1. Sentence 1 — the direct answer. Answer the question in its own words before you explain anything.
2. Sentence 2 — the definition, in the course's terms.
3. Sentence 3 — the application: "In this case that means…"
4. Sentence 4 — an example, one clause of context and then the point.

Worked shape. "A confidence rating is a self-reported measure of how sure a participant was that they completed the task correctly. It is quantitative but subjective, which is why the course pairs it with success rate rather than using it alone. In this case the 4.8/5 rating alongside an 86.7% success rate says most people succeeded and believed they had. In my own testing a high rating with a low success rate was the signal that the task wording was misleading them."

1. Definition — one or two sentences, course wording.
2. Purpose — what it is for, and when in the process it happens.
3. The named components — 3–5 of them, each with a half-sentence gloss. Use the course's own list; if the course numbers them, keep the numbers.
4. One example, with what it showed.

The trap at 8 marks is spending six of them on the definition. Definition and purpose together should be about a quarter of the answer.

1. Block 1 — Define and situate. What it is, what it is for, where it sits in the process.
2. Block 2 — Apply. Take the case's own evidence — a quote, a metric, a described failure — and run the concept over it. Name the evidence you are using.
3. Block 3 — Example and consequence. Your own example, then the "so what": what changed, or what would have changed.

If the command verb is compare, split block 2 into two halves with a stated basis of comparison, and give block 3 to the difference that matters.

1. Define and situate.
2. Apply to the case with named evidence.
3. The alternative you did not take, and the specific reason — cost, time, what it cannot show, who it excludes. This is the block that separates 12 from 10.
4. Example and consequence.

Any of the course's paired frameworks gives you block 3 for free: qualitative vs quantitative, moderated vs unmoderated, concept vs usability testing, Impact × Importance vs Impact × Effort, low vs mid fidelity, current vs future state.

1. Define and situate — brief. Two or three sentences.
2. Name the framework in full. All five heuristics, all five principles, all five Ps, all five blueprint components, the three sensemaking steps — whichever applies. Naming the whole framework costs little space and gives the rest of the answer something to work through.
3. Apply it point by point to the case. Not every point needs a paragraph; the ones the case actually breaches do.
4. Example with a consequence. Yours, with what it showed.
5. Recommendation + verification. The specific change, and how you would know it worked — a measure, a test, a hypothesis. Close the loop.

A drawn artefact may appear in one of Question 3's sub-parts, but the transcript does not identify which mark values those sub-parts carry — so do not assume it will be a 15. Where drawing is required, use the artefact as the main application block and put the reasoning into its labels and annotations.

1(a) — 10 marks, three blocks.

• Define + situate. Exploratory research is what you do before you know what the problem is — it sits in Discover on the Double Diamond and its output is understanding, not a decision. Contrast it in one clause with evaluative research, which tests something that already exists.
• Two activities, two artefacts — the question specifies a count, so label them so they are countable at a glance. Activities: semi-structured interviews; a workshop (the 4Cs Collect stage). Artefacts: a discussion guide and transcripts / participant snapshots from the interviews; clustered outputs from the workshop. Say for each what it lets you do next.
• Example + so what. Your own project: what you set out to learn, which activity produced it, and one thing you learned that you had not expected — the surprise is the evidence that it was genuinely exploratory.

1(b) — 15 marks, five blocks. This is the full-length shape from Section D.

• Situate. Criteria come after the research question and before recruitment. Say what the research question is for enrolment — e.g. "why do continuing students leave enrolment until the final week?"
• The framework in full: Who · What · How · Where · When (W2 s67), plus exclusions and limitations.
• Apply each one. Who: continuing undergraduates who have enrolled at least once, plus a second segment of first-years who have not — the behavioural qualifier, not just "students". What: numbers and quota across the two segments. How: the channel, and what it biases toward. Where: on campus vs remote. When: during the enrolment window, not after it, or they will be reconstructing from memory.
• Sample size, attributed. For a usability test of the enrolment interface, Magic of 5 — 5 users ≈ 85% of problems — attributed on W7 s35 to Jeff Sauro of MeasuringU. For exploratory interviews, say instead that you recruit until you stop hearing new themes, and that the course gives no fixed number for this.
• Two limitations, stated plainly. E.g. recruiting through a student portal reaches only students who already log in, so the people most likely to be locked out are the ones you will miss; and a single-campus sample cannot speak for distance students. Say what each limitation would stop you claiming.
• Example. The workshop where criteria were written after recruitment — and what that cost. A failure with a lesson answers this better than a clean success.

2(a) — 8 marks. A list question with an attribution requirement. Give all five, grouped by their author bands exactly as W1 s35 prints them:

• Alan Cooper — "Software should be polite": 01 Software should be forthcoming, 02 Software should be self-confident
• Jakob Nielsen — Usability Heuristics: 03 Visibility of system status, 04 Match between system & real world
• Steve Krug: 05 Don't waste my time!

