Teaching Economics at Scale: Small Changes, Big Impact

Background

Teaching economics to large undergraduate cohorts creates a particular set of challenges. These reflections draw on my experience teaching across first, second and third year economics modules, including large quantitative and applied modules with cohorts typically ranging from around 200 to 500 students.

Students often arrive with very different academic backgrounds, levels of confidence and expectations about university study. At the same time, staff time and resources are limited, and student engagement can become increasingly assessment-driven.

This raises a practical question: how can we provide a consistent and supportive learning experience at scale without simply increasing staff workload?

My experience of teaching large economics cohorts has led me to a relatively simple principle: at scale, quality depends increasingly on design rather than heroic individual effort. The solution cannot always be more contact hours, more emails, more office hours or more individually written feedback.

Instead, we need to think carefully about where staff and student effort is directed.

Over several years, I have experimented with relatively small changes to module organisation, assessment, tutorials, feedback and online learning environments. The innovation lies not in any single intervention, but in how these small changes work together as a coherent approach to teaching at scale: making expectations clearer, encouraging regular engagement and directing support towards students who need it most.

Diagnose early, but act on the diagnosis

One useful intervention is an early diagnostic activity, ideally near the beginning of the first semester.

This is particularly important in quantitative economics modules. Previous qualifications do not necessarily tell us everything about what students can actually do when they arrive at university. Students may have followed different curricula, may have gaps in particular areas of mathematics, or may differ substantially in confidence despite apparently similar qualifications.

Evidence from our study of more than 1,000 economics entrants across three cohorts demonstrates the importance of differences in pre-existing quantitative skills. Diagnostic test performance explains a substantial part of several gaps in students' subsequent early academic performance, reinforcing the importance of identifying differences in preparedness at the point of entry (Gall, Giannarakis and Lotti, 2026).

The important part, however, is what happens after the diagnostic.

A diagnostic that merely produces a score has limited value. Results should lead to action. Students can be directed towards particular resources or support sessions, while common weaknesses across the cohort can inform teaching priorities.

Diagnostics can therefore serve three purposes: helping students understand their own preparedness, helping staff identify where support is required, and providing an early opportunity to communicate expectations.

Create a predictable weekly rhythm

Large modules can easily become difficult for students to navigate. One response is to make the learning journey deliberately predictable, both in the weekly teaching pattern and in the way materials are organised online.

A typical week might include a short preparation activity, the main teaching session, tutorial problem-solving, a low-stakes test or checkpoint, feedback on common mistakes, and clear guidance about what students should do next. The VLE should mirror this structure, with materials, deadlines, practice activities and support resources appearing in consistent and predictable places.

I favour a simple weekly “Start here” section that highlights the main things students should complete that week. Support and stretch resources can then be clearly signposted for students who need additional explanation or greater challenge.

The intention is not to give students more work, but to reduce uncertainty about what they should be doing and when. This is particularly important in the first year, when students are also learning how to study independently and manage their time.

A well-designed VLE is therefore not simply an administrative tool. It is part of the teaching design, and a predictable weekly rhythm can help students develop more consistent study habits.

"The challenge is not simply to ask students to prepare, but to design tutorials in a way that makes prior engagement worthwhile."

Design for different starting points

In a heterogeneous cohort, a single learning pathway will not necessarily work equally well for everyone.

I find it helpful to distinguish between three types of resource.

• Core material covers what everyone needs to master.
• Support material provides additional explanation, foundational exercises and worked examples for students who need them.
• Stretch material provides more demanding questions and applications for students who are ready to go further.

This does not require creating three separate versions of a module. Relatively small additions, such as an extra worked example, an optional foundation activity or a more challenging extension problem, can give students different entry points into the same curriculum.

Scaffolding should also change over time. Early exercises might provide substantial guidance, while later activities progressively remove that structure.

The broader principle is the same: diversity should be designed into the curriculum rather than treated as an exception.

Tutorials should develop problem-solving habits

Tutorials are especially important in quantitative economics, but they can easily become sessions in which a tutor simply works through a question sheet while students copy the answers.

Students are expected to attempt the tutorial questions before attending the session. In practice, however, this does not always happen, particularly when preparation is optional or when students are managing competing assessment deadlines. This can easily turn tutorials into sessions in which students wait for the tutor to demonstrate the solutions.

The challenge is therefore not simply to ask students to prepare, but to design tutorials in a way that makes prior engagement worthwhile. Tutors can focus on a smaller number of carefully selected problems, common mistakes and alternative approaches rather than attempting to work mechanically through every question.

Students who have attempted the material beforehand can compare their reasoning with the discussion, while those who have not prepared are still encouraged to work through problems actively during the session rather than simply copying solutions.

