Last updated 9 min read
Key takeaways
A credible source is one you can trust for the specific claim you are using it to support. That usually means four things are visible: who wrote it and what qualifies them, where the information came from, what checks it passed before publication, and when it was produced. A peer-reviewed article in an established journal shows all four. An anonymous blog post may show none of them.
Credibility is also relative to the claim. A company press release is a credible source for what the company announced and a weak source for whether its product works. A newspaper report is good evidence of what happened at a public event yesterday and thin evidence for a scientific consensus. Before you search, decide what kind of claim you need to support — a fact, a statistic, an expert interpretation or a firsthand account — because each lives in a different place. Once you have a candidate in hand, the companion guide on how to evaluate a source covers the deeper checks.
Different questions need different starting points. The table maps the most common types of credible sources to where you find them, what they are good for, and the trap to watch for with each.
| Source type | Where to find it | Good for | Watch for |
|---|---|---|---|
| Peer-reviewed journal articles | Library databases, Google Scholar, subject databases such as PubMed (biomedicine) or ERIC (education) | Original research findings and scholarly argument | Predatory journals that charge to publish without real review |
| Scholarly books | Library catalogue, university press lists, interlibrary loan | Extended arguments, history, theory | Older editions overtaken by newer research |
| Official statistics and reports | National statistics offices, government agencies, international organizations | Numbers, definitions, policy facts | Definitions that changed between years |
| Subject reference works | Specialist encyclopedias and handbooks in your library | Background, key terms, the main debates | Too general to cite for a specific claim |
| Quality journalism | Outlets with named reporters and published corrections | Recent events and statements from people involved | Opinion pieces sitting alongside reporting |
| Preprints | Preprint servers and university repositories | The newest work in fast-moving fields | Not yet peer reviewed; findings may change |
A good search session is planned, not improvised. This sequence works for most undergraduate and postgraduate projects, and it takes you from a vague topic to a short list of sources you can defend.
Step three is where most time is saved or wasted. If you have never used search operators properly, the guide to Boolean search operators explains them with worked examples, and the guide to using Google Scholar effectively covers its “cited by” and library-access links.
A general search engine ranks pages by signals like relevance and popularity, not scholarly quality, and much of the best research sits behind journal paywalls. Your university or public library pays for databases that index journals, books and reports, and lets you filter to peer-reviewed material in one click. The single search box on most library homepages is a discovery service that covers many of those subscriptions at once, which makes it a sensible first stop.
The library also solves the paywall problem legitimately. When you find an article you cannot open, look for a link to your library's copy, request it through interlibrary loan, check the author's university repository for an accepted manuscript, or email the author, who can usually share their own work. Open-access journals are a sound route too; the Directory of Open Access Journals lists journals that meet its quality criteria.
Citation chaining is the habit that separates a thin bibliography from a strong one. Backward chaining means reading the reference list of a good source and pulling the works it relies on. Forward chaining means finding later work that cites it. Together they show you the conversation a source belongs to — whom it answers, and who answered it.
Look for names and titles that recur. If three recent articles all cite the same older study, that study is probably foundational, and you should read it yourself rather than relying on their summaries. If a well-known study has been heavily criticized since, forward chaining is how you find out before your marker does. Retracted papers are sometimes still cited, so glance at the publisher's page for a retraction notice before you rely on an older finding.
You will still use the open web, especially for current events, organizations and data. Domain endings are weak evidence. Anyone can register a .org address; .edu pages include student projects and personal pages alongside university research; and .gov is reliable for official facts but still reflects the position of the government that published it. Treat the domain as a clue that tells you which questions to ask.
The faster test is lateral reading: before you read a website closely, open new tabs and look up the organization and the author elsewhere. Who funds it? What do independent sources say about it? Professional fact-checkers work this way because a polished site can say anything about itself. Wikipedia helps here as a map — read its article on the organization, then follow the footnotes. On the page itself, look for these signals:
Finding credible sources is half the job; you also have to read them closely enough to cite them accurately. In Cavua's Work space you can upload the PDFs and Word documents you collected and research across several at once, with answers that point to the page they came from. That makes it quick to confirm a source actually says what you plan to cite it for, instead of trusting your memory of an abstract.
You can ask the tutor to quiz you on what each source argues, or talk a difficult one through by voice or video call, within your plan's allowance. If you are learning a subject rather than finishing a single project, every course in the open course catalogue comes with a 24/7 AI tutor that answers from that course's own lessons.
Markers read a bibliography as evidence of how carefully you searched. These are the patterns that undercut it most often. If you are unsure whether a source counts as original research, the guide to primary vs secondary sources explains the difference discipline by discipline.
What an AI research assistant can do for students, what it can't, how to test one before you rely on it, and what to keep out of the files you upload to it.
APA citation in the 7th edition: in-text rules for one, two and three or more authors, plus reference formats for articles, books, chapters and webpages.
Learn how to avoid plagiarism with clear rules for quoting, paraphrasing and citing, examples in APA and MLA, and a checklist to run before you submit.
How to read a research paper in three passes: what to read first, how to decode figures and methods, which questions to ask, and how to take notes that last.
How to use AI for research without being misled: where AI tools help, the five ways they mislead, a verification workflow and how to disclose AI use honestly.
How to write a literature review step by step: set the scope, search and read with a system, group sources by theme, and structure a review that argues.
Wikipedia is a useful starting point but usually not a source to cite in academic work. Anyone can edit it, article quality varies, and it is a tertiary source that summarizes other work. Use it to get oriented and learn key terms, then follow its footnotes to the books, articles and reports it relies on, read those, and cite them instead.
Not automatically. Anyone can register a .org domain, including advocacy groups, businesses and individuals. Some .org sites belong to respected professional bodies and research organizations; others exist to promote a position. Check who runs the site, who funds it, and whether its claims link to evidence you can verify, ideally by looking the organization up on independent sites.
Google Scholar is a search engine, not a source, so its results vary in quality. It lists peer-reviewed articles, books, theses, preprints and conference papers side by side and does not screen out weaker journals. Use it to find material and trace citations, then judge each result on its own merits: who published it, whether it was peer reviewed, and how later work received it.
Your assignment brief is the authority. Without guidance, aim for enough sources to cover the main positions on your question, the key evidence and the strongest objections, rather than chasing a number. A short paper may rest on a handful of strong sources, while a dissertation needs far more. Relevance and quality count for more than the length of the list.
Yes, for the right claims. Reputable outlets are good sources for recent events, public statements and context, especially when the reporter is named and the outlet publishes corrections. They are weaker for scientific or scholarly claims; for those, trace the story back to the original study or report, read it, and cite that instead of the article.
Common examples include anonymous pages with no sources, content written mainly to attract clicks, sites that sell the products they assess, social media posts that cannot be traced to an original, and journals that charge to publish without genuine peer review. Some of these can be studied as evidence of opinion, but none should carry a factual claim in an academic argument.
Every Cavua course comes with a 24/7 tutor you can video call, speak to or chat with — on the material you are actually studying.
