Smaller classes help, but not the way policymakers keep hoping

Few questions in education feel as intuitive as class size. Put fewer students in a room and each gets more of the teacher; the results should follow. The intuition is popular with parents, unions and ministers alike, which is exactly why the evidence deserves a careful reading. It turns out the intuition is partly right, and the part it gets wrong is the part that costs the most money.

The cleanest test the field has ever run started in a single American state.

What did the strongest study actually find?

Tennessee’s Project STAR, begun in 1985, did something education research rarely manages: it randomised. Students and teachers were assigned by chance to small or regular classes, which stripped out the usual confounders of richer schools and abler pupils clustering together. The finding, analysed in detail by the economist Alan Krueger, was that small-class students scored higher, and that the gains were largest for the youngest children and for disadvantaged pupils.

Project STAR

Project STAR (Student-Teacher Achievement Ratio) was a large randomised experiment run in Tennessee from 1985, assigning students and teachers to small classes of 13 to 17 or regular classes of 22 to 25. Because assignment was random, its estimate of the effect of class size on achievement is treated as unusually credible causal evidence.

That is a genuine result, and it settled the question of whether small classes can help. What it did not settle is whether cutting class size is a smart use of a fixed budget, which is a different question entirely.

Why do economists keep pushing back?

The pushback, associated most closely with Eric Hanushek of Stanford, is not that small classes fail. It is that they are a costly way to buy a modest average gain. Reducing class size across a system means hiring proportionally more teachers and building more rooms, and the bill scales with the entire student population, not just the pupils who benefit most. Measured per dollar, the achievement return often trails cheaper interventions such as improving teacher quality.

The two positions are less contradictory than they sound. STAR shows small classes work; the cost critique shows they are not the best value at scale. Both can be true, and the honest policy reading is that class-size reduction is a targeted tool, most defensible for young and disadvantaged learners, rather than a universal upgrade to spray across a system.

What happens when class size meets teacher quality

The STAR experiment isolated class size, but real systems change many variables at once. A smaller class taught by an inexperienced or unsupported teacher may still underperform a larger class with a skilled teacher. That interaction matters for India, where faculty shortages often mean that expanding seats does not automatically expand teaching quality.

Research follow-ups suggest that teacher expertise explains more variance in student outcomes than class size alone. That does not make class size irrelevant; it means that class-size policy should be paired with professional development, not substituted for it. A state that hires more teachers to cut class sizes but skips induction, mentoring, and ongoing coaching may spend heavily without capturing STAR’s gains.

India’s National Initiative for Developing and Harnessing Innovations and similar faculty-development programmes attempt to address this, but their reach is patchy. Without consistent implementation, class-size reduction becomes a numbers game rather than an evidence-based improvement.

The hybrid-class dimension

The COVID-19 era introduced hybrid and online instruction to many institutions that had never used it before. That shift changed the class-size question in two ways. First, some institutions reported maintaining or increasing class sizes while moving instruction online, assuming digital tools could compensate for less individual attention. Second, others used online components to offer smaller, more interactive synchronous sessions alongside larger lectures, effectively creating a two-tier class experience.

The evidence on hybrid class size is still forming, but early signals suggest that purely large online sections tend to produce lower completion and satisfaction rates, while blended models with smaller interactive components perform better. That matters for Indian institutions considering digital expansion: simply uploading lectures to a learning management system while keeping enrollments large is unlikely to reproduce the benefits of either in-person small classes or well-designed hybrid sections.

Where does India sit in this debate?

For most of the world’s education systems the argument is about trimming classes from, say, 24 to 18. For India it is often about getting anywhere near the official norm at all. Regulators such as the University Grants Commission and AICTE set student-teacher ratio benchmarks, but chronic faculty shortages mean many institutions run well above them. The binding problem is not fine-tuning small classes; it is staffing to the standard in the first place.

That reframes the STAR debate for an Indian reader. The marginal-value critique matters where classes are already small and the choice is how to spend the next rupee. India’s more pressing task is meeting basic staffing norms in crowded institutions, where the gap between rule and reality is large enough that even the cost-conscious economists would agree there is achievement being left on the table.

Infrastructure and space constraints

Even if India could hire enough teachers tomorrow, many institutions would still struggle to halve class sizes because they lack the physical classrooms to do it. Overcrowded campuses, double-shift schedules, and repurposed labs are common where construction has not kept pace with enrollment. That creates a hidden constraint: class-size policy is not only a staffing or budgeting decision; it is a buildings and planning decision too.

Some states have tried to ease space pressure through public-private campus partnerships and shared infrastructure across colleges. Those models can work, but they require coordination across education, urban planning, and finance ministries. In practice, class-size promises often ignore the infrastructure timeline entirely, which is why they stall after the announcement.

What the evidence does not settle

It is worth being precise about what class-size research leaves open. The evidence is strongest for early grades and for disadvantaged learners; it is thinner for higher education and for professional programmes. Much of the policy debate cites STAR as if it applies to universities, which is an overreach. The Indian higher-education context is especially different because the baseline is overcrowding and understaffing, not the marginal choice between 22 and 18 students.

That distinction matters for implementation. Policymakers who want to apply STAR-style findings to higher education need evidence specific to that level. Until they have it, the safest use of research dollars is to improve teacher quality, curriculum design, and student support alongside any class-size adjustments.

Conclusion

Class size is one of the few education questions with a clean experimental answer, and the answer is nuanced rather than convenient. Small classes lift early achievement, most for the children who need it most, which is why STAR endures. But reducing class size across a whole system is expensive enough that it rarely wins a value-for-money contest, which is why the economists endure too. For India, the debate lands differently: the target is not smaller classes but adequately staffed ones, and on that measure the evidence and the arithmetic point the same way.