Exercises — Chapter 15: Literature Reviews and Research Synthesis

Synthesis is learned by synthesizing. These exercises ask you to diagnose the difference between summary and synthesis, build a matrix, convert source-by-source prose into thematic arguments, and find real gaps — not to recall definitions. Work them with a draft 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 synthesis and what's summary, what works and what's broken. One or two sentences each.

A1. A review paragraph reads: "Chen (2020) studied X and found A. Diaz (2021) studied X and found B. Park (2022) studied X and found C." Every source is real and accurately reported. Name the problem in one phrase, and give the one-word diagnostic test from the chapter that exposes it.

A2. Two openings for the same paragraph. (i) "Several studies have examined caching." (ii) "Caching policies fall into two camps — heuristic and learned — that trade simplicity against adaptivity." Which opening sets up synthesis, which sets up a list, and why does the first sentence determine the rest?

A3. A section is titled "Causes of Burnout" and then reports three causes from three sources, one per sentence. The heading is thematic. Is the section therefore a synthesis? Explain.

A4. A literature review's gap statement reads: "No prior study has examined our specific framework on Kubernetes clusters in the European banking sector." What's the name for this kind of gap, and what question does it fail to answer?

A5. A student labels their paper's introduction "a systematic review of the literature." They selected the fifteen papers they judged most relevant. Why is the label inaccurate, and what should they call it?

A6. In a synthesis matrix, one column (theme) has only one filled cell out of twelve sources; the other columns are mostly full. The writer ignores the empty column as "not enough data." What did they miss?

A7. A Related Work section ends: "Our work also addresses this problem." The six prior citations above it were each summarized in one sentence. Name two failures here — one about the body, one about the closing sentence.

A8. A reviewer of a narrative review notices it cites only the eight studies that support its thesis and omits three well-known studies with conflicting findings. Is this a synthesis problem, an integrity problem, or both? Explain in one sentence.

Part B — Revise This ⭐⭐

Rewrite the passage. Give the scenario, not the answer. Where a passage needs sources you don't have, keep the existing citations and rebuild the structure around them.

B1. Turn this source-by-source summary into a thematic synthesis. ⭐⭐ (the chapter's core skill — do this one carefully) Rewrite the following as a single synthesized paragraph organized around a claim. The sources must become evidence, not subjects. Add a connective claim that relates them.

Hassan (2018) found that automated tests reduce regression bugs.
Lindqvist (2020) found that code review reduces regression bugs.
Owusu (2021) found that continuous integration reduces regression bugs.
Each study examined a different team and a different practice.

B2. Fix the false gap. ⭐⭐ The gap below fails the "so what?" test. Rewrite it as a real gap — invent a plausible unresolved question or tension for the topic (you may add a sentence of motivation).

"While much research exists on recommendation systems, no study has
examined recommendation systems for a mobile app aimed specifically at
amateur astronomers. This study fills that gap."

B3. Rebuild the citation dump as a positioned Related Work paragraph. ⭐⭐ Group these into categories, characterize each by a limitation, and end by positioning a (fictional) contribution.

"Many scheduling algorithms exist. Ahmed [3] proposed a priority-based
scheduler. Bianchi [4] used reinforcement learning. Cho [5] proposed a
fairness-aware scheduler. Dubois [6] studied scheduling in edge devices.
Eriksson [7] proposed an energy-aware scheduler. Our work also schedules
tasks."

B4. Make the section heading argue. ⭐⭐ A thematic section is titled "Methods Used in Prior Work." Under it, the writer lists which methods various studies used. Rewrite the heading so it states a claim about the methods (invent a plausible claim — e.g., about a shared limitation), and write the one-sentence opening that the new heading demands.

B5. Add the missing connective tissue. ⭐⭐ The two findings below are placed side by side with no relationship stated. Rewrite so the relationship (here, a tension) is the point of the sentence.

"Mbeki (2019) found that asynchronous standups improved developer focus.
Nilsson (2021) found that asynchronous standups slowed the team's
response to urgent issues."

