Architecture is one of the few disciplines where a faculty routinely examines two structurally different kinds of thesis side by side — a design thesis built around a studio project and portfolio, and a written dissertation built around a research argument. Treating both under one written-dissertation rubric produces the two failure modes every architecture programme director has seen: a design thesis marked down for thin prose when the drawings carry the argument, or a written dissertation waved through on production values it does not actually earn academically. Here is how to set up an assessment framework that scores each track on its own terms and still produces a defensible, comparable grade.
Step 1 — Decide, in writing, what a design thesis actually is at your institution
Before any studio brief goes out, the programme needs a written definition distinguishing a design thesis (a resolved architectural proposal — drawings, models, technical documentation, plus a shorter written component, sometimes called an exegesis or thesis statement) from a research/written dissertation (an argument built primarily in prose, with design work, if any, illustrative rather than the primary submission). Many programmes also run a hybrid track: a design proposal with a substantial theoretical or technical research component weighted close to 50/50. Naming the track explicitly in the programme handbook — not leaving it to informal studio convention — is what lets the assessment criteria below actually be enforced rather than negotiated case by case at the crit.
Step 2 — Build the rubric before the brief, not after the crit
A design-thesis rubric that holds up under moderation typically separates four components, each weighted and each with its own pass/fail threshold rather than one blended score: design rigour and conceptual resolution (does the proposal answer its own stated brief coherently); technical resolution (structure, environmental performance, construction logic — resolved enough to be buildable in principle, not decorative); representation and craft (drawing, model-making, digital output quality — real, but capped so it cannot substitute for the first two); and the written component (thesis statement or exegesis, typically 3,000–8,000 words depending on the programme, explaining the intent, precedents and research underpinning the design). Publishing the weighting split — for example design rigour 35%, technical resolution 25%, representation 15%, written component 25% — before the studio brief is issued is what prevents the most common appeal ground in design-thesis assessment: a student arguing the jury scored on criteria nobody told them mattered.
Step 3 — Assemble a jury, not a single marker
Architecture design theses are conventionally assessed by a jury or panel rather than a single supervisor, typically the studio tutor, one internal second marker from outside that studio, and an external examiner or invited practitioner. This is a structural safeguard, not a formality, and it is also a workload-allocation decision worth planning the same way as supervision capacity is allocated in a business school: a single marker scoring a design thesis is scoring against their own studio’s house style whether they intend to or not, and a jury with an outside voice is what catches that. For programmes seeking or holding accreditation — RIBA validation in the UK, NAAB accreditation in the US, or alignment with the EAAE’s quality frameworks in continental Europe — the external examiner or practitioner juror is often an explicit requirement, not an optional addition, and their sign-off needs to be a recorded, auditable part of the assessment file rather than an informal nod at the end of the crit.
Step 4 — Structure the crit as a defence, not a critique session
A studio crit and a thesis defence are not the same event, and a design thesis needs the second, not just the first. The distinction that matters for assessment purposes: a crit is formative, ongoing feedback during the design process; a defence is a single, scored, minuted event where the student presents the resolved proposal, the jury asks questions specifically tied to the published rubric criteria, and a decision is recorded. A defence that drifts back into open-ended crit-style critique produces exactly the kind of unrecorded, unstructured feedback that cannot support a grade if a student later appeals it. A simple fix that holds up under scrutiny: give each juror a rubric-aligned score sheet to complete during the defence itself, not from memory afterward.
Step 5 — Moderate across studios before grades are finalised
Where a faculty runs multiple studios in parallel, each with its own tutor and its own jury, calibration across studios is the single most common weak point in architecture-thesis assessment — two juries applying the same published rubric can still land on visibly different grade distributions if they never compare notes. A moderation meeting after all defences but before grades are finalised, where jury chairs bring a sample of high, middle and borderline-pass portfolios for cross-studio comparison against the rubric, is standard practice at institutions that have been through an accreditation review and is worth adopting even where no external accreditor requires it. This is the architecture-specific version of the calibration problem covered more generally in how to write a departmental writing standard for theses — the mechanism is the same, but a design-thesis rubric has more genuinely subjective criteria (conceptual resolution, craft) than a written dissertation, which makes cross-studio moderation more necessary here, not less.
Step 6 — Align the written component with your citation and integrity policy
The written exegesis or thesis statement is a full academic text and needs to sit inside the same integrity and citation policy as any other written dissertation on campus — precedent-study images and case-study material need proper attribution just as literature citations do, and a design thesis is not exempt from the faculty’s plagiarism-and-AI-use policy just because most of its content is drawings. Where the programme allows AI-assisted drafting for the written component under the same rules that apply to other faculties, that permission needs to be stated explicitly in the design-thesis handbook rather than assumed to carry over silently from the general university policy, since design-thesis students and their tutors do not always read the general academic-integrity policy as applying to them. What a similarity or AI-detection tool actually flags on a mostly visual submission with a shorter written component is also worth confirming in advance — see what academic writing tools actually check for the level-by-level comparison.