Add one sentence that the set is titled "Combining 5 heuristics for INFS3700" and is therefore this course's own combination, not any single author's list. That sentence is what turns a recall answer into a correct one.

2(b) — 12 marks. Two breaches × the five-part finding, then the comparison block.

• Breach 1 — 03 Visibility of system status. Evidence: a spinner with no text, and no state change for up to ten minutes. User consequence: the user cannot tell whether the order was placed, so they re-order or abandon. Business consequence: duplicate orders, refunds, support contacts. Recommendation: replace the spinner with a named state and a time estimate — "Order received · restaurant confirming · usually under 2 minutes" — and push the confirmation in-app, not only by email.
• Breach 2 — 01 Software should be forthcoming, or 05 Don't waste my time. Forthcoming: the system knows the order status and does not volunteer it. Don't waste my time: the user has to leave the app and check email to learn something the app already knows. Either is defensible — pick one, and say why that one rather than the other.
• The 12-mark extra block: which of the two you would fix first, and on what basis. Impact on the user, not effort for the team — and say how you would know it worked (a drop in duplicate orders, or a direct-success measure on "confirm my order was placed").

3(a) — 10 marks. Name all five, apply each in one sentence, then commit.

• People: member, branch staff, the returns/shelving team. Processes: reserve → check-in → shelve → locate → hold → collect. Places: the app and the branch collection point. Products: the app, the hold shelf, the book itself. Performance: the 72% collection rate and the 2.9/5 confidence rating.
• Commit: the failure is in Processes — specifically the handover between check-in and shelving. It presents as a Products failure (the app lies) but the app is faithfully reporting the only event the process actually records. Saying that, and saying why it is not a Products failure, is the analysis.

3(b) — 5 marks. One paragraph. State the HMW, then say what makes it good.

• "How might we give members an accurate picture of whether a reserved book can actually be collected today?"
• Good because it names the outcome (an accurate picture, today) and not the solution. It leaves open whether the fix is a new status, a delayed flag, a staff scan or a changed promise. Contrast with a poor version — "How might we add a shelving-status barcode scan to the app?" — which has already chosen the answer, and cite W3 s61 as the slide that draws this distinction.

3(c) — 10 marks. Three blocks: what each says, what neither says, what both say together.

• Quantitative alone: 72% collected, confidence 2.9/5. Those two disagree — a 72% success rate with a 2.9 confidence means members succeed more often than they expect to. That gap is the finding.
• Qualitative alone: the member quote gives the mechanism (a false "available"), the staff quote gives the cause (check-in ≠ located). Neither tells you how widespread it is.
• Together: the failure is not frequent enough to be an outage but is unpredictable, and unpredictability is what destroys confidence. Diagnose, then decide — this is an iterate, not a validate. Use the Week 10 slide-30 vocabulary (low confidence rating due to…, customers found X confusing and therefore…) and note that the course excludes calculations, so no arithmetic is expected here.

3(d) — 15 marks. Draw it. Stages across the top: reserve · notified · travel · collect. Lanes down: physical evidence, customer actions, frontstage, backstage, support processes. Lines labelled.

• Physical evidence: app screen, "available" notification, hold shelf, the book.
• Customer: reserves → receives notification → travels → asks at desk.
• Frontstage: desk staff search the hold shelf, then the trolley.
• Backstage: book checked in; book waits on returns trolley; book shelved.
• Support: the catalogue system that flips the flag on check-in.
• The annotation that completes the answer: say where the failure starts, not just where it becomes visible. The failure begins in the support process, when check-in is treated as equivalent to confirmed shelf availability. That premature status then propagates outward — across the line of internal interaction into backstage, where the book is still physically on the returns trolley and contradicts the record; through the backstage-to-frontstage handoff, where desk staff search a hold shelf the system says is stocked; and across the line of visibility into the customer-visible app notification, which is where the member finally encounters it.
• Be precise about what the line of visibility does. It separates activity the customer can see from activity they cannot. It marks where the bad status becomes visible; it is not what caused it. An answer that says "the failure is at the line of visibility" has described the symptom. Mark both points on the diagram — origin and surfacing — and label which is which. Apply the two rules the tutor corrected in class: one touchpoint per service moment, and name who performs every staff action.

3(e) — 15 marks. One change, a hypothesis, a test, a measure.