B6. Cut the sprawl. ⭐⭐ A literature review's gap is about whether a security training program changes employee behavior. One of its paragraphs reviews three studies on the history of phishing as an attack technique — interesting, but it never connects to the behavior-change gap. In two sentences, state the principle that decides whether this paragraph stays, and what the writer should do.

Part C — Write This ⭐⭐–⭐⭐⭐

Produce the document. A scenario in; a piece of real synthesis out.

C1. Build a synthesis matrix. ⭐⭐ (the chapter's core tool — required) Below are six fictional-but-realistic source notes on a single topic: whether four-day workweeks affect productivity. Build a synthesis matrix: sources down the rows, themes you derive from the notes across the columns. Fill each cell in your own words; leave cells empty where a source is silent. Then, in two or three sentences below the matrix, name (a) one theme where the sources agree, (b) one where they conflict, and (c) the gap (the sparse column).

S1 (Abara, 2019): A 4-day week at one company kept output constant;
   employees reported higher well-being.
S2 (Brandt, 2020): A 4-day week reduced output by 8% in a manufacturing
   firm; well-being rose.
S3 (Conti, 2021): Found 4-day weeks improved well-being but did not
   measure productivity at all.
S4 (Dlamini, 2022): In knowledge work, output held constant; attributed
   this to fewer low-value meetings. Did not measure manufacturing.
S5 (Eom, 2023): Output rose 2% in software teams; well-being rose;
   noted the effect may not generalize beyond knowledge work.
S6 (Farah, 2023): Reviewed prior studies and noted that almost all
   evidence comes from knowledge work, not manual/shift labor.

C2. Write the thematic synthesis from your matrix. ⭐⭐⭐ Using the matrix you built in C1, write a 150–250-word synthesized review paragraph (or two) on the four-day-workweek question. Lead with a claim. Make the sources evidence. End by naming the gap. Then, beneath it, write the bad version — the same six sources as a source-by-source list — and label the two ❌ Before and ✅ After. (You are producing the chapter's spine yourself.)

C3. Write a CS Related Work section. ⭐⭐⭐ Invent a fictional paper: it proposes a new method for detecting duplicate bug reports. Write a 120–180-word Related Work section. Group prior approaches into 2–3 categories (you may invent plausible citations like [1]–[8]), characterize each by a limitation, and end with one explicit sentence positioning your contribution. No sentence may start with a bare author/citation as its subject more than once.

C4. Write a gap statement and its motivation. ⭐⭐ Pick any topic you know. In 3–4 sentences, write a gap statement that (a) names the type of gap (unanswered question, contradiction, untested-context-that-matters, methodological, or synthesis gap), (b) states the gap, and (c) answers "so what?" — why the field needs it filled. Then add one sentence describing the specific paper you would search for to check the gap isn't already closed.

Part D — Synthesis & Critical Thinking ⭐⭐⭐

Cross-chapter integration; find-the-flaw; translate-for-audience.

D1. The grammar diagnostic at scale. ⭐⭐⭐ Take a literature review, related-work section, or "background" you have written (any length, any course/project). Count how many sentences begin with an author name or a bare citation. Then rewrite the worst three so an idea is the subject and the source becomes parenthetical. Report: the original count, and what changed in the three rewrites. Reflect in two sentences on whether the count revealed your review was closer to summary or synthesis. (If you have no prior review, use the "before" you wrote in C2.)

D2. Find the flaw. ⭐⭐⭐ A draft literature review on a medical topic claims: "No prior study has investigated whether the intervention reduces readmissions." In peer review, a reviewer writes: "See Okeke et al. (2021), which found exactly this." The author had read Okeke but found its sample small and its setting different. Diagnose what the author did wrong, and rewrite the gap sentence the honest way that engages Okeke instead of ignoring it.

D3. Translate the review for three audiences. ⭐⭐⭐ You've synthesized the literature on a topic into the claim: "The evidence that X improves Y is consistent but comes almost entirely from one population, so its generality is untested." Write that same finding three ways: (a) one sentence for a systematic-review methods section (precise, hedged); (b) one sentence for a grant proposal's significance paragraph (motivating, "so what"-forward); (c) one sentence for a blog post for practitioners (plain, concrete). Note in one line what changed across the three and what stayed constant. (Ties Ch 7 register, Ch 17 grants, Ch 28 sci-comm.)