Over time, the aim is to establish the expectation that tutorials are places for discussing attempts, difficulties and reasoning, rather than places where answers are first revealed.

Consistency also becomes increasingly important as modules grow. When hundreds of students are divided between many tutorial groups, their experience should not depend excessively on which tutor they happen to be allocated. Common resources and tutor guidance can establish shared expectations while still leaving individual tutors flexibility to respond to their groups.

"More assessment is not automatically better. The aim should be regular meaningful engagement rather than continual assessment."

Use assessment to create consistency

Assessment is one of the strongest influences on student behaviour, so rather than treating assessment-driven engagement only as a problem, it can also be used constructively.

Evidence from economics education suggests that regular assessment and repeated opportunities for practice can support student engagement and subsequent performance (Leggett, Morgan and Tang, 2022; Huskisson, O’Mahony and Lacey, 2024). I have seen similar patterns in a large first-year Mathematics for Economics module, where students completed ten online tests across the academic year. In subsequent research, increasing the contribution of these tests from 10% to 20% of the overall module mark was associated with greater test completion, stronger online test performance and higher performance in the separately assessed final examination (Giannarakis, Lotti and Psyllou, 2026).

The broader lesson is that assessment does not simply measure learning; it also helps shape when and how students engage. However, more assessment is not automatically better. The aim should be regular meaningful engagement rather than continual assessment.

Assessment design also needs to respond to generative AI. In my Applied Economics module, I incorporated a Stata assignment in which students first completed a task independently and then compared their own code, results and interpretation with an AI-generated response. They identified errors, limitations and differences in interpretation, and reflected on where AI was useful or misleading. The aim was to shift attention from simply producing an answer towards the ability to explain, verify and critique an output, while helping students develop the critical use of AI that they will increasingly need in professional settings.

Make support and feedback sustainable

Large cohorts require a more targeted approach to support and feedback. The students who need help most are not always those most likely to ask for it, while providing extensive individual feedback to every student is difficult to sustain at scale.

Diagnostics, assessment information and engagement data can help identify students who disengage early or encounter particular difficulties. Support can then be directed more effectively through targeted messages, additional sessions, clearer signposting, or strategic use of PGTAs and peer-assisted learning.

Feedback can also be made more sustainable by focusing on common patterns rather than repeating the same individual comments. Summaries of common mistakes, annotated examples, model answers, short feedback videos and reusable guidance can address issues affecting many students at once.

The aim is not to replace individual support, but to use staff time where it adds the greatest value and to intervene before difficulties become visible in the final assessment.

Student response and lessons learned

Student response has not been uniform. Activities embedded within the assessment structure have generally generated more consistent engagement, while optional preparation and support remain more difficult to sustain. Some students still arrive at tutorials without attempting the questions in advance, and optional resources are not always used by those who might benefit from them most.

There are also important trade-offs. Too many regular activities can create assessment fatigue, while too much structure can limit the development of independence. A structured VLE, online assessment and reusable feedback can reduce repetitive staff workload, but they cannot replace meaningful interaction with tutors and academics.

The main lesson has been that structure works best when expectations and incentives are clear, but it needs to be balanced with opportunities for students to take increasing responsibility for their own learning. Going forward, I would integrate preparation and support more clearly into the normal weekly structure of the module, rather than relying mainly on optional resources, while gradually encouraging students to become more independent as they progress.

Conclusion

The main lesson from teaching at scale is that improvement does not always require more contact time or more individual intervention. Small changes to structure, assessment, feedback and support can make the learning experience clearer and more consistent while also using staff time more effectively.

The challenge is to find the right balance between structure and independence, consistency and flexibility, and scalability and human contact. For me, the key principle is simple: successful teaching at scale is less about doing more and more about designing better where effort goes.

References

Gall, T., Giannarakis, P. and Lotti, E. (2026). ‘Academic performance, background and pre-existing quantitative skills of economics starters: Mind the Gap!’, International Review of Economics Education, 53, 100356. https://doi.org/10.1016/j.iree.2026.100356 .

Giannarakis, P., Lotti, E. and Psyllou, M. (2026). ‘Beyond Participation: Continuous Assessment Weighting, Engagement and Academic Performance’, SSRN preprint, Abstract 7454975. Available at: https://ssrn.com/abstract=7454975.

Huskisson, S., O’Mahony, T. and Lacey, S. (2024). ‘Improving student outcomes using automated feedback in a first-year economics class’, International Review of Economics Education, 47, 100303. https://doi.org/10.1016/j.iree.2024.100303 .

Leggett, J., Morgan, B. and Tang, K. K. (2022). ‘Repeated quizzing of basic mathematics concepts to improve grades in economics classes’, The Journal of Economic Education, 53(4), pp. 296–306. https://doi.org/10.1080/00220485.2022.2111386 .