STEM Education Needs Improvement in 2025

Science, Technology, Engineering and Mathematics (STEM) underpin the knowledge economies of today. STEM graduates command high wages and STEM jobs are projected to grow more than twice as fast as non‑STEM occupations. Yet educational outcomes and pathways have not kept pace with economic demand.

Women, racial minorities and students from low‑income backgrounds remain under‑represented in STEM fields; many students struggle to reach basic proficiency in mathematics and science; and teachers trained to teach these subjects are in short supply.

These persistent gaps limit innovation, weaken economic competitiveness and deprive young people of opportunities.

But what causes these gaps exactly? Since we focus primarily on data from the United States, the examples provided will reflect this context. However, many of the findings are equally relevant to both EU and non-EU countries.

UPDATE: In this article, New EU Approach to Education: A Skills-Based Curriculum Rules, I also explain how the EU has chosen for a skills-based curriculum, a necessity in STEM education.

Representation and equity in STEM

Not all students are represented in STEM education in the same way. There are clear gaps that need to be addressed. Diverse STEM teams produce more innovative solutions, but the current pathways exclude many who have both ability and interest. Under‑representation also widens wage gaps; in the US STEM workers earn median wages over $103 k compared with $48 k in non‑STEM jobs.

But let’s see which groups are underrepresented.

Gender disparities

Women’s participation in STEM has stagnated. Globally only 35 % of STEM graduates are women – a figure that has not changed in a decade. In the United States for instance, women earn 42 % of bachelor’s degrees in mathematics and statistics, 25 % of physics degrees and 23 % of engineering degrees. Although these shares have improved since the 1980s, women still hold only 18 % of STEM occupations.

Researchers identify several reasons: lack of access to education and job opportunities; masculine work cultures and differential treatment; insufficient positive experiences during education; and self‑selection by men into certain fields. Surprisingly since we are in 2025, stereotypes and the absence of female role models discourage girls from pursuing STEM careers.

Racial and ethnic under‑representation

In the United States, African Americans and Hispanics make up 11 % and 17 % of the workforce but only 7 % of STEM workers. The 2024 Black Students and STEM Report show enormous gaps between aptitude and exposure: 75 % of Black students with aptitude in advanced manufacturing lack exposure to the field; similar exposure gaps exist in health sciences (57 %), finance (56 %), architecture (53 %) and computing (51 %).

Black female students show even larger differences – e.g. 88 % more have aptitude than interest in advanced manufacturing. Community colleges serve a diverse student body (51 % students of colour) and are vital for diversifying the STEM workforce. Without targeted mentoring, scholarships and industry partnerships, these gaps will persist.

Other inclusion gaps

Persons with disabilities face barriers that are poorly measured. UNESCO warns that biases and social norms limit the opportunities of girls and women in STEM, and the AI literacy gap is widening a new “AI divide” that disproportionately affects women, people of colour and disabled individuals.

Representation and exposure (selected data)

Metric/Group (Year)	Key finding
Women worldwide	Only 35 % of STEM graduates are women, a level unchanged for a decade.
Women in U.S. bachelors’ degrees (2020)	Mathematics & statistics: 42 %; physics: 25 %; engineering: 23 %.
Women in the U.S. STEM workforce (2021)	Women occupy 18 % of STEM jobs.
Black & Hispanic workforce share vs. STEM share (2021)	African Americans: 11 % of workforce but 7 % of STEM workers; Hispanics: 17 % of workforce but 7 % of STEM workers.
Exposure–aptitude gap for Black students (2024)	75 % gap in advanced manufacturing, 57 % in health sciences, 56 % in finance, 53 % in architecture, 51 % in computers/technology.

Teacher shortages and quality

The quality of STEM education is directly tied to the availability of qualified educators. Yet, across both high-income and low-income regions, teacher shortages persist – particularly in mathematics, science, and digital technology subjects. Over 411,500 teaching positions in the United States alone were either vacant or filled by instructors without full certification. This trend is echoed globally, with 48 countries reporting shortages in at least one STEM subject.

