How to write a literature review in sociology

Sociology makes the literature review harder than most disciplines for a specific reason: the concepts are contested. Two studies of social capital, precarity or community may not be studying the same thing, even when they use the same word, because each has defined and measured it differently. A review that treats the shared label as a shared object produces a chapter that looks coherent and says nothing.

So the job is not to summarise what has been found. It is to show how a field has argued about a concept, which traditions frame it in which way, what each framing makes visible and invisible, and where that leaves the question you intend to ask.

This guide covers the sociological specifics: mapping traditions, tracking operationalisation, searching across a discipline with porous borders, combining qualitative and quantitative evidence, and being honest about your own selection. The general AI-assisted workflow lives in how to write a literature review with AI.

What a sociology literature review is arguing

Your review has a thesis of its own, and it is usually one of a small number of shapes. The field has treated X as a property of individuals when it is better understood as a property of relations. Two traditions have talked past each other because they operationalise the same concept differently. A well-studied mechanism has never been examined in the setting you care about.

Write that sentence early, even badly, and revise it as you read. Without it you will produce coverage, and coverage is the failure mode examiners describe as descriptive. With it, every study you include has a job: it either supports the claim, complicates it, or represents the position you are arguing against.

The distinction shows up in your topic sentences. Descriptive reviews open paragraphs with authors. Argumentative reviews open with claims and use authors as evidence inside the paragraph.

Expect the claim to change. The version you write in month one is a hypothesis about the field, and reading will falsify parts of it. What matters is that a claim exists at every stage, because it is what tells you whether the article you just found is relevant or merely interesting.

Map the theoretical traditions before the studies

Before you sort papers, sort schools. Most sociological questions sit inside two or three broad traditions that disagree about what counts as an explanation, and readers in your subfield will recognise those traditions immediately. Your review needs to show you can locate work within them.

A practical way in is to build a short map, no more than a page, of the traditions relevant to your question.

You do not have to adjudicate between traditions, and attempting it inside a literature review usually overreaches. What you do have to do is locate yourself. Say which tradition frames your question, what that framing brings into view, and what it will consequently leave out, because an examiner working in a different tradition will ask exactly that.

• What each tradition takes as the unit of analysis: individuals, interactions, organisations, fields, structures.
• What each treats as an explanation: mechanisms, meanings, structural position, historical process.
• What methods follow from that: ethnography, interviews, survey and regression, comparative-historical, computational.
• Where the tradition is strongest and what it habitually leaves out.
• Which contemporary authors are working in it now, not only its founders.

Track how each study operationalises the contested concept

This is the analytic core of a sociology literature review, and it is the part that most reliably impresses an examiner. For every empirical study you include, record the definition used and the actual measure or method behind it, in the study's own terms.

When you do this systematically, the mixed findings in your field often stop being mixed. Studies that measure trust with a single survey item and studies that infer it from observed reciprocity are not producing contradictory results about one concept; they are producing consistent results about two. Being able to say that, with the evidence to back it, is a genuine contribution and it frequently becomes the justification for your own design.

The mechanics are straightforward: an evidence matrix with a column for stated definition and another for operational measure. Synthesis Lab builds that matrix across the papers in your project library, and because extractions bind to the sentences they came from, you can quote a definition without having to hunt the page again.

Do the same for qualitative work, adapting the fields. Record the setting, the sampling logic, the form of data and the analytic approach, because a concept built from interviews with people who volunteered is not the same object as one built from sustained observation of an entire workplace.

Searching a discipline with porous borders

Sociological questions rarely stay in sociology journals. Depending on your topic, the important work may sit in public health, criminology, education, geography, social policy, anthropology, economics or communication studies, and each of those fields will use different vocabulary for a concept you consider yours.

Search for the phenomenon under several disciplinary labels, not only the one you were taught. Housing insecurity, residential instability and forced mobility may name overlapping literatures that barely cite each other. Finding the disconnect between them is often where a thesis comes from.

The reverse problem is real too. A term that is precise in your subfield may be generic elsewhere, and searching it alone will bury you. Pair the broad term with a second block naming your population, setting or mechanism, so the results come back at a size a person can actually screen.

Cast the net wide first and filter afterwards. Literature Finder reaches an internal corpus plus live OpenAlex, PubMed, Semantic Scholar and arXiv, which spans several of those neighbouring fields in one pass. Then work by hand: follow citations backwards from a good recent review, follow them forwards to find responses, and check the key journals in your subfield directly, since a keyword search will never fully substitute for knowing where your field publishes.

Grey literature and policy sources

Government statistics, third-sector research, inquiry reports and think tank publications matter in sociology in a way they do not in laboratory disciplines. They are often the only source of population-level data on your topic, and excluding them wholesale can leave a review oddly detached from the phenomenon it describes.

