Unit 1: Literature Review Techniques and Tools

A literature review is the systematic identification, evaluation and synthesis of the recorded body of work produced by scholars on a topic. In dissertation research it is not a summary of readings but an argued account of what is known, how it is known, and where the gaps lie that the candidate's own study will address. The review governs everything that follows: it justifies the research question, frames the methodology, and supplies the benchmark against which findings are later interpreted.

Defining properties the rest of the unit relies on:

• Purpose: to map the intellectual territory — establishing context, avoiding duplication, and locating the research gap that motivates the dissertation.
• Systematic character: the search must be reproducible; another researcher following the same protocol should retrieve substantially the same corpus.
• Critical, not descriptive: sources are weighed for validity, relevance and authority, not merely listed.
• Iterative: searching, screening and synthesis loop back on one another as new keywords and citations surface.
• Source hierarchy: primary sources (original studies, experiments, texts) outrank secondary sources (reviews, textbooks), which outrank tertiary sources (encyclopaedias, indexes) as evidence.
• Currency and coverage: the corpus must be both recent enough to reflect the current state and deep enough to include seminal foundational work.

II. Procedure of Literature Search and Review

The end-to-end workflow from question to synthesis

The procedure is a defined sequence that converts a broad topic into a defended, written synthesis. It proceeds through problem definition, source location, retrieval, screening, critical reading, organisation, and finally the writing of a coherent narrative or argument.

A. Defining the scope and search question

Every search begins by narrowing an open topic into a bounded, answerable question.

• Topic to question: convert "renewable energy" into "What factors affect rooftop solar adoption in urban households since 2015?" — bounded by subject, population and timeframe.
• Framework for structure: use PICO (Population, Intervention, Comparison, Outcome) in clinical/experimental work, or SPIDER (Sample, Phenomenon of Interest, Design, Evaluation, Research type) for qualitative work, to decompose the question into searchable concepts.
• Inclusion and exclusion criteria: fix these before searching to keep screening objective:
  • Inclusion: peer-reviewed, English-language, published 2015–2025, empirical studies.
  • Exclusion: editorials, non-peer-reviewed preprints, opinion pieces, studies outside the target population.
• Scope boundaries: decide the date range, geography, discipline and study types up front so the search is reproducible.

B. Identifying keywords and search strategy

The question's core concepts are translated into a controlled vocabulary of search terms.

• Concept mapping: break the question into 2–4 concept blocks; for the solar example the blocks are rooftop solar, adoption, and urban households.
• Synonyms and variants: list alternatives per block — "photovoltaic", "PV", "solar panel" — to avoid missing indexed work.
• Boolean operators: combine terms to widen or narrow retrieval:

("rooftop solar" OR photovoltaic OR "solar PV")
AND (adoption OR uptake OR diffusion)
AND (urban OR "city households")

• AND: narrows — all concept blocks must appear.
• OR: widens — any synonym within a block qualifies.
• NOT: excludes — e.g. NOT rural (used sparingly, as it can drop relevant records).
  • Truncation and wildcards: adopt* retrieves adopt, adopts, adoption, adopting; wom?n retrieves woman and women.
  • Phrase searching: quotation marks force exact adjacency — "renewable energy" excludes documents where the words appear apart.
  • Field and proximity searching: restrict terms to title/abstract fields, or require terms within n words (e.g. NEAR/3) to raise precision.

C. Selecting databases and sources

The strategy is executed across sources chosen for coverage of the discipline.

• Multidisciplinary indexes: Scopus and Web of Science offer citation data and broad coverage; Google Scholar maximises recall but includes non-vetted material.
• Subject-specific databases: PubMed/MEDLINE (biomedicine), IEEE Xplore (engineering), JSTOR (humanities), ERIC (education), PsycINFO (psychology).
• Grey literature: theses (ProQuest, Shodhganga), conference proceedings, government and institutional reports — needed to counter publication bias, the tendency for positive results to dominate the published record.
• Search tools within databases: thesauri and controlled vocabularies such as MeSH (Medical Subject Headings) map free-text terms to standardised subject tags for consistent retrieval.
• Record management from the start: log each database, the exact query string, the date searched and the hit count, so the search is auditable and repeatable.

D. Screening and selecting relevant literature

Raw results are filtered down to the studies that actually meet the criteria.

