Quiz — Chapter 9: Visuals and Data

Target: 70%+ before moving on. Answers and explanations are hidden—try each before expanding.

Section 1 — Multiple Choice

1. The single most under-taught data-display skill, per this chapter, is: - A) Choosing the right color palette - B) Writing a caption that interprets the figure, not just labels it - C) Using software to generate charts - D) Putting figures in an appendix

Answer**B.** The chapter's spine is the interpretive caption—stating what a figure *means*, not what it *is* (§9.2, §9.5). A label ("Figure 3: Churn by cohort") leaves the conclusion to a reader who may not reach it, especially since many readers read *only* the figures and captions. Color (A) and tooling (C) are secondary; appendices (D) are a placement decision, not the core skill.

2. You need to convey that average response time dropped from 1,400 ms to 180 ms. The best tool is: - A) A bar chart - B) A 12-row table - C) Prose (a sentence) - D) A pie chart

Answer**C.** Two numbers and a simple before/after comparison belong in a sentence—"response time dropped from 1,400 ms to 180 ms, an 87% reduction" (§9.1). A chart would be decoration around one fact; reserve figures for *shapes* (trends, correlations) and tables for *exact lookup across many rows*.

3. Tufte's "data-ink ratio" principle says you should: - A) Add gridlines so values are easy to read - B) Use 3-D effects to make charts engaging - C) Maximize the proportion of ink that represents actual data; erase the rest - D) Use as many colors as there are categories

Answer**C.** Data-ink is the ink representing data; the principle is to maximize its share and delete non-data-ink—gridlines, 3-D, gradients, redundant legends (§9.3). A, B, and D all *add* non-data-ink that competes with (and can distort) the data.

4. "Figure 3 shows the relationship between onboarding time and churn." The weakness is: - A) It's too long - B) It's about the figure instead of the finding, and wastes the topic-sentence slot - C) It uses passive voice - D) Nothing—it's a good data sentence

Answer**B.** It announces that a figure exists and makes the reader go look, instead of stating the finding (§9.6). The fix leads with the claim and demotes the figure to a citation: "Slow-onboarding customers churn 7× as often (Figure 3)." It also wastes the paragraph's most valuable real estate—the first sentence—on an administrative announcement ([Ch 4](../../part-01-writing-is-thinking/chapter-04-structure/index.md)).

5. Which is a Level-3 (interpretive) caption? - A) "Figure 4." - B) "Figure 4. Revenue by quarter." - C) "Figure 4. Revenue declined over the four quarters." - D) "Figure 4. Revenue fell 23% in Q4—the first decline in nine quarters—driven by enterprise non-renewals."

Answer**D.** It states what the data *means* (cause + magnitude + context), not just what it is (B, a label) or what it shows (C, an observation). A is a non-caption (Level 0). The progression is *what it is → what it shows → what it means* (§9.5).

6. A chart that helped you find an insight (dense, multi-series, default styling) is called ___ and should usually ___ in the final report: - A) explanatory; appear prominently - B) exploratory; be rebuilt as a simpler explanatory figure (or moved to an appendix) - C) chartjunk; be deleted entirely - D) a small multiple; be enlarged

Answer**B.** Exploratory graphics are for the analyst during analysis; publishing them is a classic failure because the reader lacks your context (§9.4). Rebuild the one point as a clean explanatory figure for the reader; keep the exploratory version for an appendix or a technical peer.

7. Starting a bar chart's y-axis at 90 instead of 0: - A) Is fine for any chart - B) Misleads, because a bar encodes value by length, so proportions are distorted - C) Is required by Tufte - D) Only matters for line charts

Answer**B.** A bar's length encodes its value, so truncating the axis makes small differences look large—an ethically fraught error (§9.7, §9.9, Mistake 6). Start bars at zero. (Line charts, which encode value by position not length, can sometimes start elsewhere to show a trend—if you say so.)

8. "Interpret, don't present" (§9.7) means: - A) Show every figure you have and let the reader find the story - B) State the finding, then use figures as evidence in an order that builds the case - C) Never use figures - D) Always use a pie chart

Answer**B.** Presentation hands the reader a warehouse of exhibits; interpretation makes a claim and marshals the figures that support it (cutting ones that don't), ending on an action. Same data, but an argument instead of an inventory (§9.7).