• Change: split the status — the flag flips to "Returned — being processed" at check-in and to "Ready to collect" only when the item is scanned onto the hold shelf. Notification fires on the second event, not the first.
• Hypothesis: We believe that members lose trust because "available" is set before the book is collectable. So if we notify only on shelf-scan and show a distinct in-process state, we will see confidence rise and wasted trips fall.
• Test: concept-test the two-state notification with members using a low-fidelity wireframe (a storyboard would do the same job if you want to show the wasted journey), then a moderated usability test on the task "find out whether you can pick up your book today". Participants: members who have reserved in the last month — the behavioural qualifier from topic 01. Five users, attributed to Sauro's Magic of 5.
• How you would know: confidence rating moves off 2.9; direct success on the "can I collect today" task; and the operational measure — collections attempted before the item was shelved. Say which of these you would consider decisive, and say what result would prove you wrong.

1(a) — 10 marks. The verb is compare, so state the basis first.

• Basis: what kind of question each can answer. Qualitative answers why and how; quantitative answers how many, how often, how much. Say this in one sentence before anything else — a comparison with no stated basis is the classic way to miss what this verb is asking for.
• Where each sits: the course maps qualitative into Discover and Develop, quantitative into Define and Deliver (W2).
• Examples with their outputs. Qualitative: semi-structured interviews → discussion guide, transcripts, participant snapshots; contextual observation → observed vs inferred behaviour. Quantitative: an online survey → distributions and segment sizes; usability metrics → direct/indirect success, ratings and scales, time to complete.
• How you would collect both — the question asks this explicitly, so answer it explicitly: recruit against criteria, run the survey first to size the problem, then interview a subset to explain the pattern. Name this as triangulation and say what it buys you.
• Close with the verdict a comparison needs: neither is sufficient; the pairing is what makes a finding defensible.

1(b) — 15 marks. Full-length shape.

• Situate. This is Define — between fieldwork and any design work.
• The framework in full. The three sensemaking steps, then: observations → affinity mapping → themes → insights → problem statement → How Might We → prioritisation.
• Define insight properly. W3 s16: insights are groupings of observations that bubble up into a clear theme that is IRA — Interesting, Relevant, Actionable. Give all three tests, and give the s17 gloss on Interesting: how did the user's perspective change yours?
• The worked chain — one line each, using a real example. Observation: "I checked the group chat four times before the meeting to see if anyone had started." Theme: effort spent checking, not doing. Insight: members cannot see progress without asking, so they substitute checking for working — interesting (it reframes the problem from motivation to visibility), relevant, actionable. Problem statement: team members have no shared view of progress and fall back on interruption. HMW: "How might we let a team member see what has moved without asking anyone?"
• Prioritisation + naming rule. Name one matrix and define its axes — Impact × Importance (W3 s64, customer benefit × business benefit, 1–5) or Impact × Effort (W3 s65, Start Here / Do Next / Proceed Carefully / Avoid). And state the affinity naming rule: clusters are named after the experience, not the proposed fix.

2(a) — 8 marks. Definition, purpose, components, contrast.

• Definition, quoting W4 s43: storyboards convey social or functional messages through entertainment and engagement in a visual format — showing, in a linear and sequential fashion, how a customer may interact with your experience.
• Purpose: it answers "why would anyone want this?" A storyboard carries a person, a situation and an emotion.
• Components: the three acts, and two or three of the eight elements (characters, speech bubbles, devices, signs, arrows…).
• The contrast the question asks for: a wireframe shows one screen's structure — layout, hierarchy, functionality, interaction options. It cannot show context outside the screen, emotion, or the trigger that made someone open the app at all. Name what each artefact does not show; that is the core comparison the question requires.

2(b) — 12 marks. Draw six frames, three acts, captioned, emotion marked.

• Frame 1 — actor, context, trigger. A named resident walking home in the dark past a dead streetlight. Emotion: uneasy.
• Frame 2 — the pain point. They have reported it before by phone, waited on hold, and never heard back. Emotion: resigned. This frame is what makes it a storyboard rather than a wireflow.
• Frame 3 — the user action. Phone out, photo taken on the spot.
• Frame 4 — the intervention / touchpoint. The app captures the location automatically and shows a reference number. Emotion: relief.
• Frame 5 — the outcome. A notification: "crew scheduled Thursday". Emotion: trust.
• Frame 6 — what it means beyond this. They report the next fault without hesitating — and tell a neighbour.
• Annotations: against frame 2, note that the pain is the silence, not the pothole; against frame 4, note that automatic location removes the step most likely to make someone abandon; against frame 5, that the outcome is a closed loop, which is what turns one report into a reporting habit. Apply the three principles by name — Stay Authentic · Keep it Simple · Highlight Emotion.

3(a) — 10 marks. Name all five, then commit to two.

• The two: 4 Provide feedback — "It didn't say anything, so I signed up again" is a direct quote of an absent confirmation, and it is producing the duplicate records the administrator reports; and 1 Perceivable and predictable — three visually identical lists mean the volunteer cannot tell which one is this week without opening all three.
• Justify the choice. These two are the ones with evidence attached in the case and a measurable consequence: duplicates for the first, 3m 40s and a 54% direct success for the second. Note briefly that 2 Consistent and conventional is also arguably breached, and say why you ranked it below the other two — that ranking is the evaluation.
• Do not write about signifiers or F-patterns. Neither is course vocabulary.

