Step-by-Step Guide to Conducting a Systematic Literature Review for Your Thesis

A systematic literature review is one of the most demanding pieces of methodological work in graduate research, and also one of the most frequently underestimated. Unlike a traditional narrative literature review, a systematic review follows an explicit, reproducible protocol, one that another researcher could, in theory, follow and arrive at broadly the same set of included studies. That reproducibility is exactly what makes it rigorous, and exactly what makes it time-consuming to do well.

This guide walks through how to conduct a systematic literature review from initial planning through final synthesis, covering the PRISMA guidelines most committees now expect, along with the practical decisions that trip up first-time reviewers.

Systematic review vs. traditional literature review: knowing which you need

Before starting, confirm with your supervisor which type of review your thesis actually requires, since the terms are often used loosely. A traditional, or narrative, literature review synthesizes relevant literature to build an argument and identify a gap, without a formal, reproducible search and screening protocol. A systematic review is a pre-specified, documented procedure for identifying, selecting, and synthesizing the literature on a well-defined question in an effort to include all relevant studies. A scoping review uses systematic search methods but aims to map the breadth of literature on a topic, rather than answering a narrow, focused question.

A systematic review methodology is typically required in health sciences, evidence-based practice fields, and increasingly in social sciences, education, and other disciplines where reproducibility is valued. If your field or program does not require this level of rigor, a well-structured narrative review may be entirely appropriate, and attempting a full systematic review when it is not required can waste significant time.

Step 1: Formulate a focused, answerable research question

Every systematic review starts with a tightly defined question, since a vague question makes every subsequent step, search strategy, inclusion criteria, and synthesis, unmanageable. Most systematic reviews use a structured framework to build this question, with PICO being the most common in health and clinical research: the Population being studied, the Intervention or exposure being examined, the Comparison it is being measured against where applicable, and the Outcome or effect being measured. Other frameworks exist for different fields, such as SPIDER for qualitative reviews, covering sample, phenomenon of interest, design, evaluation, and research type, or PEO for questions without a comparison group, covering population, exposure, and outcome. Choose the framework that matches your question type, and use it to keep your search strategy focused throughout.

Step 2: Register your protocol before you begin

For many systematic reviews, particularly in health sciences, pre-registering your protocol is expected practice, and increasingly required by journals and some graduate programs. Registering with PROSPERO, the international registry for systematic review protocols, establishes a public, dated record of your planned methodology before you begin screening studies.

Registration matters for two reasons: it protects against accusations of post-hoc cherry-picking, since your methods are documented before you have seen the results, and it reduces the chance that someone else is conducting an identical review simultaneously. Even where formal registration is not required, writing out a detailed protocol before starting your search, and sharing it with your supervisor for sign-off, is strong practice that will save you significant rework later.

Step 3: Develop and document your search strategy

This is where systematic review methodology diverges most sharply from narrative reviews. Your search strategy needs to be reproducible, meaning another researcher could run the same search and get the same results. A thorough search strategy typically includes database selection, searching multiple databases such as PubMed, Scopus, Web of Science, PsycINFO, or CINAHL, depending on your field, rather than relying on a single source, since no single database indexes all relevant literature. It also requires carefully developed search terms and Boolean logic, using keywords and controlled vocabulary, such as MeSH terms in medical databases, connected with Boolean operators. Document the exact search string used in each database, since minor syntax differences between databases mean your search terms often need adjustment for each platform.

Decide and justify any limits on publication date or language, since overly restrictive limits can introduce bias, while no limits at all may be impractical for large literature bases. Depending on your field, consider searching conference proceedings, dissertation databases, or hand-searching the reference lists of key included studies to catch relevant work that database searches miss. Save your full search strategy, including exact strings, dates searched, and number of results per database, in a supplementary document or appendix. This documentation is what makes your review reproducible and is typically required for PRISMA compliance.

Step 4: Establish clear inclusion and exclusion criteria

Define your inclusion and exclusion criteria before you begin screening, not partway through, since criteria developed mid-screening risk being shaped by the results you are seeing, which undermines the review's objectivity. Common criteria categories include study design, such as randomized controlled trials only or a broader range of designs, population characteristics, publication date range, language, peer-reviewed status, and outcome measures reported. Write these criteria into your protocol with enough specificity that a second researcher applying them to the same set of studies would make largely the same inclusion decisions you would.

