Exercises — Chapter 11: Citing Sources and Avoiding Plagiarism

Writing is learned by writing. These exercises ask you to diagnose, convert, and fix real citation and integrity problems — not to fill in bubbles. Work them with a source open and a reference manager handy where noted. Selected solutions and rubrics live in appendices/answers-to-selected.md; open-ended tasks carry a self-assessment rubric here.

Part A — Analyze This ⭐

Identify what's right or wrong. One or two sentences each.

A1. A student paper contains: "Studies have shown that agile teams deliver faster." What's wrong with this sentence as a piece of cited writing, and what's the name for the phrase doing the damage?

A2. Here are two references from the same paper's reference list:

[1] Smith, J. (2019). Database indexing strategies. ACM Computing Surveys.
[2] R. Patel, "Caching at scale," IEEE Trans. Software Eng., 2021.

Name the integrity-adjacent problem (it isn't dishonesty — it's a craft failure) and what it signals to a reviewer.

A3. A writer cites a finding to a blog post that summarizes a journal article, rather than to the article itself. Why is this risky even though the blog post is publicly available and the citation is real?

A4. True or false: "Water is composed of hydrogen and oxygen (Pauling, 1960)." Is the citation appropriate? Explain.

A5. A chemistry student submits a manuscript to an ACS journal with all references in APA author–date format. The information is all correct. Why might an editor still send it back?

A6. A researcher writes: "Transformer models use self-attention to weigh the relevance of different input tokens." In an ML conference paper, does this sentence need a citation? Does your answer change if the same sentence appears in a magazine article for a general audience? Explain the principle.

A7. An AI tool produces this reference for a writer's literature review:

M. Chen and L. Okafor, "Reproducibility in distributed systems benchmarks,"
ACM Trans. Computer Systems, vol. 38, no. 2, pp. 112–140, 2020.

It's formatted flawlessly. What is the first thing the writer must do, and why does the flawless formatting tell them nothing about whether it's safe to use?

A8. A paragraph in a paper has a citation at the end only, but draws on that one source across all five sentences, with several sentences echoing the source's phrasing. Which two problems from the chapter does this risk, and how would you tell them apart?

Part B — Revise This ⭐⭐

Rewrite the passage to fix the problem. Give your revision, then one line naming the principle.

B1 — Fix the patchwriting (core skill). Here is a real, citable source and a student's "paraphrase." The paraphrase is patchwriting. Rewrite it as an honest paraphrase with a citation. Then, separately, produce a version that uses a short direct quote for the most distinctive phrase.

SOURCE (William Zinsser, On Writing Well, 2006):
"The secret of good writing is to strip every sentence to its cleanest
components. Every word that serves no function, every long word that
could be a short word, every adverb that carries the same meaning that's
already in the verb, every passive construction that leaves the reader
unsure of who is doing what — these are the thousand and one adulterants
that weaken the strength of a sentence."

STUDENT'S PATCHWRITING (fix this):
"The key to strong writing is to reduce every sentence to its cleanest
parts. Every word that serves no purpose, every long word that could be
a short one, every adverb that conveys the same meaning already in the
verb, every passive construction that leaves the reader unsure who is
doing what — these are the many adulterants that weaken a sentence."

B2 — Fix the patchwriting (your turn, harder). A second patchwritten passage. Source and bad paraphrase below — produce an honest paraphrase with a citation.

SOURCE (Frederick P. Brooks Jr., The Mythical Man-Month, 1975):
"Adding manpower to a late software project makes it later."

STUDENT'S PATCHWRITING (fix this):
"Putting more workers on a delayed software project makes it more
delayed (Brooks, 1975)."

(Hint: this one is hard because the source is already maximally concise — think about what an honest paraphrase of a nine-word aphorism even looks like, and whether quotation might be the more honest choice here.)

B3 — De-fake the citation. A draft sentence reads: "Reviews catch roughly 60% of defects before release (Anderson, 2018)." The writer admits they don't actually have a source for the 60% figure — they remember reading something like it somewhere. Rewrite the sentence honestly, two ways: (a) if they can't find any source, and (b) what they should do before publishing it at all.

B4 — Right-size the citations. This paragraph is over-cited to the point of noise. Revise it to cite only what genuinely needs citing.

"Python (van Rossum, 1991) is a programming language (Wikipedia, 2023)
widely used in data science (O'Reilly, 2020). It is interpreted
(Smith, 2019) rather than compiled (Jones, 2018). Our analysis used
pandas, a Python library for data manipulation (McKinney, 2010), which
reported a 23% increase in null values (our data) after the migration."

