Chapter 31 Quiz: Communicating Results: Reports, Presentations, and the Art of the Data Story

Instructions: This quiz tests your understanding of Chapter 31. Answer all questions before checking the solutions. For multiple choice, select the best answer. For short answer questions, aim for 2-4 clear sentences. Total points: 100.

Section 1: Multiple Choice (10 questions, 4 points each)

Question 1. What is the key difference between a "finding" and an "insight"?

  • (A) A finding is quantitative; an insight is qualitative
  • (B) A finding describes what the data shows; an insight adds context, interpretation, and suggests action
  • (C) A finding is preliminary; an insight is final
  • (D) A finding is wrong; an insight is correct
Answer **Correct: (B)** A finding is a fact extracted from data ("churn increased 8%"). An insight places that finding in context and suggests what to do about it ("churn increased 8%, driven by discount customers, suggesting we need engagement checkpoints"). Insights drive action; findings just describe data.

Question 2. The Pyramid Principle recommends that data communications should be structured:

  • (A) Chronologically — first what you did, then what you found
  • (B) From simple to complex concepts
  • (C) With the conclusion first, then supporting evidence, then details
  • (D) With the methodology first to establish credibility
Answer **Correct: (C)** The Pyramid Principle, developed at McKinsey, inverts the natural chronological order. You lead with the answer (conclusion/recommendation), then provide supporting evidence, then provide detailed methodology for those who want it. This respects the audience's time and ensures the key message is received even if they stop reading early.

Question 3. Which of the following is an example of an assertion-evidence slide title?

  • (A) "Q3 Sales Data"
  • (B) "Sales increased 15% in Q3, driven by the new enterprise tier"
  • (C) "Overview of Results"
  • (D) "Figure 7: Revenue by Quarter"
Answer **Correct: (B)** An assertion-evidence title makes a claim (an assertion) that is then supported by the visual evidence on the slide. It is a complete sentence that states the insight. (A), (C), and (D) are descriptive or topic-based titles that tell the viewer what the slide is about, but not what to conclude.

Question 4. When presenting to a non-technical executive audience, which approach is most appropriate?

  • (A) Show all your code so they can verify your work
  • (B) Focus on methodology to establish credibility
  • (C) Lead with the recommendation and business impact, using plain language
  • (D) Present the full statistical analysis including p-values and confidence intervals
Answer **Correct: (C)** Executive audiences want to know what you found, what it means for the organization, and what they should do. They do not need code, methodology details, or statistical jargon. Lead with the recommendation, support it with business-relevant evidence in plain language, and keep methodology in backup slides.

Question 5. What is the "curse of knowledge" in the context of data communication?

  • (A) Having too much data to present
  • (B) Once you know something, it becomes hard to imagine not knowing it, causing you to skip context and use jargon
  • (C) The risk of sharing confidential data
  • (D) The tendency for knowledgeable presenters to be boring
Answer **Correct: (B)** The curse of knowledge is a cognitive bias where experts forget what it is like to be a novice. Data scientists who have spent weeks with a dataset assume their audience shares their context, leading to skipped explanations, unexplained jargon, and assumptions that charts are self-explanatory. Recognizing this bias is the first step to overcoming it.

Question 6. Which annotation type is MOST effective for helping a viewer quickly understand a trend chart?

  • (A) A decorative border around the chart
  • (B) A 3D effect on the data points
  • (C) A text callout marking the key inflection point with an explanation
  • (D) A legend with 15 color-coded categories
Answer **Correct: (C)** Text annotations — callout boxes, labeled data points, event markers — guide the viewer's attention to what matters. They do the interpretive work for the viewer rather than expecting them to find the insight on their own. Decorative elements (A, B) are chartjunk. An overly complex legend (D) adds cognitive load.

Question 7. The four parts of a data story narrative arc are:

  • (A) Introduction, Methods, Results, Discussion
  • (B) Situation, Complication, Resolution, Call to Action
  • (C) Abstract, Data, Analysis, Conclusion
  • (D) Hook, Evidence, Rebuttal, Summary
Answer **Correct: (B)** The data story narrative arc follows: Situation (establish context and baseline), Complication (what changed or what is wrong), Resolution (what the analysis reveals), and Call to Action (what the audience should do). This mirrors classic storytelling structure and is designed for persuasive, audience-centered data communication. (A) is the IMRAD format for scientific papers, not data storytelling.

Question 8. A dashboard is the WRONG choice when:

  • (A) Different users need to filter data by different dimensions
  • (B) Data updates regularly and users need current information
  • (C) You need to tell a specific story with a specific conclusion
  • (D) Users want to explore data and answer their own questions
Answer **Correct: (C)** Dashboards are designed for monitoring and exploration, not for telling a specific story. When you need to convey a particular narrative with a conclusion and recommendation, a presentation or report is more appropriate. Dashboards work best when data updates regularly (B), users need different views (A), and self-service exploration is valuable (D).

