (52 score points), Peru (55 score points), and Hungary (57 score points); and they are smaller (but still significant) in Kazakhstan, at 20 score points.

Much of these raw differences are accounted for by differences in other personal characteristics, as the effect of socio-economic background mainly works through the intergenerational transmission of educational attainment: adults with highly educated parents are more likely to attain higher levels of education themselves. In most countries, the adjusted differences are about half the size of the unadjusted ones. Kazakhstan is an exception, as adjusting for observable characteristics has almost no impact on the size of the socio-economic gap. Adjusted differences are slightly below the OECD average in Mexico (18 score points), and remain higher than the average in Ecuador (24 score points), as well as in the United States (27 score points), Hungary and Peru (29 score points).

copy the linklink copied! Summary

This chapter has highlighted the proficiency levels of different subgroups of the population, defined according to a number of socio-demographic characteristics. The analysis has confirmed a number of expected results, such as the association between proficiency and educational attainment, the age-proficiency profile, and the extent to which men and women tend to perform differently in different domains.

While these associations hold across most countries and economies, a number of peculiarities have emerged from the analysis, and some can be traced back to the individual history of development and the policies adopted in different countries.

For example, the analysis has shown that, while Latin American countries in PIAAC tend to have lower performance across the board, they seem to be benefiting from the recent expansion in access to education, as more highly educated young cohorts show greater proficiency than older adults. This is not the case in Kazakhstan, where the expansion of tertiary education has not brought about the expected benefits in terms of higher proficiency. At the same time, young early school leavers in Ecuador and Peru demonstrated an extremely low level of proficiency, which calls for targeted policies to address the needs of this particularly vulnerable group of people.

Adults in Hungary, on the other hand, tend to score roughly at the same level as the OECD average and outperform many countries when the analysis is restricted to tertiary-educated adults. Moreover, Hungary stands out as a country where there are no gender gaps in numeracy, thanks to exceptionally strong performance of Hungarian women in that domain.

copy the linklink copied!

A note regarding the Russian Federation

The sample for the Russian Federation does not include the population of the Moscow municipal area. More detailed information can be found in the Technical Report of the Survey of Adult Skills, Third Edition (OECD, 2019[6]).

References

[5] OECD (2019), OECD Skills Outlook 2019: Thriving in a Digital World, OECD Publishing, Paris, https://doi.org/10.1787/df80bc12-en.

[6] OECD (2019), Technical Report of the Survey of Adult Skills, Third Edition, OECD, http://www.oecd.org/skills/piaac/publications/PIAAC_Technical_Report_2019.pdf.

[3] OECD (2016), Skills Matter: Further Results from the Survey of Adult Skills, OECD Skills Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264258051-en.

[4] OECD (2013), OECD Skills Outlook 2013: First Results from the Survey of Adult Skills, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264204256-en.

[1] Paccagnella, M. (2016), “Age, ageing and skills: Results from the Survey of Adult Skills”, OECD Education Working Papers, No. 132, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jm0q1n38lvc-en.

[2] Paccagnella, M. (2016), “Literacy and numeracy proficiency in IALS, ALL and PIAAC”, OECD Education Working Papers, No. 142, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jlpq7qglx5g-en.

Note

← 1. Adjusted differences are estimated through an ordinary least squares regression (one regression per country). The dependent variable is the literacy (or numeracy) score of each individual respondent, and the independent variables are the gender, age class, educational attainment, socio-economic background (measured by the highest level of education attained by either parents), and immigrant and language background. The dots in Figure 3.1 report the estimated regression coefficient for the various independent variables.

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

https://doi.org/10.1787/1f029d8f-en

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at http://www.oecd.org/termsandconditions.

╳

OECD Skills Studies: Skills Matter: Additional Results from the Survey of Adult Skills

Do you want to open the publication page on OECD-ILIBRARY for more details on:

available formats

purchase of a print copy

or to publication metadata?

Cancel

3. The socio-demographic distribution of key information-processing skills