B5 — Fix the AI-fabrication risk. A colleague's draft paragraph (AI-assisted, by their own admission) contains three citations to support a claim about edge-computing latency. Rewrite the colleague a two-sentence instruction telling them exactly what to do before this paragraph can go in the report, and why "the citations look real" isn't enough.

B6 — Disclose the reuse. A writer wants to paste a strong methods paragraph from their own published conference paper into a new journal manuscript, unchanged. Write the one or two sentences of guidance you'd give them so the reuse is honest rather than self-plagiarism.

Part C — Write This ⭐⭐–⭐⭐⭐

Produce the document. A scenario, not an answer.

C1 — Convert this citation between styles (required). Here is one real source. Produce its reference-list entry in all four styles — IEEE, APA, Chicago (notes-and-bibliography footnote form), and ACS — and the matching in-text form for each.

THE SOURCE:
Edward R. Tufte, "The Visual Display of Quantitative Information,"
2nd edition, Graphics Press, Cheshire, Connecticut, 2001.

After you produce all four, write two sentences identifying what changes across the four versions and what stays constant — and connect that observation to Chapter 2's central idea about audience.

C2 — Convert a second citation (a journal article). Same task, harder source type — a journal article with volume, issue, and page range. Produce IEEE, APA, and ACS reference-list entries (skip Chicago footnote here) and the in-text form for each.

THE SOURCE:
Peter C. Rigby and Christian Bird, "Convergent Contemporary Software
Peer Review Practices," published in the Proceedings of the 2013 9th
Joint Meeting on Foundations of Software Engineering (ESEC/FSE 2013),
pages 202 to 212.

C3 — Write a paraphrase from scratch. Find a real source on a topic you know (a paper, a textbook chapter, a standard). Read one paragraph until you understand it. Now apply the close-the-source method: close it, write the idea from memory in your own words and structure, reopen to check, and cite. Submit (a) the source paragraph, (b) your paraphrase, and (c) a one-sentence reflection on whether anything in your draft drifted toward the source's wording when you reopened it.

C4 — Write the AI-disclosure line. Imagine your context (a course or a journal) permits AI assistance with disclosure. You used an AI tool to brainstorm an outline and to rephrase two of your own paragraphs (you verified every fact yourself, and used no AI-generated citations). Write the one-to-three-sentence disclosure statement you'd include, honestly describing the tool's role.

Part D — Synthesis & Critical Thinking ⭐⭐⭐

D1 — The translation trap. A writer reads an excellent technical explanation in Spanish, translates it carefully into English, and presents the English as their own original writing — no citation, because "I wrote these English words myself; I did the translation." Make the case, in a short paragraph, for why this is plagiarism. Then state what they should have done.

D2 — Where's the line? (build a rule). AI tools sit on a spectrum from "clearly fine" to "clearly plagiarism." Place these five uses on that spectrum and justify your ordering: (a) AI fixes your grammar and typos; (b) AI suggests a clearer structure for your argument; (c) AI rephrases your paragraph and you verify it still says what you mean; (d) AI writes a paragraph of new content you couldn't have written and you paste it in, verified; (e) AI writes the whole document and you submit it unread. What principle separates the acceptable from the unacceptable? (Connect to Chapter 1's "writing is thinking.")

D3 — Translate the citation norm for three audiences. In three short paragraphs, explain why citation matters to (a) a first-year undergraduate who thinks it's a pointless formatting rule, (b) a working engineer who says "I don't write papers, so this doesn't apply to me," and (c) a manager deciding whether to fund a plagiarism-detection subscription. Same core idea, three different framings — this is Chapter 2 applied to a meta-topic.

D4 — Diagnose the honesty tier. For each claim below, decide which of the book's three tiers (§11.7) applies and how you'd handle it: (a) "Strunk and White's Elements of Style devotes a rule to omitting needless words." (b) "Research suggests that reviewers read tired and skim, which is why structure matters." (c) "Our fictional data scientist, Dana, wrote her churn memo three ways." (d) "The transformer architecture was introduced in a 2017 paper by Vaswani and colleagues." For each, say which tier and what the honest citation move is.

D5 — The self-plagiarism edge case. A PhD student is writing a "stapler thesis" — a dissertation assembled from three of their own already-published papers. Their advisor says this is standard and fine. A friend says "but you can't reuse your own published work, that's self-plagiarism." Reconcile these: under what conditions is reusing your own published work legitimate, and what makes the difference between the legitimate version and actual self-plagiarism?

