How to Write a Systematic Literature Review: A Thesis Guide

Tesify Team Avatar

·

How to Write a Systematic Literature Review: A Thesis Guide

A systematic literature review (SLR) is fundamentally different from a traditional narrative literature review: where a narrative review synthesises the literature selectively, a systematic review follows a pre-specified, replicable protocol to identify, select, and synthesise all relevant evidence on a research question. Used widely in health sciences, education, psychology, and social policy research, the systematic review is increasingly expected in PhD theses and master’s dissertations in empirically oriented disciplines. This step-by-step guide walks you through the PRISMA-guided process from search strategy to final synthesis.

Quick Answer: A systematic literature review follows these steps: (1) formulate a focused research question using PICO or SPIDER; (2) develop a comprehensive search strategy with Boolean operators; (3) search multiple databases and grey literature; (4) screen titles/abstracts, then full texts against pre-specified inclusion/exclusion criteria; (5) extract data systematically; (6) assess quality using validated tools; (7) synthesise and report using the PRISMA flow diagram. Pre-register your protocol on PROSPERO for maximum credibility.

SLR vs Narrative Review: When to Use Each

Not every thesis requires a systematic review. Use a systematic review when:

  • Your research question asks “what is the evidence for…?” or “what is the effect of X on Y across studies?”
  • You are working in health, social care, education, or policy where evidence synthesis is expected
  • Your supervisor or department explicitly requires it

Use a narrative (traditional) review when:

  • Your research is exploratory or theoretical
  • You are working in humanities or interpretive social sciences
  • The literature is too sparse, heterogeneous, or context-dependent for systematic synthesis

For a worked example of a narrative review, see how to write the review of related literature for a thesis on remote work. This article covers only the systematic variant.

Student screening research articles against inclusion and exclusion criteria for a systematic literature review
Title and abstract screening against pre-specified criteria is the backbone of a systematic review.

Step 1: Formulate Your Review Question Using PICO or SPIDER

PICO (for clinical and health sciences):

  • Population: Who is being studied? (e.g., postgraduate students)
  • Intervention: What is the exposure/intervention? (e.g., structured AI writing tools)
  • Comparison: What is it compared to? (e.g., unassisted writing)
  • Outcome: What is measured? (e.g., thesis completion rates)

SPIDER (for qualitative and mixed methods):

  • Sample, Phenomenon of Interest, Design, Evaluation, Research type

Using a structured framework ensures your search terms are comprehensive and your inclusion criteria logically derived.

Turn each framework element into a concept block. In the PICO example above, the review question becomes: “Among postgraduate students, do structured AI writing tools, compared with unassisted writing, affect thesis completion rates?” Each element then generates its own list of synonyms for the search string, and each element becomes an inclusion criterion. Write this question, the criteria and your planned synthesis method into a short protocol before you run a single search. The protocol is what your supervisor and examiners will compare your final methods section against, so any later deviation must be reported and justified.

Step 2: Develop Your Search Strategy

A systematic search strategy must be documented, reproducible, and comprehensive. The key elements:

  1. Identify all relevant databases: A minimum of three for most topics — typically one general (Web of Science or Scopus), one discipline-specific (PsycINFO for psychology, ERIC for education, MEDLINE for health), and one grey literature source (PROQUEST dissertations, government databases)
  2. Develop your search string: Using Boolean operators (AND, OR, NOT) and truncation (*)
    Example: (thesis* OR dissertation* OR “postgraduate research”) AND (“artificial intelligence” OR “AI” OR “large language model”) AND (“academic writing” OR “academic integrity” OR “plagiarism”)
  3. Test and refine: Run a preliminary search; if you get 50,000 results, narrow; if you get 50, broaden
  4. Document everything: Date of search, database, exact search string, number of results — this goes in your methodology and PRISMA diagram

For advanced database search techniques, see our guide to Google Scholar advanced search. For AI-assisted literature discovery, see best AI tools for literature reviews.

Hand-drawn PRISMA-style flow diagram showing the stages of study selection
Sketching your PRISMA flow diagram early helps you document every screening decision.

Step 3: Screen Studies Systematically

Screening happens in two stages:

  1. Title and abstract screening: Apply your pre-specified inclusion/exclusion criteria to each title/abstract. Document how many are included/excluded and why.
  2. Full-text screening: For included abstracts, obtain the full text and apply criteria again. Document reasons for exclusion.

Inclusion/exclusion criteria must be pre-specified before you screen. Common criteria categories:

  • Publication language (e.g., English only)
  • Date range (e.g., 2015–2026)
  • Study design (e.g., primary empirical studies only, no opinion pieces)
  • Population (e.g., postgraduate students at degree-awarding institutions)
  • Outcome (e.g., must report thesis completion data)

For PhD systematic reviews, dual independent screening (two reviewers screening independently) is usually required for methodological rigour. For master’s theses, single screening with documented decision criteria is typically acceptable — check with your supervisor.

Step 4: Extract Data Systematically

Create a data extraction form before you start — a table capturing the same information from every study. Minimum fields:

  • Author(s), year, country, publication type
  • Study design and sample size
  • Population characteristics
  • Intervention/exposure details
  • Outcome measures and results
  • Key limitations noted by authors

Tools for systematic data extraction: Covidence (paid, the standard for clinical reviews), Rayyan (free, good for title/abstract screening), Elicit (AI-assisted extraction — see AI literature review tools), or Microsoft Excel.

