Literature Review Rules

A literature review's value is the synthesis — the argument about what the body of work as a whole shows, where it disagrees, and what it hasn't answered yet — not a list of paper summaries; a review organized paper-by-paper has done the reading but skipped the actual work of reviewing it.

Synthesize, don't list-summarize

• Never structure a review as "Paper A found X. Paper B found Y. Paper C found Z." with one paragraph per source in sequence — that structure signals summary, not synthesis, no matter how well each individual summary is written.
• For every claim in the review, ask what it says about the field's state of knowledge, not just what one paper found — a synthesized sentence integrates multiple sources making or complicating the same point ("Several studies converge on X (Smith 2020; Lee 2021), though Chen 2022 found this held only under Y condition").
• Write topic sentences that state a claim about the literature, then use citations as evidence for that claim — not topic sentences that name a paper and then describe its contents.

Organize thematically, not by source

• Structure the review around themes, questions, methods, or theoretical positions relevant to the research question — not around the chronological order sources were published, nor around one section per author.
• Decide the organizing structure (thematic, methodological, or theoretical) before drafting, based on what best serves the review's argument — a review of a mature field with competing schools of thought organizes around those schools; a review of an emerging area might organize around open questions instead.
• A single source will often appear in multiple sections if it's relevant to multiple themes — don't force one paper into exactly one bucket just because that's cleaner to write; follow the ideas, not the source list.

Map contradictions explicitly

• When sources disagree, state the disagreement directly and by name — don't quietly cite only the sources that support the review's eventual argument while omitting or burying contradicting findings; that misrepresents the field's actual state of knowledge.
• Investigate why sources disagree before just reporting that they do — different populations, methods, time periods, or definitions of the same construct often explain apparent contradictions and are worth surfacing as part of the synthesis.
• Distinguish genuine empirical contradiction (two studies with comparable methods reaching opposite results) from merely different scope (two studies answering related but not identical questions) — conflating the two overstates the field's actual disagreement.

Identify the gap

• Every literature review supporting a research proposal or thesis must end by identifying a specific, well-evidenced gap the review itself has demonstrated — not an assumed gap declared before doing the reading, and not a vague "more research is needed" placeholder.
• Ground the gap claim in what the review has already shown: a question no cited source has addressed, a population or context existing studies haven't covered, or a contradiction existing studies haven't resolved — the gap section should feel like a direct consequence of the synthesis above it, not a new argument introduced from nowhere.
• Avoid claiming a gap that a broader search would immediately falsify — verify with an actual search (not just the sources already collected) that the specific angle being claimed as novel hasn't already been directly addressed elsewhere.

Source quality tiers

• Weight and describe sources according to their evidentiary tier: peer-reviewed empirical studies and systematic reviews/meta-analyses outrank single case studies, which outrank conference proceedings and preprints, which outrank non-peer-reviewed or trade sources — and say so explicitly when a claim rests on a lower tier ("early evidence from a single case study suggests...").
• Note methodological limitations of key sources when they matter to the claim being drawn from them (small sample size, non-representative population, correlational rather than causal design) — citing a source without noting a limitation that materially weakens the claim it's being used to support is misleading.
• Prefer recent sources for fast-moving fields and be explicit when relying on an older foundational source that newer work has built on or partially superseded — don't let a review's evidentiary base silently skew outdated in an area that has since moved on.

Scope discipline

• Define the review's scope (time range, population, methodology, geographic or disciplinary boundary) explicitly before starting the search, and state that scope in the review itself — an unscoped review either becomes unmanageably broad or arbitrarily narrow.
• Exclude sources outside the defined scope even when they're interesting or well-known — scope creep from including "just one more relevant paper" repeatedly is how reviews balloon past their useful length without adding proportional value.
• Re-check scope discipline at the end: if a section has drifted into a tangential subtopic that doesn't serve the research question the review supports, cut it or move it to a "future directions" note rather than leaving it bloating the main synthesis.