The Two Skills That Shape Different Futures
This series examines the math–verbal divide: the different ways mathematical and verbal skills are connected to education, fields of study, work, and earnings. Its central question is not which skill is simply “better,” but why the same early skill profile can open one kind of opportunity and pay off through another. Understanding that distinction gives readers a better way to interpret education statistics, gender gaps, and policy choices without turning an average pattern into a verdict about any individual.
Imagine two students who are both doing well at school. One is especially strong at understanding and expressing language; the other is especially strong at quantitative reasoning. It would be easy to assume that the student with the stronger academic profile will have better outcomes everywhere. But “better outcomes” can mean several different things: qualifying for the next stage of education, finishing a degree, choosing a particular subject, finding work, or earning more once employed. Those are related outcomes, not interchangeable ones.
The evidence assembled here makes that separation unusually clear. In a large English study that follows school cohorts into adulthood, verbal skill is more strongly associated with educational attainment. Mathematics skill, however, has the larger association with earnings in the early thirties. The difference is not a paradox produced by one number contradicting another. It reflects a pathway: skills influence educational decisions and destinations, and those destinations meet institutions and employers that reward some combinations of knowledge more than others.
One word, “skill,” hides two different capacities
Mathematical and verbal skills overlap, but they are not the same measurement. In the English study, mathematics is represented by performance in GCSE mathematics, while verbal skill is represented by English performance, combining language and literature grades. The researchers convert grades into ranks within each exam cohort. A person’s rank therefore says how they performed relative to other students taking exams in the same period; it is not a timeless measure of an innate quantity called intelligence.
That distinction matters because the two domains connect to different tasks. Quantitative skill can be useful for working with rates, proportions, formal relationships, and data. Verbal skill can be useful for understanding difficult language, communicating an argument, and navigating institutions whose requirements are expressed through text. Neither description exhausts what people can do, and neither predicts a life on its own. The point is narrower and more useful: a school system and a labour market can place different demands on different parts of a person’s skill profile.
The relevant comparison is often comparative rather than absolute. A student may be capable of succeeding in more than one field, yet find one route more attractive or attainable because it fits their relative strengths. That choice can then expose them to a different set of completion rates, occupations, and pay levels. A difference in outcomes may therefore arise not because one student possesses a universally higher ability, but because skill profiles guide people toward different environments.
This is why the math–verbal divide should not be read as a contest between two kinds of intelligence. It is a question about how measured capacities travel through a sequence of institutions. At one stage, reading and communication may help someone meet educational requirements or persist through them. At another, quantitative expertise may be more directly rewarded in fields and jobs with high earnings. Strong ability in both domains may be especially valuable where analysis must be explained to other people, but that combination need not be necessary merely to obtain employment.
The puzzle: the route to college is not the route to pay
The English evidence follows people who completed their GCSEs between 2002 and 2004. It records educational milestones and later labour-market outcomes for the same cohorts, rather than comparing one group of students with a separate group of workers. The results show a clear first contrast. After the researchers account for measurement error, verbal skill has the larger estimated relationship with taking A-levels, entering college, graduating, earning a good degree, and pursuing postgraduate study.
College attendance is the most intuitive example. The paper translates a ten-percentile-point increase in the skill rank into an attendance increase of about seven percentage points for verbal skill, compared with about three points for mathematics. The exact coefficient depends on the outcome and specification, but the direction is consistent: verbal skill is closely tied to crossing and progressing through educational gates.
Yet educational participation is not the final destination. Students choose subjects and programmes, and those programmes differ in their typical completion rates and earnings returns. The study finds that stronger mathematics predicts entry into STEM and other higher-return fields, while stronger verbal skill predicts entry into humanities, arts, and social-science fields and other lower-return destinations on the relevant comparisons. Verbal skill can thus support more education while also being associated with choices whose average labour-market payoff is smaller. Mathematics can have a weaker relationship with entering college while guiding those who do enter toward fields where earnings are higher.