Question 9. "The regression coefficient was 0.43 with p < 0.01." Which is the best translation for a non-technical audience?

  • (A) "The number was 0.43 and it was very small."
  • (B) "For every additional clinic per 100,000 residents, vaccination rates increased by about 4 percentage points, and we're highly confident this is a real pattern."
  • (C) "The statistical analysis was positive."
  • (D) "The data was significant."
Answer **Correct: (B)** Effective translation gives the coefficient a real-world meaning (what happens when X increases by one unit), uses natural units the audience understands (percentage points, not regression coefficients), and conveys statistical significance in plain language ("highly confident this is a real pattern" instead of "p < 0.01"). The other options lose critical information.

Question 10. What is "false precision" in data communication?

  • (A) Rounding numbers too aggressively
  • (B) Reporting numbers with more decimal places than the data supports, implying unwarranted accuracy
  • (C) Using precise numbers instead of estimates
  • (D) Making predictions that turn out to be wrong
Answer **Correct: (B)** False precision means reporting "The vaccination rate is 78.3472%" when "about 78%" would be more honest. Extra decimal places imply a level of measurement precision that the data does not support. This can mislead audiences into thinking estimates are more certain than they are. Rounding to meaningful precision is a sign of good communication, not laziness.

Section 2: True or False (4 questions, 4 points each)

Question 11. True or False: A narrative notebook and a lab notebook serve the same purpose and should be structured the same way.

Answer **False.** A lab notebook documents the messy process of exploration (useful for the analyst). A narrative notebook tells a coherent analytical story (useful for the audience). Lab notebooks often have cells in random order, dead code, and minimal narration. Narrative notebooks read top-to-bottom with Markdown prose guiding the reader through question, analysis, and interpretation.

Question 12. True or False: If your chart is clear enough, you do not need annotations — the data should speak for itself.

Answer **False.** Data rarely "speaks for itself." Even well-designed charts benefit from annotations — assertive titles, labeled data points, reference lines, and event markers. Annotations guide the viewer's attention and prevent misinterpretation. The belief that data is self-explanatory is one of the most common communication mistakes in data science.

Question 13. True or False: When you discover a finding that contradicts your recommendation, the ethical approach is to include it in your communication and explain how it fits into the overall picture.

Answer **True.** Cherry-picking — presenting only the evidence that supports your conclusion while hiding contradictory findings — is an ethical violation. The honest approach is to present all relevant findings, including those that complicate your narrative, and explain why your recommendation still holds (or acknowledge that the evidence is mixed).

Question 14. True or False: The most important element on a dashboard should be the largest and placed at the top.

Answer **True.** Dashboard design follows the inverted pyramid principle. The most important metric — the "big number" — should be the largest, most prominent element, placed at the top of the page. This ensures that users who glance at the dashboard for a few seconds still get the most critical information. Detail and exploration options go further down the page.

Section 3: Short Answer (4 questions, 6 points each)

Question 15. Explain the difference between the three audience types (technical, managerial, public/executive) and give one example of how you would present the same finding differently to each.

Answer **Technical audiences** (data scientists, engineers) want to know *how* you did the analysis — they care about methodology, code, and reproducibility. **Managerial audiences** (directors, product managers) want to know *what you found* and *what to do* — they care about business impact and recommendations. **Public/executive audiences** (C-suite, general public) want to know *why they should care* — they care about the big picture and human impact. Example finding — a churn prediction model: - **Technical:** "The random forest achieved AUC 0.89 with feature importance dominated by session_frequency (0.31) and days_since_login (0.27)." - **Managerial:** "We can predict 82% of customers who will churn within 30 days. The biggest warning signs are login frequency and recency. I recommend integrating alerts into the CRM." - **Executive:** "We've built a system that identifies at-risk customers before they leave, which could save an estimated $2M in annual revenue."

Question 16. What are the five components of an effective executive summary?

Answer 1. **The headline** — one sentence capturing the key finding or recommendation. 2. **The context** — two to three sentences explaining why the analysis was done. 3. **The key findings** — three to five bullet points, each stating an insight (not a raw finding), in plain language. 4. **The recommendation** — a specific, actionable next step. 5. **The caveats** — honest acknowledgment of limitations and what would strengthen the analysis. The structure follows the Pyramid Principle: conclusion first, evidence second, details third. An effective executive summary can be read in under two minutes and provides enough information for a decision-maker to act.

Question 17. Describe two ways that data visualization can be ethically misleading even when the underlying data is accurate.