Step 5: Assess Study Quality

Quality assessment tools by study design:

  • Randomised controlled trials: Cochrane Risk of Bias 2.0 (RoB2)
  • Observational studies: ROBINS-I or Newcastle-Ottawa Scale
  • Qualitative studies: CASP Qualitative Checklist
  • Mixed methods: Mixed Methods Appraisal Tool (MMAT)

Report quality assessment in a table. Studies of low quality are not necessarily excluded — they may be included with their limitations noted in your synthesis.

As with screening, appraisal is stronger when a second reviewer independently rates at least a sample of the included studies and disagreements are resolved by discussion. State in your methods chapter which tool you used for which design and how quality judgements fed into the synthesis, for example by running a sensitivity check that drops high-risk studies. If your review informs practice recommendations, consider rating the overall certainty of evidence for each outcome with the GRADE approach, which many health and social care examiners will recognise.

Step 6: Synthesise and Write Up with PRISMA

The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 flow diagram must appear in your thesis to show the screening process:

  • Records identified through database searching
  • Records after duplicates removed
  • Records screened (title/abstract)
  • Records excluded (with reasons)
  • Full-text articles assessed
  • Full-text articles excluded (with reasons)
  • Studies included in synthesis

PRISMA diagrams can be generated free at prisma-statement.org. Pre-register your review protocol on PROSPERO (prospero.york.ac.uk) before starting — this demonstrates rigour and prevents duplication.

For writing up your synthesis, narrative synthesis (grouped by theme or outcome) is appropriate when studies are too heterogeneous for statistical meta-analysis. Meta-analysis (statistical pooling of effect sizes) requires comparable outcome measures across studies and is typically reserved for PhD theses.

A narrative synthesis still needs structure. Group studies by outcome, population or intervention type, summarise each group in a table, describe the direction and consistency of findings, and explain why studies disagree. The SWiM (Synthesis Without Meta-analysis) reporting guideline gives a useful checklist for describing how you grouped studies and what standardised metric you used. Close the chapter with a discussion of strengths and limitations of the review itself, such as language restrictions or databases you could not access, and the implications for your own empirical study.

When writing up all stages of your systematic review, Tesify can help you structure the methodology and results write-up with academic clarity, ensuring your systematic review documentation meets examiner expectations.

Frequently Asked Questions

What is the difference between a systematic and a narrative literature review?

A systematic review follows a pre-specified, transparent, and replicable protocol to identify and synthesise all relevant evidence on a question, minimising bias through documented inclusion/exclusion criteria, comprehensive search strategies, and quality assessment. A narrative review synthesises the literature selectively, guided by the author’s expertise and argument, with no requirement for comprehensiveness. Systematic reviews are required in health sciences and evidence-based disciplines; narrative reviews are standard in humanities and interpretive social sciences.

Do I need to use PRISMA for my thesis systematic review?

Yes, for any thesis claiming to be a systematic review. PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 provides the reporting standard that examiners and journal reviewers expect. The PRISMA flow diagram documenting your screening stages is a required component. If your review is a scoping review rather than a systematic review, use the PRISMA-ScR extension. Download the PRISMA checklist free at prisma-statement.org.

How many studies should be included in a systematic review?

There is no minimum number. Systematic reviews have been published with as few as 3–5 included studies (when the literature is genuinely sparse) and as many as 500+. What matters is comprehensiveness (you have searched systematically and included all eligible studies) and transparency (you have documented why each excluded study did not meet your criteria). For a master’s dissertation systematic review, 8–25 included studies is typical; for a PhD, 20–50+ is more common.

What is PROSPERO and should I register my review?

PROSPERO (prospero.york.ac.uk) is the international prospective register of systematic reviews, run by the Centre for Reviews and Dissemination at the University of York. Pre-registering your systematic review protocol before conducting it demonstrates rigour, prevents inadvertent duplication, and provides a citeable record that your protocol was pre-specified (not adapted post-hoc to fit your results). For health and social care systematic reviews, registration is now standard practice. For master’s students, pre-registration is best practice but not always required — check with your supervisor.

How long does a systematic literature review take for a thesis?

It depends on the breadth of your question, the number of records retrieved and whether you screen alone or with a second reviewer. For most thesis students it takes several months, with screening and data extraction consuming the most time. Build the timeline backwards from your submission date, leave time for updating the search before submission, and agree milestones with your supervisor for the protocol, screening, extraction and write-up.

Can I use AI tools in a systematic literature review?

Yes, as support rather than as a replacement for your judgement. AI tools can help prioritise records during screening, draft extraction tables or tidy your write-up, but you remain responsible for every inclusion decision and every extracted value. Check your university’s AI policy, verify all AI-assisted outputs against the original studies, and report which tools you used and for which stage in your methods section.

Write Up Your Systematic Review with Tesify

From search strategy to PRISMA flow diagram write-up, Tesify provides the structural framework and AI-assisted writing support to produce a rigorous, well-documented systematic review chapter — with auto-bibliography for every included study in your required citation format.

Structure Your Literature Review →

Write your thesis with AI

Structure, draft, cite, and format your thesis faster with Tesify’s AI writing tools, automatic bibliography, and plagiarism checker. Free to start, no credit card required.

Tesify Team Avatar

Leave a Reply

Your email address will not be published. Required fields are marked *