• Two-stage screening:
  1. Title and abstract screening: discard clearly irrelevant records against the inclusion criteria — fast, high-volume.
  2. Full-text screening: read the remaining papers in full to confirm eligibility and extract detail.
• De-duplication: remove records retrieved from multiple databases before screening; reference managers automate this.
• PRISMA flow accounting: in systematic reviews, report the numbers at each stage — identified → after duplicates removed → screened → excluded (with reasons) → included — as a transparent audit trail.
• Citation chaining: widen coverage beyond keyword search:
  • Backward chaining (snowballing): mine the reference lists of key papers for older foundational work.
  • Forward chaining: use "cited by" links to find newer work building on a key paper.

E. Critical evaluation and appraisal of sources

Selected sources are judged for quality before their claims are trusted.

• Authority and provenance: check the author's credentials, the publishing journal's peer-review status and its impact factor or quartile ranking.
• CRAAP test: appraise each source on Currency, Relevance, Authority, Accuracy and Purpose.
• Methodological rigour: examine sample size, study design, controls, statistical treatment and whether conclusions are supported by the data presented.
• Bias detection: watch for funding conflicts, selective reporting and unrepresentative samples.
• Distinguishing evidence weight:
  1. Primary empirical study: direct evidence — carries most weight for factual claims.
  2. Secondary review or commentary: useful for context and synthesis but is one step removed from the data.

F. Organising and managing the literature

Appraised sources are arranged so patterns and relationships become visible.

• Reference management tools: Zotero, Mendeley and EndNote store PDFs, auto-generate citations in styles such as APA, IEEE or MLA, and insert in-text references while writing.
• The synthesis matrix: a grid with sources as rows and themes as columns, so agreements, disagreements and gaps across studies are read at a glance.

Source     | Theme: cost | Theme: policy | Theme: awareness
-----------|-------------|---------------|-----------------
Smith 2019 | high barrier| subsidies +   | low
Rao 2021   | falling     | net-metering  | rising

• Thematic vs chronological organisation:
  1. Thematic: group by concept or debate — best for showing the structure of an argument.
  2. Chronological: order by date — best for tracing how a field or method evolved.
• Note-taking discipline: record annotations, direct quotes with page numbers, and one's own critical remarks against each source to prevent misattribution and accidental plagiarism later.

G. Synthesising and writing the review

The organised evidence is woven into a single argued narrative that leads to the research gap.

• Synthesis, not summary: integrate sources around ideas — "three studies link cost to adoption while two emphasise policy" — rather than describing papers one after another.
• Structure of the written review:
  • Introduction: states scope, question and organising principle.
  • Body: organised thematically, each theme presenting agreements, contradictions and trends.
  • Conclusion: names the identified gap and states how the dissertation will address it.
• Identifying the gap: locate the unanswered question, contested finding or untested population that justifies new research — the review's ultimate deliverable.
• Citation integrity: attribute every idea; use direct quotation only for precise, memorable wording and paraphrase accurately elsewhere, always with a citation.
• Maintaining a critical voice: signal evaluation with hedging and contrast — "although", "however", "in contrast", "the evidence remains inconclusive" — so the review argues rather than reports.

III. Types of Review and Tool Support

Matching the procedure's rigour to the review's purpose

The same procedure is applied with varying strictness depending on the review type, and dedicated tools support each stage.

A. Forms of literature review

The depth and formality of the procedure scale with the review's aim.

• Narrative (traditional) review: broad, flexible overview of a topic; useful for background but less reproducible.
• Systematic review: exhaustive, protocol-driven search answering a precise question, following PRISMA; the most rigorous and reproducible form.
• Scoping review: maps the breadth and nature of evidence on a wide topic to identify what exists rather than to answer a narrow question.
• Meta-analysis: statistically pools quantitative results across comparable studies to produce a combined effect estimate — an extension of the systematic review.

B. Software tools across the procedure

Each stage of the search-and-review procedure is supported by dedicated software.

• Search and discovery: database platforms (Scopus, PubMed) and search engines (Google Scholar) with alerts for new publications.
• Screening: Rayyan and Covidence manage blinded title/abstract screening for systematic reviews and track inclusion decisions.
• Reference management: Zotero, Mendeley, EndNote for storage, de-duplication and citation formatting.
• Analysis and visualisation: VOSviewer and CiteSpace produce bibliometric maps of co-citation and keyword co-occurrence, exposing intellectual clusters and emerging trends across a large corpus.
• Qualitative synthesis: NVivo and ATLAS.ti code and organise themes when synthesising qualitative literature.