Part M — Mixed Practice (Interleaved) ⭐⭐–⭐⭐⭐

These mix this chapter with earlier ones, so you have to choose the right tool.

M1 — (Ch 11 + Ch 3, Clarity). Here is a wordy, patchwritten sentence. It has two separate problems — a clarity problem and an integrity problem. Fix both, and label which fix addresses which.

SOURCE: "Vigorous writing is concise." — Strunk & White
DRAFT:  "It is generally the case that writing which is vigorous in
         nature tends to be concise in its overall character."

M2 — (Ch 11 + Ch 9, Visuals). You're writing a data memo and want to include a chart that you built by replotting another team's published dataset in your own style. Two questions: (1) Does this need attribution, and where does it go? (2) Write the interpreting caption (Chapter 9 skill) with the attribution (Chapter 11 skill) for this figure, inventing a plausible finding.

M3 — (Ch 11 + Ch 2, Audience). You're explaining the same finding to two audiences: an academic paper (which expects formal citation) and an internal Slack message to your team. How does the citation differ — or should it — between these two? Is "I read in Rigby & Bird that smaller PRs get reviewed better" an acceptable "citation" for the Slack message? Argue your answer.

M4 — (Ch 11 + Ch 4, Structure). A literature review paragraph cites six sources but reads as "Author A said X. Author B said Y. Author C said Z." It's all correctly cited — no plagiarism. Yet it fails. Using Chapter 4's ideas about structure serving the reader (and previewing Chapter 15), explain what's wrong and sketch how you'd restructure it. (Note: this is the difference between correct citation and good use of sources — both matter.)

M5 — (Ch 11 + Ch 10, Design). A reference list runs two full pages of dense, undifferentiated text. Apply Chapter 10's design principles: name two formatting choices that make it more usable, and explain why a well-designed reference list serves the same goal (findability) as a well-cited claim.

Part E — Extension ⭐⭐⭐⭐ (optional, for the Deep Dive track)

E1 — Build a personal citation workflow. Install Zotero (or Mendeley). Capture ten real sources on a topic you're researching. For each, write an own-words note. Then draft a 200-word synthesis paragraph that cites at least four of them — and that you wrote from your notes, with all sources closed. Reflect: did working from your own notes (rather than the sources) change how easy it was to avoid patchwriting?

E2 — Audit a published paper's citations. Take any open-access paper in your field. Pick five of its citations and try to (a) locate each source, and (b) confirm the source actually supports the claim it's cited for. Report what you found. (You will be surprised how often a citation doesn't quite say what the citing sentence implies — this is why "cite the primary source you actually read" matters.)

E3 — Write the integrity policy. You're the editor of a student technical journal. Write a one-page "citation and AI policy" for contributors that covers: required style, what counts as plagiarism (including patchwriting and self-plagiarism), the common-knowledge boundary, and the rules for AI assistance. Make it clear enough that an honest first-year contributor couldn't accidentally violate it.

Self-Assessment Rubrics (for open-ended tasks)

Patchwriting fixes (B1, B2, M1): A strong fix (1) preserves the source's idea accurately, (2) uses your own sentence structure — a reader couldn't reconstruct the original's wording from yours, (3) includes a citation, and (4) quotes-and-marks any phrase too distinctive to paraphrase. If your "paraphrase" tracks the source clause-by-clause, you've patchwritten again — close the source and write from memory.

Style conversions (C1, C2): Score yourself on (1) correct information in every version (authors, title, year, venue, pages all present and right), (2) correct mechanics per style (brackets vs. parentheses vs. superscript; year placement; author abbreviation; punctuation), and (3) consistency within each style. The metalesson lands if your two-sentence reflection connects "same info, different surface" to audience.

AI/integrity reasoning (B5, C4, D1, D2, D5): A strong answer names the specific integrity principle (authorship, fabrication, outsourced thinking, novelty/disclosure) rather than a vague "it's cheating," and lands on an actionable rule (verify, disclose, cite yourself). The deepest answers connect back to Chapter 1: the line is crossed when the thinking you present as yours wasn't yours.

Selected solutions and full rubrics: appendices/answers-to-selected.md. For the conversion tasks (C1/C2), check your mechanics against the authoritative manual for each style (see Further Reading) — and against your reference manager's output, which you should still verify by hand.