Until the mid-2010s, most GCSEs and A-Levels in England were modular: students sat and banked unit exams progressively across the course, often resitting individual units to improve a mark. Reforms rolled out from 2015 to 2017 replaced that structure with linear, terminal exams — a single set of papers at the end of the course covering everything. Other systems, including the International Baccalaureate and parts of Germany's Abitur, still combine ongoing assessed work with a final exam, and the evidence on which approach is better for learning or wellbeing doesn't point cleanly one way.
If your own GCSEs or A-Levels were sat before roughly 2017, the exam your child now sits is genuinely structured differently — not just re-branded with new grade numbers, but built on a different theory of when and how a subject should be tested.
What "modular" used to mean
Under the modular system that preceded the current reforms, an A-Level or GCSE course was broken into discrete units, each separately taught, examined and graded — often across AS-level exams at the end of Year 12 and A2 exams at the end of Year 13, with individual units resat if a student wanted to improve a specific mark before the final grade was calculated. The final qualification grade was effectively a running total of unit results banked over the two years, rather than a single measurement taken at the end.
What changed, and why
The reforms rolled out under Ofqual's GCSE, AS and A-Level reform timetable moved almost all GCSEs and A-Levels in England to a linear structure: teaching still runs across the full course, but assessment happens in one block of exams at the end, covering the whole specification. AS-Levels were decoupled from A-Levels so they no longer automatically count towards the final A-Level grade in most subjects, and the practice of resitting individual units mid-course was largely removed, because there are no longer separate unit exams to resit — only the whole subject, retaken as a full re-entry.
The Department for Education's stated rationale, carried through in Ofqual's implementation, was twofold: to reduce a resit culture built around repeatedly re-sitting small units for marginal gains, and to encourage teaching that builds connected understanding across an entire subject rather than treating it as a sequence of separately examinable chunks. Whether that rationale has been borne out in practice is a matter of ongoing educational debate rather than a settled empirical fact, and this article doesn't claim otherwise.
Where modular and continuous assessment still exist
The change was a UK-specific policy decision, not a universal shift — several major systems elsewhere in this cluster kept a more modular or continuously-assessed shape.
The International Baccalaureate Diploma Programme explicitly combines internal assessment — coursework, projects and orals marked by the student's own teachers and moderated externally — produced across the two-year course, with final external exams at the end. The final grade in most IB subjects is a genuine blend of both, not a pure terminal-exam result.
In Germany, the Abitur is organised by each Land (state) rather than nationally, and several states build a substantial ongoing component into the final result: assessed work and grades accumulated during the final two years of school (the Kursstufe) sit alongside the terminal written and oral exams, rather than the exam alone deciding the outcome. This varies by state, which is itself a structural point worth knowing — Germany doesn't have one Abitur system, it has sixteen, coordinated but not identical.
A worked example: what actually changed for A-Level Maths
It's easier to see what "modular" meant in practice with a concrete example. Under the old system, A-Level Maths was commonly built from six discrete units — typically three "Core" pure maths modules and three applied modules (drawn from mechanics, statistics or decision maths) — each sat as a separate exam, often at different points across Year 12 and Year 13, and each separately graded. A student unhappy with one applied module's result could resit that single unit without retaking the rest of the qualification. Under the linear reform, A-Level Maths became a single specification assessed by exams sat at the end of Year 13 covering pure maths, mechanics and statistics together — there is no longer a "Core 3" or "Statistics 1" exam a student can bank early and never revisit; the whole subject is examined as one, in one exam season, and a resit means resitting the entire A-Level rather than one component of it.
Another contrast: the US model of continuous assessment
A different kind of alternative to England's terminal-exam model is the way US high schools weight the grade point average (GPA) — a running average built from grades earned in coursework, tests and assignments across all four years of high school — as the primary record most colleges see, alongside standardised tests such as the SAT (College Board) or ACT sat at one or more points during those years. This is a more extreme version of "continuous" than either the IB or the German Abitur: there is no single terminal exam covering an entire subject's content at all in most US states, and a poor performance early in high school can meaningfully affect a student's overall academic record in a way no single UK GCSE or A-Level resit fully mirrors, because there's no equivalent "whole subject, one sitting, clean slate" reset available. The trade-off runs the other way from England's: continuous, cumulative pressure across four years, rather than concentrated pressure at one end-of-course sitting — a genuinely different distribution of stakes, not simply a longer version of the IB's blended model.
Comparison: how different systems distribute assessment
| System | Model | Where the pressure sits |
|---|---|---|
| England (post-reform) | Linear — one terminal exam set per subject | Concentrated at the end of the two-year course |
| England (pre-2015) | Modular — separate unit exams, resittable individually | Spread across the course, with repeat opportunities per unit |
| IB Diploma | Blended — internal assessment plus final exams | Spread across the course, with a final exam component |
| Germany (Abitur) | Blended, varies by Land — Kursstufe grades plus terminal exams | Spread across the final two years, plus an exam sitting |
| United States (typical) | Continuous — GPA built from four years of coursework, plus standardised tests | Spread across the entire high-school career, cumulatively |
What the evidence actually says on stress and learning
This is one area where the honest answer is "it's genuinely contested" rather than "here's the finding". A single terminal exam concentrates the entire stakes of two years of study into one sitting (or a short exam season), which many students, parents and student-wellbeing organisations describe as intensely stressful precisely because there's no earlier result to fall back on and no way to recover from a bad day except a full resit. That's a real and widely reported concern about linear exams, even without a single settled statistic to attach to it.
Modular and continuous-assessment systems trade that concentrated stress for something more diffuse: pressure spread across the whole course, rather than compressed into one result. Whether that's genuinely less stressful overall, or simply the same stress spread more thinly (with the added burden of assessed coursework recurring throughout the two years), isn't something the research base settles cleanly — reasonable people, and reasonable education systems, have reached different conclusions, which is exactly why both models still coexist internationally rather than one having displaced the other.
The takeaway
England chose a linear, terminal-exam model in the 2010s specifically to move away from modular unit-resitting, while the IB and several German states kept a more continuous, blended approach for the opposite reason — to spread stakes and assessment across the whole course rather than concentrate them at the end. Both are defensible, considered designs rather than one representing an outdated approach the other has superseded; the honest comparison is about what each trades away, not which one is "right".