Answer **Truncated y-axis:** Starting a y-axis at a value other than zero can make small changes appear dramatic. For example, showing revenue from $9.8M to $10.2M makes a 4% increase look like a doubling. The data is accurate, but the visual impression is misleading. **Cherry-picked time range:** Choosing a start date that makes a trend look better (or worse) than the longer-term pattern. For example, showing stock performance from a market low to today creates an impression of strong growth, while showing it from a market high would create an impression of decline. Both are "accurate" but tell different stories depending on the chosen window. Other valid answers include: dual y-axes creating false visual correlations, inconsistent scales across compared charts, and using area/volume encodings that distort proportional differences.

Question 18. Explain the concept of "progressive disclosure" in slide design and why it is effective.

Answer Progressive disclosure means revealing information gradually rather than all at once. In slide design, this means building up a complex chart step by step — first showing the axes, then one data series, then a comparison series, then annotations. Each step adds one layer of information. This technique is effective because it guides the audience's attention, prevents cognitive overload (the audience does not have to parse a complex chart all at once), and allows the presenter to narrate each element as it appears. It mirrors the way people naturally process information — one piece at a time — rather than dumping everything at once and hoping the audience figures out what to focus on.

Section 4: Applied Scenarios (2 questions, 8 points each)

Question 19. You have discovered that employee satisfaction scores correlate strongly with the number of plants in the office (r = 0.72). Your manager asks you to present this to the leadership team as evidence that the company should buy more plants.

Describe how you would present this finding responsibly. Address: (a) the correlation-causation issue, (b) how you would frame the finding for a non-technical audience, and (c) what recommendation (if any) you would make.

Answer **(a) Correlation-causation:** This is a classic confounding situation. Offices with more plants likely have other features that contribute to satisfaction — natural light, modern design, better maintenance, larger floor space, or a management culture that invests in employee wellbeing. Plants may be a proxy for overall office quality, not a cause of satisfaction. Presenting this as "plants cause satisfaction" would be misleading. **(b) Framing:** "We found an interesting association between office environment and satisfaction. Offices that score highest on satisfaction also tend to have more natural elements like plants. However, this does not mean that adding plants alone will boost satisfaction — these offices also tend to have better lighting, more space, and more modern amenities. The plants may be a signal of a broader investment in the work environment." **(c) Recommendation:** "Rather than simply buying plants, I'd recommend a broader office environment assessment. If we want to test whether specific environmental changes improve satisfaction, we could pilot upgrades (including but not limited to plants) in a few offices and measure satisfaction before and after. This would give us much stronger evidence before investing across all locations." This approach is honest about uncertainty, avoids the causal claim, and still offers a constructive path forward.

Question 20. You need to communicate the results of a student performance analysis to three different stakeholders in the same week: (1) the data team in a Monday meeting, (2) the school principal on Wednesday, and (3) parents at a Thursday evening open house.

For each audience, specify: the format you would use, the level of detail, one thing you would definitely include, and one thing you would definitely leave out.

Answer **1. Data team (Monday):** - **Format:** Jupyter notebook shared via repository, discussed in meeting - **Level of detail:** Full — methodology, code, statistical tests, limitations - **Include:** Reproducible code, data sources, confidence intervals, alternative models tested - **Leave out:** Policy recommendations (that is for the principal to decide, not the data team) **2. School principal (Wednesday):** - **Format:** 8-10 slide presentation with one-page executive summary handout - **Level of detail:** Moderate — key findings with supporting charts, recommendations - **Include:** Specific, actionable recommendations (e.g., "redirect resources from X to Y") with estimated impact - **Leave out:** Code, statistical test details, methodology specifics (available in appendix if asked) **3. Parents (Thursday):** - **Format:** 4-5 slides with large, simple charts, plus a one-page handout to take home - **Level of detail:** High-level — big-picture trends and what they mean for children - **Include:** What the school is doing in response and how parents can help - **Leave out:** All statistical terminology, complex charts, anything that could be interpreted as blaming parents or specific teachers The key principle: same analysis, three completely different presentations, because the audiences have different knowledge levels, concerns, and information needs.

Scoring Guide

Section Points
Multiple Choice (10 × 4) 40
True/False (4 × 4) 16
Short Answer (4 × 6) 24
Applied Scenarios (2 × 8) 16
Total 96

Note: Scores above 96 are not possible; the remaining 4 points are reserved for exceptional depth in short answer or applied scenario responses. Passing score: 70/100.

Communication is a skill, not a talent. Like any skill, it improves with practice. The fact that this quiz asked you to think about communication rather than do it is a limitation — the real test is your next presentation, your next report, your next notebook. Apply what you have learned.