The impact is most pronounced in high-need and rural areas, where turnover rates are highest and resources are most limited.

Let’s look into detail what is really going on.

National teacher shortages (United States)

The Learning Policy Institute’s 2025 scan of state‑level data paints a rather bleak picture: an estimated 365 967 teachers were not fully certified for their assignments and 45 582 teaching positions were vacant. Together these 411 549 positions represent 1 in 8 of all teaching positions.

The shortage affects more than 6 million students. Thirty‑five states and the District of Columbia saw increases in the number of teachers not fully certified between the 2023 and 2025 scans, while the number of unfilled positions increased in 12 of 18 reporting states.

This shortage hits STEM fields hardest. For the 2024‑25 school year, 45 states reported shortages in special education, 41 in science and 40 in math. Nearly every state reported shortages in at least one of these areas, and the deepest vacancies were in special education, elementary education, language arts and career and technical education.

Causes of the shortage

Low entry into the profession: Interest in teaching among high‑school and college students is at its lowest level in decades. U.S. teacher preparation enrolments fell by about 100 000 candidates between 2012–13 and 2014–15; from 2016–17 to 2020–21, 27 states saw further declines of 5 % or more.

High attrition: Attrition accounts for 90 % of annual teacher demand. Fewer than one‑fifth of teachers leaving the profession retire; most cite low pay, dissatisfaction or better opportunities elsewhere. Each teacher departure costs districts $12 000–$25 000 in recruitment and training. Rural districts are especially affected due to smaller tax bases and distance from preparation programs.

Pipeline diversity and preparation: Women now make up a larger share of STEM teachers compared with a decade ago, and more new teachers hold comprehensive credentials. However, the total number of STEM teachers produced annually in the U.S. has fallen from about 31 000 to 20 000 over the past decade. Research shows that teacher residencies and mentoring increase retention, yet many districts lack funding for these programmes.

Mississippi for instance reduced the number of teachers on emergency or provisional licences by 54 % between 2021–22 and 2023–24 but still reported nearly 3 000 unfilled positions due to a high turnover rate above 23 %. Arizona saw a 60 % rise in teachers not fully certified from 2020–21 to 2023–24 and had 2 261 unfilled positions by September 2024.

International context

Teacher shortages are not unique to the United States. Data from the Trends in Mathematics and Science Study (TIMSS) show that almost 30 % of schools worldwide lack adequate mathematics and science teachers; in Malaysia and Türkiye that figure exceeds 80 %. In sub‑Saharan Africa only 30 % of short‑cycle tertiary enrolment is in STEM (25 % for females and 34 % for males), making it impossible to meet future teacher needs even if all graduates went into teaching.

A global shortfall of 6 million computing and mathematics workers is expected in the United States and 1 million in Germany by 2030. Countries have experimented with salary supplements, alternative certification pathways and regional training hubs such as the African Institute for Mathematical Sciences (AIMS) to improve recruitment and retention.

Teacher shortage overview in the U.S.

Measure (U.S., 2025)	Data	Implications
Teachers not fully certified	365 967 teachers without full certification	Suggests large reliance on emergency/temporary credentials
Vacant teaching positions	45 582 positions	Vacancies often filled by substitutes, larger class sizes
Total positions affected	411 549 positions, ≈ 1 in 8 of all teaching posts	More than 6 million students impacted
States reporting special‑education shortages	45 states	Chronic shortage since at least 1990
States reporting science shortages	41 states	STEM teachers in short supply
States reporting math shortages	40 states	Undermines algebra & calculus pathways
Annual production of STEM teachers	Fell from ≈31 000 to ≈20 000 per year	Fewer new teachers entering STEM fields

Skills and curriculum gaps

Despite growing demand for STEM competencies, many national education systems continue to deliver outdated or uneven curricula. Key 21st-century skills – such as data literacy, computational thinking, and interdisciplinary problem-solving – are very often underrepresented.

Here’s what the status is today, and we will use EU data for it as this has been very well researched.