This gives us three distinct questions:
- Who gets through the educational gate? Verbal skill is especially important in the English estimates.
- Where do those students go? Mathematics and verbal skill point toward different fields, consistent with comparative advantage.
- What does the destination pay, and when? Mathematics has the stronger later earnings association, while verbal returns are shaped by education timing and occupational sorting.
The distinction also changes how we read gender differences. Women in the English cohorts are more likely than men to attend college, and their relative advantage in verbal skill explains much of that difference in the study’s specifications. The gender gap in STEM participation is also partly reduced when both skills are taken into account. But the gender earnings gap does not shrink in the same way. The education benefit associated with stronger verbal skill and the lower average returns associated with some chosen fields work in opposite directions. A single “gender ability gap” cannot capture all three outcomes.
The same logic applies to policy. If the goal is wider access to higher education, verbal skill may be a particularly important margin. If the goal is higher earnings, mathematics and the field choices associated with it may matter more. If the goal is employment among people at the bottom of the skill distribution, basic proficiency may be more consequential than high-end specialization. If the goal is a productive and satisfying job, the match between a person’s skills, qualifications, field, and actual work also matters. These are different objectives, so they should not be evaluated with one universal score.
The diagram is an analytical map, not a deterministic pipeline. It says that skills can operate through several links and that context can shape each link. It does not say that a particular student with a particular score must attend a particular college, choose a particular major, or earn a particular wage. Nor does it say that the route from skill to pay is entirely indirect. The later evidence suggests that mathematics continues to matter after education and field are considered, while verbal skill remains valuable in some communication-intensive work.
What this comparison can—and cannot—establish
The core English study is unusually informative because its skill measures precede labour-market entry. It uses Key Stage 2 assessments taken around age 11 as instruments for GCSE mathematics and English measures taken around age 16. In simple terms, the earlier assessments help address the problem that exam grades are imperfect measurements and might otherwise be entangled with later experiences. The researchers then follow a population-scale set of records through education and into employment, with earnings observed annually for the relevant cohorts.
That design improves the ordering of the story: the measured skills come before the educational and work outcomes. It does not remove every uncertainty. The interpretation relies on assumptions about how the earlier assessments relate to the later measures and outcomes. The sample also excludes people who emigrated and people who did not sit their GCSEs in an English school. Most importantly, the results describe estimated relationships for particular English cohorts under a specified design. They are not promises about what every individual will earn after studying mathematics or English, and they do not by themselves test a particular classroom intervention.
The second source broadens the setting but changes the question. The OECD’s Survey of Adult Skills reports literacy, numeracy, and adaptive problem solving among US adults aged 16 to 65 in 2022–23. These are adult proficiency domains assessed through tasks designed to reflect complex, often digital situations. They are related to verbal and mathematical capability, but they are not the same as GCSE English and mathematics ranks. The US survey is useful for showing the scale and distribution of adult skills, their associations with wages and wellbeing, and the costs of mismatch between qualifications and jobs. It is not a continuation of the English administrative cohorts or a direct replication of the IZA estimates.
Keeping those boundaries visible is part of understanding the subject. A result can be important without answering every nearby question. The English evidence is strongest for the pathway from adolescent subject-specific performance through education, field choice, and early adult labour-market outcomes in England. The OECD note is strongest for contemporary US adult proficiency and the broader conditions under which skills are used or left underused. Combining them as if they measured one population would create a false comparison; placing them side by side can instead reveal which parts of the pathway are specific to schooling and which concern adult life more generally.
The central lesson is therefore methodological as well as substantive. Do not ask only whether mathematics or verbal skill has the larger “return.” Ask which return: access to education, completion, choice of field, employment, earnings, wellbeing, or a good match between capability and work. The answer changes with the margin being measured. To read the evidence responsibly, we first need to examine how the two studies turn skills into scores, and how those scores become estimates rather than direct observations of a person’s future.