9. A report shows the same monthly revenue numbers in a table, a line chart, and spelled out in the prose. This is: - A) Thorough and good - B) The "redundant figure" mistake—each element should carry what the others don't - C) Required for accessibility - D) An example of small multiples

Answer**B.** Delivering one fact three ways wastes the reader's time; each element must earn its place (§9.9, Mistake 4; theme 6). If the point is a trend, keep the line chart; if exact values are needed for reference, keep one table—not all three plus the prose.

10. For accessibility, a figure should at minimum: - A) Be large and colorful - B) Have alt-text describing what it shows and means, and not rely on color as the only channel - C) Always be a pie chart - D) Never appear in a printed document

Answer**B.** Alt-text (conveying content and meaning), plus redundant encoding so color isn't the only carrier of information (use labels/patterns/position), are the baseline (§9.9). When exact values are the point, a table can be *more* accessible than a figure.

11. The Challenger charts failed, in data-display terms, because: - A) The data was wrong - B) The decisive temperature-vs-damage relationship was spread across many exhibits and never isolated into one clean figure that made it unmissable - C) They used too few colors - D) The engineers didn't have the data

Answer**B.** Per Tufte's analysis, the data existed but no single figure isolated "colder → more damage, and the forecast is colder than any prior flight"; scattered across charts and tables, the fatal pattern stayed invisible to time-pressured readers (§9.8). The failure was display, not knowledge.

12. The reliable template for an interpretive caption is: - A) "Figure N. [Topic]." - B) "[Finding], shown by [comparison]: [values], which means [implication]." - C) "As you can see in Figure N…" - D) "[Long description of the chart's every wiggle]."

Answer**B.** This template forces the two things weak captions omit—the *claim* and the *implication*—around the key *values* (§9.5). A is a label; C buries the finding; D narrates geometry instead of meaning.

Section 2 — True/False with Justification

T1. "A clear, accurate chart communicates its point on its own; a caption is just a label."

Answer**False.** This is the exact misconception the threshold concept overturns (§9.2). Data is not a conclusion; a reader—especially a scanner or non-expert—sees shapes, not your insight. The caption supplies the meaning, and for caption-only readers it *is* the entire message. Accurate ≠ self-explanatory.

T2. "You should always include every figure you generated during analysis so the reader can see your full work."

Answer**False.** Those are *exploratory* graphics for you, the analyst (§9.4). Publishing them dumps your search on a reader who lacks context. Include *few* explanatory figures, each carrying one point; relegate exploratory detail to an appendix or a technical peer review. (Also theme 6: every element earns its place.)

T3. "A truncated y-axis is always wrong and can never be used honestly."

Answer**False, with nuance.** For *bar* charts it's misleading, because bars encode value by length, so truncation distorts proportion—start bars at zero (§9.7). But *line* charts encode value by position, so a non-zero axis can legitimately reveal a trend *if you disclose it.* The rule is honesty and matching the encoding, not a blanket ban.

T4. "In every field and journal, captions must interpret the data."

Answer**False.** Some journals require *descriptive* captions and reserve interpretation for the Results/Discussion text (§9.9, "It depends"; [Ch 35](../../part-07-writing-for-specific-fields/chapter-35-writing-for-science/index.md)). Know your venue. Even then, make the *prose* interpretive and the figure clean—the interpretation has to live *somewhere*; the convention just dictates where.

T5. "Writing about data well means describing what the chart looks like so the reader can picture it."

Answer**False.** Describing the chart's geometry ("the line rises then dips then rises") is a mistake (§9.9, Mistake 7)—the reader can see the line. Describe what the shape *means* ("growth stalled, then resumed after the redesign"). Lead with the finding, not the picture.

T6. "A table is sometimes more accessible than a figure."

Answer**True.** When *exact values* are the point, a table is often better for screen-reader users (who navigate tables well) and preserves precise numbers a chart only approximates (§9.9). Accessibility, like tool choice, depends on what the reader needs to do—see a shape (figure) or retrieve a value (table).

Section 3 — Short Answer

S1. State the three "levels" a caption can occupy and the one-line difference between them.