3(b) — 5 marks. One paragraph, the recipe verbatim.

• We believe that volunteers abandon sign-up because they cannot tell which roster is current and get no confirmation when they commit. So if we show a single dated "this week" roster as the default view and confirm the sign-up inline with the shift name and date, we will see direct success rise, time on task fall, and duplicate sign-ups drop.
• Add one sentence on why it is a good hypothesis: it is specific, it names the variable being changed, and it predicts a measurable outcome — which means it can be shown to be wrong.

3(c) — 10 marks.

• The gap is the finding. 54% direct + 21% indirect = 75% eventually succeeded. A high indirect proportion means the task is possible but not discoverable — people find it by backtracking. If the failure were structural, indirect success would be near zero too. The 3m 40s corroborates it: that is the cost of the backtracking, on a task that took seconds on paper.
• Pair with the qualitative. The two quotes name the mechanism: three identical lists (discovery) and no confirmation (feedback). Diagnosis, then decision — this is an iterate.
• What the figures cannot tell you. Who stopped volunteering entirely — they are not in the numbers, because they never started the task. Also nothing about the older volunteers who left the app for SMS. The metrics measure people who tried; the drop in volunteer numbers is about people who did not. Say what you would do to reach them, and note that the course's vocabulary for these measures is Week 10 slide 30's (direct/indirect success, confidence rating), with Week 7's success rate as the equivalent.

3(d) — 15 marks. Two screens, low fidelity, labelled, with an arrow from the tapped element.

• Screen 1 "This Saturday — 12 April": a dated title bar so the current roster is unmistakable; a single list of shifts with role, time and slots remaining; each row showing its own state (open / full / you're on this); one primary action per row. Placeholders only — images as X, text as lines.
• Screen 2 "You're on — Canteen, Sat 12 April, 9–11am": reached by tapping the row's action. Confirmation stated in words, with an undo, and the roster row now showing your name.
• Three annotations, each with its evidence: (1) the date in the title — because "I don't know which one is this week until I've opened all three"; (2) the per-row state including "you're on this" — because "I signed up. I think." and because duplicates are appearing in the administrator's data; (3) one primary action per row, visually dominant — because indirect success at 21% shows people are hunting for the commit step.
• Close by naming what the wireframe does not settle — whether older volunteers can read it, and whether "you're on this" is legible at a glance — and say that is what 3(e) tests.

3(e) — 15 marks. Follow the course's four steps: Plan, Prepare, Moderate, Outcomes.

• Participants and criteria. Current and lapsed volunteers who have signed up for a shift in the last season — the behavioural qualifier. Include the older-volunteer segment deliberately, since they are the group the case says gave up. Five to six users for the usability test, attributed to Sauro's Magic of 5 (W7 s35) — 5 users ≈ 85% of problems.
• Tasks, in the user's language, not the interface's. "You've been asked to help at the canteen this Saturday morning — do what you'd normally do to put your name down." Then: "Check whether you're actually on the roster." Never say "tap the sign-up button".
• Measure: direct and indirect success on each task; time on task; a confidence rating after task 2; and the qualitative three — what they liked, what they did not, what they would suggest. Add the operational check: duplicate sign-ups.
• Two biases from the course's own list, both of which genuinely fit this scenario. (i) Social desirability bias — you are testing a club's app with the club's own volunteers, who have every reason to be polite about it; ask what they did, not whether they liked it, and take the behaviour over the compliment. (ii) Hawthorne effect — a volunteer who knows they are being watched will persist through a confusing screen far longer than they would alone on a Saturday morning, which would conceal exactly the abandonment this case is about; use unmoderated sessions for the discoverability task, or say plainly that observed persistence is not evidence of usability. Week 7 slide 39 names four biases — Confirmation bias · Social desirability bias · Hawthorne effect · Culture bias — and gives no definitions, so name them and show you understand them through the mitigation rather than inventing a textbook definition.
• Two further risks, correctly labelled. Moderator influence — a facilitator who nudges or explains contaminates a discoverability test, which is what the 8 Rules guard against: be curious, listen more than you talk, base it in actual behaviour rather than "would you…?". And sampling — recruiting only from the app's active users excludes precisely the volunteers who left, so recruit lapsed volunteers by SMS. Neither of these two is one of the four biases named on slide 39. They are practical testing risks, and worth raising as such — just do not present them as course-named biases.
• Close the loop. State what result would make you reject the hypothesis — e.g. direct success rises but duplicates do not fall, which would mean the confirmation was still not being read.