Include them deliberately rather than accidentally. Say in your methods which kinds of non-academic sources you drew on and why, and treat provenance as part of the analysis: who commissioned the report, what it was for, and what that implies about its framing. Handled this way, grey literature strengthens a review. Handled casually, it reads as padding.

Keep them in the same place as everything else so nothing gets lost between a folder and a bibliography. The Library takes PDF uploads alongside papers pulled by DOI, so a mixed corpus of journal articles and reports stays searchable together, with reading statuses to show you what you have actually processed.

Cite them precisely. Reports are often undated on the face of them, revised without notice and republished at new addresses, so record the publisher, the version and the date you consulted it at the moment you add the source rather than a year later.

Combining qualitative and quantitative evidence

Many sociology reviews have to hold both kinds of evidence at once, and the temptation is to segregate them into two halves of the chapter. That produces a review with no argument in the middle.

Organise by question instead, and let the two kinds of evidence answer different parts of it. Quantitative work is usually better at establishing that a pattern exists and how it is distributed. Qualitative work is usually better at explaining how it works and what it means to the people living it. Written well, a section states the pattern, then explains the mechanism, then notes what neither has established.

Do not force qualitative findings into the language of effect sizes, and do not dismiss survey evidence as thin because it is not close-grained. Both moves signal that you are more comfortable in one tradition, which is exactly the impression a review is supposed to correct.

Where the two literatures genuinely conflict, resist the urge to declare one of them right. Ask what each is measuring instead. A survey capturing reported attitudes and an ethnography capturing observed practice may both be accurate about different things, and that divergence is frequently the sociologically interesting result.

Organise by debate, not by decade

Chronological reviews are the default draft and almost always the wrong final structure. They imply that the field progressed steadily to the present, which is rarely how sociology works, and they leave the reader with a timeline rather than a position.

Structure around the live disagreements instead. One section per contested question, competing positions inside it, and a closing paragraph on what remains unsettled. Use chronology only where the sequence is the argument, for instance when a concept was introduced, then critiqued, then reformulated.

Balance canonical and current work. A review with only recent citations reads as unmoored, and a review resting mainly on mid-century theorists reads as out of touch. The usual rule is that canonical work establishes the framing and current work establishes the state of play, so your most recent citations should cluster around the specific question you are asking.

A cheap test for structure: read only your section headings, in order. If they name topics, the chapter is descriptive. If they name questions or tensions, the chapter argues. Rewriting headings is a fast way to discover that a section has no point to make, since a section without an argument cannot be given an argumentative heading.

Reflexivity about what you selected

Sociology takes seriously the idea that a researcher's position shapes what they see, and that applies to your review as much as to your fieldwork. You made choices about which traditions to foreground, which journals to trust, which languages you can read and which geographies you covered, and those choices shaped your conclusions.

A short paragraph acknowledging this is not a weakness. It is a demonstration that you understand your own method. Say how you searched, what you prioritised, and what the review consequently underrepresents, for example scholarship not published in English or work from outside the regions your databases cover well.

The point is to make the selection auditable. If an examiner can see how sources came in, an omission becomes a boundary rather than an oversight.

The same logic applies to what you found persuasive. A review that quietly favours one tradition while presenting itself as neutral is weaker than one that says which account it finds convincing and why. Take the position openly, defend it with evidence, and give the strongest counter-argument its own paragraph rather than a dismissive subordinate clause on the way past.

Citation style and the final pass

Sociology programmes commonly use APA, Harvard or Chicago author-date, all three of which are among the twelve styles Reference Manager supports. Confirm the requirement in your handbook, set it once, and let the bibliography follow rather than formatting entries by hand.

Then run a final pass against the checks below before the chapter leaves your desk.

• Each section makes a claim, and the studies inside it are evidence for that claim.
• Every empirical study cited has its concept definition and measure recorded, at least in your notes.
• Theoretical traditions are named and located, not just quoted in passing.
• Canonical and current work are both present, with the recent citations closest to your own question.
• Grey literature is included deliberately, with its provenance discussed.
• The chapter ends on the unresolved question your study takes up, phrased in the field's own vocabulary.

Common mistakes in a sociology literature review

The recurring weaknesses in this chapter are less about missing literature than about how the literature is handled once it is found.

• Treating a shared word as a shared concept, so studies with incompatible measures are reported as agreeing.
• Name-dropping theorists without using their framework to do any analytical work.
• A chronological march through the field that ends without a position.
• Citing a classic through a textbook summary, which is how misreadings of canonical arguments propagate.
• Restricting the review to sociology journals when the question is being answered next door in public health, geography or education.
• Presenting qualitative and quantitative literature in separate silos and leaving the reader to integrate them.