Why We Grade Projects, Not Quizzes
Multiple-choice tests measure recognition. Jobs demand building. Here's why ALCI assesses students on real work — and how it works in practice.
ALCI Editorial Team2 min read
When we designed how ALCI assesses students, we started with a simple question: what would convince an experienced engineer that someone can do the job? The answer wasn't a quiz score. It was seeing what they've built, and how they talk about it.
What quizzes miss
Quizzes are useful for checking understanding along the way, and we still use short checks inside lessons. But as a final measure of skill they fall short. Choosing the right answer from four options is very different from designing a database, debugging a failing deployment or explaining a trade-off to a teammate.
How project assessment works
- Each module ends with a project brief that mirrors a real task.
- Students build the project and submit it with a short write-up of their decisions.
- A mentor reviews the work line by line, the way a senior colleague reviews a pull request.
- Students revise based on the feedback. Revision is expected, not a penalty.
- The final version is assessed against a published rubric.
What the rubric looks at
- Does it work, including edge cases?
- Is the code clear and reasonably structured?
- Is it tested where it matters?
- Can the student explain why they made their choices?
Why revision matters
In professional work, nobody's first draft is final. Feedback and iteration are how quality happens. By building revision into every project, students practise the loop they'll use for the rest of their careers — and the final grade reflects what they can actually do, not what they got right on a first attempt.
What about integrity?
Projects are harder to fake than quizzes because they're discussed. Mentors ask students to walk through their code, explain decisions and extend their work. It's difficult to explain code you didn't write — and easy to tell when someone genuinely understands it.