Low proficiency and declining performance

The European Commission’s 2025 preview of the STEM Education Strategic Plan warns that many EU students fail to reach minimum proficiency in mathematics and science. Covid‑19‑related disruptions deepened inequalities; performance deteriorates when pupils move from integrated primary STEM to discipline‑specific secondary courses.

As a result we also see that the share of top performers is shrinking while under‑achievement is concentrated among disadvantaged students. PISA 2022 results echo this trend: the United States ranked 28th among 37 OECD countries in mathematics and 12th in science; the average U.S. math score dropped by 13 points since 2018. Only 28 % of Americans rate U.S. K‑12 STEM education as above average.

Early introduction and curriculum integration

Survey data from Ireland show that educators value early exposure: 66 % rate teaching STEM and digital skills as very or extremely important (up from 60 % in 2024). 40 % of respondents believe STEM should begin at ages 4–6, and 65 % of institutions teach STEM and digital skills (66 % of primary schools, 75 % of secondary schools, 33 % of youth centres and 57 % of libraries).

However, 81 % of educators cite lack of knowledge/training, 76 % cite insufficient resources/funding and 54 % cite lack of time as barriers. Only 15 % of institutions report sufficient STEM funding.

The EU report further notes that integrated STEM/STEAM approaches are gaining popularity but are defined inconsistently, complicating curriculum design and evaluation. Longitudinal research on the effectiveness of these approaches remains scarce unfortunately.

Barriers at the classroom level

Four‑fifths of Irish educators report lacking training to teach STEM; three‑quarters cite insufficient resources or funding; over half mention time constraints. Similar patterns appear internationally – many teachers lack expertise in computer science or advanced mathematics. Without targeted professional development, technology integration can and definitely will widen achievement gaps.

Many systems focus on memorization rather than inquiry. Secondary courses often silo math, science and technology, missing opportunities to show real‑world connections. Integrated approaches and project‑based learning remain exceptions rather than the norm.

Digital divide and skills gap

Access to digital infrastructure and skills training remains a major barrier to equitable STEM education. The digital divide – defined by disparities in connectivity, device availability, and digital literacy – disproportionately affects rural schools, low-income households, and marginalized communities.

Digital skills demand vs. supply

The National Skills Coalition (NSC) found that 92 % of U.S. jobs require digital skills, yet one‑third of workers possess low or no digital skills. Workers with at least one digital skill earn 23 % more; those with three or more digital skills earn 45 % more. Demand spans all sectors, not just technology; digital skills boost wages even in manufacturing and healthcare. And with the rise of AI, these skills will become even more important in order to stay afloat in the job market.

Digital literacy and effective tool use

Providing devices alone does not close opportunity gaps. A Digital Promise study analyzing how eighth‑grade students used digital tools during the National Assessment of Educational Progress showed that students from historically excluded groups tended to rely on assistive features like text‑to‑speech, while higher‑achieving peers used more advanced features such as digital pencils and answer‑elimination tools.

Students who used advanced tools had better math performance. The study argues that digital equity requires intentional teaching of digital literacy and professional development for educators. Community partnerships can extend digital skills instruction beyond school.

Access and infrastructure

Unequal access persists across countries and communities. In 2019 only 38.5 % of households in developing countries had a computer at home compared with 82.3 % of households in developed countries.

In India, the Tamil Nadu Free Laptop Scheme distributed over 5 million laptops to upper‑secondary students. A 2025 study using ASER and IHDS data found that access to laptops improved students’ foundational math proficiency, especially for economically disadvantaged households. The programme also increased study time, improved language comprehension and reduced private tutoring; it helped close economic and gender divides.

AI literacy and the emerging AI divide

We touched it briefly under the digital skills section: rapid advances in artificial intelligence have created a new digital divide centered on AI literacy. A global survey across 31 countries found nearly equal shares of adults feeling nervous (52 %) and excited (54 %) about AI. Marginalized communities are more likely to fear AI and less likely to understand its implications. UNESCO urges policymakers to support AI literacy programs, allocate resources to trusted local organizations, and ensure materials are inclusive and multi‑lingual.