Step 5: Screen studies using the PRISMA process

PRISMA, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses, is the most widely adopted reporting standard for systematic reviews, and most committees expect PRISMA compliance even when a formal meta-analysis is not being conducted. The usual screening process runs through identification, the total number of records identified from all databases plus any additional records identified through other sources, duplicate removal, typically using reference management software, title and abstract screening against your inclusion and exclusion criteria, full-text screening of the remaining records, and final inclusion of the studies that survive full-text screening.

Best practice, and often a program requirement, is to have at least two independent reviewers conduct screening, particularly at the full-text stage, with a documented process for resolving disagreements. If a second reviewer is not available for your thesis, discuss with your supervisor how to address this limitation, since single-reviewer screening is a common and reportable limitation in student systematic reviews.

Step 6: Evaluate the quality of included studies

Once you have your final set of included studies, evaluate the methodological quality of each study using a tool relevant to your study designs. Available options include the Cochrane risk of bias tool for randomized trials, the Newcastle-Ottawa Scale for cohort and case control studies, CASP checklists for different qualitative and quantitative designs, and the Mixed Methods Appraisal Tool for reviews combining quantitative, qualitative, and mixed methods studies. Present the results of quality assessment systematically, usually in a summary table, and discuss how variation in quality of studies might affect the confidence in your overall findings. A common issue examiners find in students' systematic reviews is skipping quality assessment, or doing it in a superficial way.

Step 7: Extract data systematically

Develop a standardized data extraction form before you begin extracting data from your included studies, capturing consistent information across every study: author, year, country, sample size, population characteristics, methodology, key findings, and any other variables relevant to your specific research question. Using a consistent extraction form, rather than pulling different details from each paper, is what makes synthesis across studies possible.

Step 8: Synthesize your findings

How you synthesize your findings depends on the nature of your included studies. Meta-analysis is appropriate when included studies are sufficiently similar in design and outcome measures to statistically pool their results, producing a combined effect size. Narrative synthesis is used when studies are too heterogeneous for statistical pooling, organizing findings thematically or by outcome and discussing patterns, consistencies, and contradictions across studies in prose. Thematic synthesis is common in reviews including qualitative studies, identifying recurring themes across included studies' findings. No matter what approach you take, your synthesis should go beyond sequentially summarizing each study. It should be able to recognize patterns in the body of evidence, account for contradictions where they exist, and evaluate the overall strength and consistency of the evidence base.

Writing the literature review chapter: structure and reporting

When writing a literature review chapter based on systematic methodology, most committees expect an introduction and rationale establishing why a systematic review was appropriate for your research question, a methodology section detailing your search strategy, inclusion and exclusion criteria, screening process, and quality assessment approach, a PRISMA flow diagram showing the full screening pipeline from identification to final inclusion, a results section presenting characteristics of included studies, often in a summary table, and your quality assessment findings, a synthesis organized thematically or by outcome, and a discussion of limitations addressing gaps in the evidence base, limitations of your search strategy, and any constraints in your screening process, such as single-reviewer screening.

Common mistakes in student systematic reviews

Failing to document exact search strings, databases, and dates searched makes it impossible for anyone, including your own future self, to verify or replicate your search. Changing inclusion or exclusion criteria partway through screening, without documenting and justifying the change, undermines the review's claimed rigor. A results section that walks through included studies one by one, without identifying cross-study patterns or contradictions, is a narrative review dressed up as a systematic one, genuine synthesis integrates findings into a coherent picture. And quality assessment is not optional in a rigorous systematic review, omitting it, or including it without meaningfully discussing how quality variation affects your conclusions, weakens the chapter significantly.

Rigor takes time, and that is the point

A well-executed systematic review is genuinely time-intensive, often taking months rather than weeks, and that investment is exactly what gives it credibility. Committees and examiners can tell the difference between a review that followed a documented, reproducible process and one that retrofitted a PRISMA diagram onto an essentially narrative approach. Building in adequate time for search strategy development, careful screening, and genuine synthesis from the start will save you far more time than it costs.