D4. Choose the review type. ⭐⭐⭐ For each scenario, name the appropriate review type (narrative / systematic / scoping) and justify in one sentence: (a) the introduction to a conference paper proposing a new algorithm; (b) a healthcare body deciding whether a treatment works, needing all the evidence and a reproducible method; (c) a new interdisciplinary field where you need to map what research exists before formulating a question; (d) a thesis literature-review chapter arguing for the student's study.

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

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

M1. ⭐⭐ (Ch 11 + Ch 15) A writer's synthesis-matrix cell for a source reads, word for word, the source's abstract sentence. When they draft the review from that cell, what two distinct problems are they set up to commit — one from Chapter 11, one from this chapter — and what single habit prevents both?

M2. ⭐⭐ (Ch 8 + Ch 15) A synthesized paragraph jumps awkwardly between sources even though it's organized by theme. Using Chapter 8's given-new contract, explain what's likely wrong with the sentence openings, and how to fix the flow without changing the content.

M3. ⭐⭐⭐ (Ch 14 + Ch 15) A paper's introduction has a beautifully synthesized literature review but the reader still doesn't know what the paper contributes. Using Chapter 14's hourglass, diagnose which part of the structure is missing and where the gap-to-purpose handoff should happen.

M4. ⭐⭐ (Ch 3 + Ch 15) A gap statement is technically a real gap but is buried in a 60-word sentence full of nominalizations ("the investigation of the determination of…"). Which two chapters' skills does fixing it require, and in what order would you apply them?

M5. ⭐⭐⭐ (Ch 13 + Ch 15) A lab report's Introduction reviews prior work source-by-source, then states a purpose that doesn't obviously follow from it. Explain how converting the review to thematic synthesis would also fix the disconnected purpose statement. (Ties the IMRaD Introduction to the gap.)

Part E — Extension ⭐⭐⭐⭐

For motivated readers and the Deep Dive track.

E1. Reverse-engineer a real review. ⭐⭐⭐⭐ Find a published literature review or a survey paper in your field (or the Related Work section of a well-regarded paper). Reconstruct its implied synthesis matrix: what are its themes (columns), and which sources populate each? Then assess: does it synthesize or summarize? Where is its gap, and is it a real one (passes "so what?") or a false one? Is it fair to conflicting findings? Write a one-paragraph critique using this chapter's vocabulary.

E2. The honest systematic-review protocol. ⭐⭐⭐⭐ For a narrow question in your field (e.g., "does pair programming reduce defects?"), draft — don't execute — a one-page systematic-review protocol: the specific databases you'd search, the search string, your inclusion/exclusion criteria, and what you'd record for each included study. Then write two sentences on why writing this protocol before searching is what makes the result reproducible and cherry-proof — and why a narrative review can't make the same claim.

E3. The disappearing gap. ⭐⭐⭐⭐ Take a gap you've claimed (or one from this chapter's examples). Spend twenty minutes genuinely trying to close it — search for the paper that already answered it. Report what you found: did the gap survive, narrow, or vanish? Reflect in two sentences on why this adversarial search is part of intellectual honesty, not an optional nicety. (Ties Ch 38 ethics.)

Self-assessment rubric for the open-ended tasks (B1–B6, C1–C4, D1–D4, E1–E3). A strong response: (1) makes ideas the subjects of sentences, with sources as parenthetical evidence — not author-name openings; (2) states relationships between sources (agree / conflict / extend), not just adjacency; (3) organizes by theme/claim, not by source; (4) where a gap is involved, names a real gap that passes "so what?" and is honest about work that might close it; (5) models the book's own clarity — no nominalizations, every sentence earning its place. For the before/after tasks, the "after" should be one where a reader could no longer shuffle the sentences without losing the argument. Full rubrics and worked solutions for B1, C1, C3, and D2: appendices/answers-to-selected.md.