Optional vs core modules — what the data actually says

UK universities pitch optional modules as a chance to specialise — to develop your interests, to differentiate your degree. That's true, but it's a strangely bloodless way to describe what's actually happening on the page when you tick the boxes.

Optional modules are where the variance lives.

In the data we've collected — multi-year FOI disclosures across UK Russell Group institutions — the spread of grades within a year is dramatically higher across optional modules than across core modules. The same student, taking different sets of optionals, can graduate with materially different classifications. We're not talking about marginal differences; we're talking about the difference between a high 2:1 and a first.

Why core modules cluster

Core modules — the ones every student on a degree has to take — tend to have compressed distributions. There are a few reasons:

1. The cohort isn't self-selecting. Everyone takes it, including people who aren't great at it.
2. Marking is heavily moderated. Cores get more eyes, more second-marking, more external scrutiny.
3. Assessment formats are conservative. Core modules tend to use established formats — exams, structured essays — which produce predictable distributions.
4. The syllabus is stable. Once a core module is established, departments don't change it lightly.

The result is what we see in the FOI archive: core module distributions tend to bunch tightly around the cohort mean, with relatively few firsts and relatively few fails.

Why optional modules spread

Optional modules go the other way. The cohort is self-selecting, often small, and motivated. Assessment formats are more varied — projects, dissertations, presentations, group work. Marking is sometimes done by a single specialist with strong views. Year-on-year variation is higher.

The result: distributions are wider. More firsts. More middling 2:1s. More 2:2s if the module didn't suit you. The ceiling and floor are both lower — there's more dispersion.

The "self-selection" effect

There's a subtler thing going on with optional modules: the cohort that turns up is usually self-selecting around interest or perceived suitability. People who pick "Advanced Cryptography" tend to be the people who think they'll do well at it. The mean of that cohort is therefore higher than it would be if everyone had to take it.

But — and this is the bit that surprises people — the distribution within the optional cohort isn't necessarily compressed. There's still a tail of people who picked it because the module name sounded cool and got rinsed.

This is why the average mark on an optional module isn't a great proxy for "how easy is this module to do well in". The cohort is biased, but unevenly. Looking at the upper-quartile mark and the first-rate gives you much more useful signal.

What this means for how you pick

Three operational implications:

1. Spend your optional budget on modules with a high ceiling. Look at top-quartile marks, not just means. A module where the top end stretches into the high 70s is more likely to give you a first than a module with a higher mean but a flat ceiling.

2. Don't mistake a high mean for an easy module. A high mean often reflects a high-effort, motivated cohort. If you're not in that cohort, the mean doesn't apply to you.

3. Stable beats trendy. A module that has been running for ten years with consistent distributions is much more predictable than a brand-new module with two years of data. Predictable matters when you're trying to plan a target classification.

What this looks like in the FOI archive

In our Exeter Computer Science dataset (2014–2023), core modules sit in a fairly narrow band: most have means between 60 and 65, with first-rates between 18% and 28%. Pretty consistent, year on year.

The optional modules look completely different. We have one module with a mean of 71 and a first-rate of 41% across a decade. We have another that nominally covers similar material with a mean of 58 and a first-rate of 11%. Same university, same department, same year level — and an order-of-magnitude difference in outcome.

If you graduated in the cohort that took the second module instead of the first, you didn't have a worse degree experience. You almost certainly had a different one. But on paper, you came out behind.

The framework, in one paragraph

When you're picking optional modules, you're allocating risk and reward. Optional modules are where the variance lives. The same effort applied to different optionals produces different outcomes — and the historical FOI data tells you which is which. Spend that budget thoughtfully.

This is the bit GradeHack is being built to handle: enter your degree, your goals, and your constraints, and the system surfaces the optional combination historically associated with your target. We're scaling up the dataset right now. Join the waitlist and you'll be one of the first to use it when it lands.