# How to Conduct a Systematic Review > A step-by-step guide for researchers on how to conduct a systematic literature review, from formulating a question and developing a protocol to screening. Source: https://www.alfredscholar.com/blog/how-to-conduct-a-systematic-review/ Published: 2026-10-18 Author: Alfred Scholar Team ## Your Research Needs a Backbone, Not Just a Bookshelf Many researchers treat a literature review as a necessary chore: read a pile of papers, summarize them, find a gap, and move on. That approach is fine for a course paper, but for a thesis or a high-impact publication, you need something more rigorous. You need a systematic review. Unlike a traditional narrative review, a systematic literature review is a form of research in itself. It uses a transparent, methodical, and reproducible process to find, evaluate, and synthesize all the available evidence on a specific question. Think of it less like building a bookshelf and more like conducting an experiment where the existing literature is your dataset. This guide provides a step-by-step walkthrough of how to conduct a systematic literature review, turning a potentially chaotic process into a manageable, high-value research project. ## Step 1: Formulate an Answerable Research Question Everything flows from the question. A vague or overly broad question will lead to a sprawling, unfocused search and an impossible synthesis task. A common problem reviewers encounter is a question that is really a topic in disguise. The most widely used framework for structuring a question, especially in health sciences, is **PICO**: * **P** - Population/Patient/Problem: Who are you studying? * **I** - Intervention or Exposure: What treatment, factor, or program are you investigating? * **C** - Comparison or Control: What is the alternative to the intervention (e.g., a placebo, standard care, no intervention)? * **O** - Outcome: What are you trying to measure, improve, or change? For example, a vague topic like "AI in education" becomes a focused PICO question: "In *university students* (P), does the use of *AI-powered writing assistants* (I) compared to *standard human feedback* (C) affect *final essay grades and writing confidence* (O)?" This structure does more than just clarify your thinking; it becomes the foundation for your search strategy and eligibility criteria. ## Step 2: Develop and Register Your Protocol A systematic review protocol is your roadmap. It explicitly states your research question, search strategy, inclusion and exclusion criteria, and planned methods for data extraction and synthesis. Writing it before you begin is non-negotiable; it prevents "scope creep" and reduces the risk of bias by forcing you to make methodological decisions upfront. Your protocol should detail: * **Rationale:** Why is this review needed? * **Research Question(s):** Your PICO-driven question. * **Inclusion/Exclusion Criteria:** Based on your PICO, what types of studies (e.g., randomized controlled trials only), publication dates, languages, and settings will you include or exclude? * **Search Strategy:** Which databases will you search (e.g., MEDLINE, Scopus, Web of Science)? What keywords and subject headings will you use? * **Study Selection:** How will reviewers screen titles, abstracts, and full texts? The standard is independent screening by two reviewers, with a third to resolve conflicts. * **Data Extraction:** What specific information will you pull from each included study? * **Risk of Bias Assessment:** What tool will you use to appraise the quality of the included studies (e.g., Cochrane RoB 2, ROBINS-I)? * **Synthesis Plan:** How will you synthesize the findings? Will it be a narrative synthesis or a quantitative meta-analysis? Once complete, register your protocol in a public registry like [PROSPERO](https://www.crd.york.ac.uk/PROSPERO/). Registration promotes transparency, prevents unintentional duplication of work, and is increasingly expected by journals. ## Step 3: Conduct the Comprehensive Literature Search This is the most time-consuming phase. The goal is to identify *all* relevant studies, not just the easy-to-find ones. Your search must be exhaustive and meticulously documented. 1. **Select Databases:** Use multiple databases relevant to your field. Relying only on Google Scholar or a single database is a common but critical error. 2. **Build Search Strings:** Translate your PICO components into keywords and database-specific subject headings (like MeSH in MEDLINE). Combine them using Boolean operators (AND, OR, NOT) to build a precise search string for each database. This process is iterative and often benefits from consulting with a subject librarian. 3. **Search "Grey Literature":** Look beyond academic databases for unpublished studies, conference proceedings, dissertations, and clinical trial registries. This helps mitigate publication bias, where studies with positive or significant results are more likely to be published. 4. **Document Everything:** Save every search string exactly as you ran it, along with the date and the number of results from each database. This information is required for your PRISMA flow diagram. Managing the thousands of citations you find requires a robust system. Use a good citation manager and consider a dedicated review platform. A unified research workspace like Alfred Scholar can help keep your library, notes, and manuscript drafts organized from the start, preventing the chaos of juggling multiple tools and files. ## Step 4: Screen Studies and Select for Inclusion With your search results compiled and deduplicated, the screening process begins. This is typically a two-stage process managed by at least two independent reviewers. * **Stage 1: Title and Abstract Screening:** Reviewers read the title and abstract of each record, applying the predefined inclusion and exclusion criteria. Each record is marked as "include," "exclude," or "maybe." * **Stage 2: Full-Text Screening:** For all records marked "include" or "maybe," the reviewers retrieve and read the full-text article to make a final inclusion decision. Disagreements between reviewers are common and should be resolved through discussion or, if necessary, by a third, senior reviewer. Meticulous record-keeping is crucial. You must track how many studies were excluded at each stage and the primary reason for exclusion (e.g., wrong population, wrong intervention, wrong study design). These numbers will populate your PRISMA flow diagram. ## Step 5: Extract Data and Assess Quality Once you have your final set of included studies, you move to data extraction. Using a pre-piloted data extraction form, you and another reviewer will independently pull the relevant information from each paper. Typical data to extract includes: * Study characteristics (author, year, country, study design) * Participant characteristics (sample size, age, demographics) * Intervention and comparison details * Outcome measures and results (e.g., means, standard deviations, p-values) At the same time, you'll assess the methodological quality or risk of bias for each study using the tool you specified in your protocol. This isn't about whether you "like" the study; it's a formal appraisal of its internal validity. The results of this assessment will help you interpret the overall strength of the evidence in your synthesis. ## Step 6: Synthesize, Interpret, and Report the Findings The final step is to bring it all together. How you synthesize the data depends on the nature of your findings. * **Narrative Synthesis:** If the studies are too different in their methods, populations, or outcomes to combine statistically (a common scenario), you will perform a narrative synthesis. This involves grouping studies by theme or outcome, summarizing their findings in text and tables, and identifying patterns or discrepancies across the evidence base. * **Meta-Analysis:** If your included studies are sufficiently similar (homogeneous), you can perform a meta-analysis. This uses statistical methods to combine the results from individual studies, producing a single, more precise estimate of the effect. Your report should present the findings clearly, starting with the PRISMA flow diagram showing the journey from search results to final included studies. Discuss the main findings, the overall quality of the evidence, limitations of the review, and implications for future research and practice. Conducting a systematic review is a marathon, not a sprint. But by following these structured steps, you can produce a high-quality, impactful piece of research that truly contributes to your field, far beyond what a standard [literature review](/blog/how-to-write-a-literature-review-that-proves-novelty/) can achieve.