Model answer**Level 1 (label):** says what the figure *is* ("Churn by cohort"). **Level 2 (observation):** says what it *shows* ("Churn rises with onboarding time, 3% to 22%"). **Level 3 (interpretation):** says what it *means*, with the implication ("Onboarding, not price, drives churn… intervene in the first week"). *Rubric:* names all three and captures the *is → shows → means* progression.

S2. Give the formula for writing about data in prose, and rewrite "Table 2 shows the benchmark results" using it (invent a plausible result).

Model answer**Formula:** *claim + key number + (Figure/Table N)*—lead with the finding, state the load-bearing number in words, demote the reference to a parenthetical citation. **Rewrite:** "Postgres sustained 14,000 inserts/sec, well ahead of MySQL (9,000) and SQLite (6,500) on write-heavy workloads (Table 2)." *Rubric:* finding leads; number present; table cited parenthetically.

S3. A colleague asks, "Should this go in a figure, a table, or prose?" What's the single diagnostic question you'd tell them to ask?

Model answer**"What will the reader *do* with this—see a shape, look up a value, or follow an argument?"** Shape → figure; exact lookup → table; one/two numbers or reasoning → prose (§9.1). *Rubric:* frames the decision around the reader's task, not "what looks professional."

S4. In one sentence each, name two distinct things "chartjunk" does to a chart (beyond wasting ink).

Model answer(1) It **obscures the data**—heavy gridlines, busy backgrounds, and dense hatching compete with the signal for the reader's attention. (2) It can **distort the data**—3-D tilt makes pie slices and bars misrepresent their true values, so the decoration actively misleads (§9.3). *Rubric:* obscures + distorts (or equivalents).

S5. What's the difference between presenting data and interpreting it?

Model answer**Presenting** shows the reader your figures and numbers and trusts them to find the story (a warehouse of exhibits). **Interpreting** states the finding and uses the figures as evidence, in an order that builds the case to an action (§9.7). One is an inventory; the other is an argument. *Rubric:* captures "show-and-trust" vs. "claim-then-evidence."

Section 4 — Applied Scenario

A1. (Caption + tool choice.) You have this result: customers who get a response from support within 24 hours renew at 88%; 2–3 days, 71%; 4+ days, 52%. (a) Decide figure/table/prose and justify in one sentence. (b) Write a Level-3 interpretive caption (if you chose a figure) or the data sentence (if prose).

Rubric**(a)** A *figure* (bar chart, three cohorts) is defensible because the point is a *shape*—renewal falls as response time grows; prose is also defensible since it's three numbers carrying an argument. Either is fine if justified by the reader's task. **(b)** Level-3 caption, e.g.: *"Figure X. Fast support response drives renewal: customers answered within 24 hours renew at 88%, versus 52% for those waiting 4+ days—a 36-point gap that makes first-day response the highest-leverage retention lever."* Score: finding named (2) · key numbers in words (2) · implication explicit (2) · self-contained (2). 7–8 = excellent.

A2. (Fix the passage.) Rewrite this so it leads with the finding, interprets, and cites the figure properly; then write a Level-3 caption for Figure 1.

"Figure 1 shows our experimental results. As the figure demonstrates, there are differences between the groups. Overall the data shows our intervention had an effect." (Result: treatment improved 18% vs. 4% control; effect held across subgroups.)

Rubric**Strong rewrite:** "Our intervention worked: the treatment group improved 18% versus 4% for controls (Figure 1), and the effect held across every subgroup—so the result isn't driven by one segment." **Caption:** "Figure 1. The intervention produced a 4.5× larger improvement than the control (18% vs. 4%), consistent across subgroups—evidence the effect is real and general." Score: finding leads (2) · number in words (2) · figure demoted to citation (2) · caption reaches Level 3 (2). Deduct for any surviving "Figure 1 shows…" or vague "there are differences."

Scoring & Next Steps

Score What it means Do this
< 50% Core ideas not yet solid Re-read §9.2 (caption threshold) and §9.5 (interpretive captions); redo Section 1.
50–70% Recognize the ideas, shaky on production Redo Exercises Part B (Fix the caption, Lead with the finding).
70–85% Solid—proceed Continue to Chapter 10; keep the Project Checkpoint going.
> 85% Strong Try Exercises Part E (small multiples, the ethics of the axis) and the Deep Dive in Case Study 2.

Review: Key Takeaways · Further Reading · back to Chapter 9