The hybrid track: when design and research are weighted close to even
A growing number of programmes, particularly at master’s and professional-doctorate level, run a third track between the two: a hybrid thesis where a design proposal and a substantial research or technical investigation are weighted close to 50/50 rather than one clearly leading the other. This track needs its own explicit rubric rather than being treated as a design thesis with extra pages or a written dissertation with extra drawings — the two most common assessment failures on hybrid submissions are a jury defaulting to design-thesis criteria and skimming the research chapter, or defaulting to written-dissertation criteria and treating the design proposal as an illustration rather than a scored deliverable in its own right. Naming the hybrid track explicitly in the handbook, with its own weighting table separate from the two in the section above, closes this gap before it reaches a jury.
Common rejection and revision grounds, and how a published rubric prevents most of them
Four grounds account for most design-thesis revision or fail decisions that later face an appeal, and a published, pre-brief rubric addresses each directly: the proposal does not resolve technically at the level the brief demanded (addressed by an explicit, weighted technical-resolution criterion rather than folding it into a vague “quality” score); the written component does not actually explain the design intent, reading as an afterthought rather than integral to the submission (addressed by setting a minimum word count and a specific content requirement — precedents, research basis, intent — rather than leaving the exegesis open-ended); the jury’s stated reasons for the grade do not match what is on the rubric score sheet (addressed by requiring jurors to complete the rubric-aligned score sheet during the defence itself, not reconstruct it afterward); and cross-studio grade drift, where an identical portfolio would have scored differently in a different studio (addressed by the moderation meeting in Step 5). A programme that can point to all four safeguards in its assessment file is in a materially stronger position if a grade is formally appealed or externally reviewed.
A template weighting table you can adapt
| Component | Illustrative weighting | What the jury is scoring |
|---|---|---|
| Design rigour and conceptual resolution | 30–40% | Does the proposal coherently answer its own stated brief? |
| Technical resolution | 20–30% | Structure, environmental performance, construction logic — resolved in principle |
| Representation and craft | 10–20% | Drawing, model-making, digital output quality — capped, not decisive alone |
| Written component (exegesis/thesis statement) | 20–30% | Intent, precedents, research underpinning the design, properly cited |
These weightings are an illustrative starting point for a programme committee to adapt against its own accreditation requirements, not a fixed external standard — RIBA, NAAB and national programme requirements vary, and the split needs sign-off from whoever owns your accreditation relationship before it is published to students.
Where Tesify fits
The written exegesis is where most design-thesis academic-integrity questions actually surface, not the drawings. The Tesify Integrity Suite for Institutions screens the written component with sources named and a batch view across a studio cohort, so a jury chair moderating across studios can see integrity flags alongside the rubric scores rather than chasing them separately after the fact. A free departmental pilot lets one studio test this against a single design-thesis cycle before any procurement decision.
Frequently asked questions
Should a design thesis and a written dissertation use the same rubric?
No. A shared rubric under-scores design work that argues primarily through drawings and models, and over-credits production values in a written dissertation that should be judged on argument. Publish a separate, explicit rubric for each track.
Who should sit on a design-thesis jury?
Conventionally the studio tutor, an internal second marker from outside that studio, and an external examiner or invited practitioner. The outside voice is a structural safeguard against a single studio’s house style being scored as if it were the only valid approach.
Does RIBA or NAAB require an external examiner on the jury?
Requirements vary by validation or accreditation body and by programme level — confirm the current criteria with whichever body your programme is validated or accredited under before finalising the jury composition, since this is exactly the kind of requirement that changes between review cycles.
How long should the written component of a design thesis be?
Programmes vary widely, typically in the 3,000–8,000 word range, well short of a full written dissertation, because the design work itself carries a substantial share of the argument. Set and publish your own programme’s expectation rather than importing another institution’s figure.
What is the difference between a crit and a thesis defence?
A crit is formative, ongoing feedback during the design process. A defence is a single, scored, minuted event where the jury scores against the published rubric and the decision is recorded — the crit format alone cannot support a grade if a student later appeals it.
Does the written exegesis need to go through the same plagiarism and AI-use policy as other theses?
Yes. State this explicitly in the design-thesis handbook — design-thesis students and studio tutors do not always assume the general academic-integrity policy applies to a mostly visual submission unless it is spelled out.
Why is cross-studio moderation more important in architecture than in a single-supervisor dissertation model?
Because more of a design-thesis rubric is genuinely subjective — conceptual resolution and craft — than a written dissertation’s criteria, two studios applying the same published rubric can still diverge without a moderation meeting comparing sample portfolios before grades are finalised.