Digital divide and skills overview

Indicator	Key data
Jobs requiring digital skills (U.S.)	92 % of jobs require digital skills; one‑third of workers have low or no digital skills.
Wage premium	Workers with digital skills earn 23 % more; those with ≥3 digital skills earn 45 % more.
Device access	In 2019, 38.5 % of households in developing countries had a computer vs. 82.3 % in developed countries.
Tamil Nadu Free Laptop Scheme	Distributed >5 million laptops and improved math proficiency, particularly for disadvantaged students.
Digital tool use in U.S. classrooms	Students from historically excluded groups rely on assistive features; higher‑achieving peers use advanced tools; effective tool use improves math performance.
Irish educators’ barriers	81 % lack training; 76 % cite insufficient resources; 54 % lack time; only 15 % have sufficient funding.

Policy and practice recommendations

Tackling STEM education challenges requires systemic reforms aligned with evidence-based policy and scalable practices. International organizations such as UNESCO, OECD, and the European Commission have emphasized the need for long-term strategies that address structural inequalities, teacher capacity, and curriculum modernization.

Key policy levers include incentivizing STEM teacher recruitment through scholarships and service-based loan forgiveness, embedding digital and green skills in national standards, and supporting data-driven monitoring of learning outcomes. Successful models demonstrate the value of coordinated, multisectoral approaches.

In short, we could identify 6 recommendations:

1. Expand early and integrated STEM education. Surveys show strong support for introducing STEM from infancy. Policymakers should incorporate inquiry‑based STEM experiences starting in early childhood and maintain integrated approaches in secondary school to prevent proficiency declines. Curricula should connect mathematics, science, technology and engineering through real‑world projects. Research on the effectiveness of STEAM programs needs to be scaled up.
2. Invest in the teacher pipeline. States should create high‑quality, affordable pathways into teaching (e.g. teacher residencies, “grow‑your‑own” program) and provide induction and mentoring supports. Competitive salaries and improved working conditions are essential to reduce attrition. Targeted scholarship and loan‑forgiveness program can attract candidates to shortage areas (math, science, special education). Data systems tracking vacancies and certification status should be strengthened.
3. Promote diversity and belonging. Outreach program must encourage girls, women, Black and Hispanic students and other under‑represented groups to pursue STEM. Mentorship networks connecting students with diverse role models can close exposure gaps. Scholarships, internships and partnerships with minority‑serving institutions and community colleges will help build inclusive pipelines. Employers should review pay and promotion practices to close gender and racial wage gaps.
4. Bridge the digital divide and build digital/AI literacy. Governments should invest in broadband infrastructure and device access, but equal emphasis must be placed on digital literacy training. Teachers need professional development to integrate advanced digital tools and AI responsibly. Community partnerships (libraries, non‑profits) can extend learning opportunities. Program like the Tamil Nadu laptop scheme show that well‑designed interventions can improve foundational skills and narrow divides.
5. Align education with labor‑market needs. Regularly update curricula to include emerging fields like artificial intelligence, data science, cybersecurity and sustainable engineering. Partnerships between schools, universities and industry can provide internships, apprenticeships and project‑based learning. AI literacy initiatives should target marginalized groups to prevent a new AI divide.
6. Improve research and data. The EU report notes the scarcity of research on integrated STEM/STEAM education and inconsistent definitions. Governments and funding agencies should support longitudinal studies assessing learning outcomes and equity impacts. Data on STEM participation by gender, race, disability and socioeconomic status should be disaggregated and publicly available.

Demand for STEM skills is surging, time to act

The demand for STEM skills is surging, yet the pathway from school to the workforce remains very uneven. Persistent gender and racial disparities, severe teacher shortages, declining proficiency, and a widening digital divide highlight areas where improvement is essential.

The good news is that proven strategies do exist: early and integrated curricula, high‑quality teacher preparation, mentoring, digital literacy instruction, equitable access to resources, and inclusive policies can transform outcomes.

---

---