You have a question, a deadline and a search page full of results. The difficult part is not finding something that looks academic; it is finding evidence you can justify using. This guide shows you how to find peer-reviewed sources, check what you have found and turn a promising result into a useful research note.
The workflow below suits a taught university assignment. A dissertation, systematic review or clinical evidence review may require a more comprehensive search protocol, so follow your supervisor’s instructions where they differ.
How do you find peer-reviewed sources?
Start with your university library’s subject databases, translate the assignment into search concepts, and apply an appropriate peer-reviewed filter. Then check the journal’s review policy and the individual article type. Finally, read the methods, findings and limitations to decide whether the evidence supports the claim you want to make.
The University of Sydney’s peer-review guidance explains why both checks matter: editorials and other items in a peer-reviewed journal may not themselves have undergone peer review. A filter is a useful starting point, not a complete quality assessment.
Start with the evidence your question needs
Before opening a database, write a one-sentence description of the evidence you need. “Sources about remote work” is too broad. “Research comparing how early-career employees learn informally in remote and office settings” gives you something to screen against.
Underline the task’s population, context, period and required source types. If the brief specifies primary empirical research, a textbook or review article may help you understand the topic without meeting that requirement. If it asks for a policy analysis, official documents may be essential alongside research articles.
Use the assignment command-word guide to check whether you are being asked to describe, compare or evaluate. That decision changes what counts as useful evidence.
Build a search you can improve
A search strategy combines concepts deliberately. Sydney’s search-strategy guide recommends defining the scope, selecting terms and databases, running the search and refining it. Keep a short record so you can see which change improved the results.
For our illustrative remote-work question, try these concept groups. They are a starting query, not a tested or exhaustive search:
- Working arrangement: “remote work”, telework, “hybrid work”.
- Population: “early career”, newcomers, graduates.
- Outcome: “informal learning”, mentoring, “knowledge sharing”.
(“remote work” OR telework OR “hybrid work”) AND (“early career” OR newcomers OR graduates) AND (“informal learning” OR mentoring OR “knowledge sharing”)
Check the database’s help before pasting this query: phrase marks, truncation and other syntax can differ. If results are scarce, remove the population group temporarily and screen it manually. If results are overwhelmingly broad, restrict a concept to the title or abstract. Change one thing at a time and note what happened.
Check the journal and the article separately
- Open the publisher’s article record rather than relying on a search snippet or an isolated PDF.
- Confirm the title, authors, journal and publication date match the document you are reading.
- Find the journal’s stated peer-review process. Your library may also provide a journal directory.
- Check the item’s label: research article, review, editorial, letter or another publication type.
- Look for linked corrections, expressions of concern or retraction notices before using its claims.
A DOI identifies an object; it does not certify research quality. Similarly, a polished website, a familiar institution in the author list or a high citation count does not remove the need to read critically. Record uncertain status as uncertain and ask a librarian rather than guessing.
Read for usefulness, not just credibility
A credible study can still be unsuitable for your question. The University of Leeds recommends evaluating relevance, expertise, currency and evidence, then examining how the author’s argument reaches its conclusion.
Apply that distinction to our example. A study about experienced software developers may inform your discussion, but it does not directly establish what new graduates experience. A survey of perceived learning measures something different from a test of retained skills. A short-term study may not show what happens after onboarding ends.
Write the boundary next to the finding: “Useful for perceived access to colleagues; does not measure skill development.” This is more valuable than simply labelling the source “good”. It tells you exactly how far your paragraph can go.
Keep a compact source log
Use a spreadsheet, reference manager or ordinary document. The following is our suggested working record; adapt it to your discipline:
- Identity: complete reference, DOI or stable link, and saved filename.
- Discovery: database, query and date searched.
- Selection: why the source belongs in this assignment.
- Evidence: the relevant finding in your own words, with its page or section.
- Boundary: the main limitation for your intended claim.
- Use: the paragraph or theme it might support or challenge.
Mark copied wording as a quotation immediately. Keep your interpretation in a separate field so that it cannot later be mistaken for the author’s conclusion. If you are comparing several sources, expand the log into a literature-review matrix.
Troubleshoot the common dead ends
Everything is behind a paywall
Try your university’s access link, an author’s institutional repository or the library’s document-delivery service. Ask library staff what is available before paying personally. Do not treat an abstract as a substitute for the full methods and results when those details are central to your argument.
The search returns hundreds of irrelevant papers
Inspect a few results before adding filters. Perhaps one keyword has another disciplinary meaning, or your outcome is too vague. Replace the ambiguous term, add a meaningful phrase or choose a more suitable subject database. An arbitrary date restriction may hide useful material without fixing the actual problem.
Every result supports the same view
Search for competing explanations, alternative measures and different populations. Check whether several papers analyse the same dataset. Five publications are not automatically five independent pieces of evidence. Make room for relevant disagreement without manufacturing a false balance.
Frequently asked questions
Is everything in Google Scholar peer reviewed?
No. Check the publication and article type instead of treating the search engine as a quality label. Use the publisher’s record and your library’s guidance to establish what you are looking at.
How many peer-reviewed sources do I need?
Follow the brief’s requirement. Where none is given, justify coverage through the claims and perspectives your answer needs, not a universal sources-per-thousand-words rule. Ask your tutor if the expected scope is unclear.
Can I use an older source?
Yes, when it is relevant—for example, an original theoretical work. Explain its role and check whether later evidence changes the claim. Publication age alone does not tell you whether a source remains appropriate.
Need help understanding the next step?
Ask about tutoring or feedback on your own draft. Tell us what your course permits; you remain responsible for the research, reasoning and submitted work.
Discuss study support