Skip to main content
Guides

Are Blogs Credible Sources? 7 Checks and How to Cite Them

Karan Singh BhakuniUpdated
  • 21 min read
  • 11 sections
  • 8 comparison tables
  • 11 questions answered

The short answer

Blogs can be credible sources, but most count as popular rather than scholarly sources, and the format alone tells you little about whether a particular claim is reliable. Evaluate the exact post, its evidence and your intended use. A firsthand account can document one person’s experience; it does not automatically establish what happens to everyone. A company can accurately explain its own product while still having an interest in selling it.

Want the site rather than the reading? Try Grigora free for 14 days

Contents11 sections

For academic work, also follow the assignment’s source requirements. A useful blog may help you find an original study without satisfying a requirement for peer-reviewed research. The University of Cambridge’s research integrity guidance on citing blogs notes that blogs are typically not peer reviewed, and that where the author has published the same ideas in a peer-reviewed format, you may be able to cite that version instead. A blog can also be the material you analyze when your question concerns what its author said. Credibility, suitability and permission to cite are related but separate decisions.

Are blogs really credible sources? Evaluate the claim, author and evidence before deciding.

Read nextBlog vs Article: The Difference Between Them, With Examples

Are blogs scholarly or peer-reviewed sources?

Usually not. University library guides class blogs as popular sources, written for a general audience, rather than scholarly ones. The University of North Carolina’s guide to source types lists bloggers among the authors of popular sources, and UC Merced Library tells students to treat all blogs as popular rather than scholarly sources. A post by an academic is still a blog post unless it went through a reviewed channel: a blog post is not the same as a journal article, even when both describe the same research.

Source typeReviewed before publication byCounts as scholarly?Best use
Peer-reviewed journal articleIndependent experts in the fieldYesEvidence for research findings
Preprint or working paperNo formal peer review yetUsually not; check your assignmentEarly findings, stated as preliminary
News or magazine articleEditors, and at many outlets fact checkersNo (popular)Recent events, background and quotes
Organization or company blogThe organization’s own staff, if anyoneNo (popular)Its own statements, data and policies
Personal blogUsually no oneNo (popular)A firsthand account, or the blog itself as the object of study

Start with the claim and the purpose

Write the claim you want to support in one sentence before judging the website. “This person described a difficult commute” needs different evidence from “this transport policy reduces congestion.” Reading a whole site as either trustworthy or untrustworthy can hide that distinction. A strong post may contain a weak statistic, and a biased source may still provide a verifiable original document.

Your purposeWhat a blog may provideWhat still needs checking
Understand a topicAn accessible explanation and links to useful evidence.Whether the explanation represents the underlying work accurately.
Describe someone’s experienceA firsthand account of what the author reports experiencing.Authenticity, context and the limits of a single account.
State a company’s current policyThe company’s own announcement or documentation.Date, current terms, product/market scope and whether the policy has changed.
Make a general factual claimA lead to data, reporting or research.The original evidence, methods, qualifications and independent coverage.
Meet an academic requirementA source or object of analysis, if the assignment permits it.The actual required source type and citation rules; ask your instructor or librarian when unclear.

The Rhode Island College library’s blog-evaluation guide recognizes that both edited and amateur blogs can have research uses, depending on evaluation. In particular, an author without traditional expert credentials may have a valuable firsthand perspective. That does not turn the account into evidence for an unrelated technical conclusion.

Which kinds of blogs are more credible?

The publisher and the author’s track record change what a post can support. Wikipedia applies a strict public standard to its own sourcing. Its verifiability policy accepts blogs run by newspapers and magazines when the writers are professionals, while warning that such posts may skip the outlet’s normal fact-checking. Its policy on self-published sources accepts a personal blog only from an established expert whose work in the relevant field has previously been published by reliable, independent publications. UC Merced Library makes a similar split: blogs on organizational websites have some research credibility, while personal blogs may have little.

Blog typeEditorial controlWhat it can supportCheck before relying on it
News outlet blogThe outlet’s editors, though blog posts may get lighter fact-checkingReporting and named analysisIs it news or opinion? Does the outlet publish corrections?
Organization or company blogThe organization itselfIts own policies, announcements, data and documentationThe date, the scope and any commercial or political interest
Expert’s own blogUsually none beyond the authorAnalysis inside the author’s fieldEarlier work in the field published by reliable, independent outlets; claims outside that field
Personal blogNoneWhat the author says they experienced or thinkAuthenticity, context and the limits of one account
AI content farmNone or minimalNothing you should citeNo named author, undisclosed AI text, copied stories

For that last category, NewsGuard lists a site as an AI content farm when a substantial share of its content is produced by AI, it is published without significant human oversight, it is presented so that a reader could assume human journalists wrote it, and it does not disclose the AI. As of the June 23, 2026 update of its AI Tracking Center, NewsGuard had identified 3,749 such sites in 16 languages.

Blog types and credibility: what news, organization, expert, personal and AI content farm blogs can support, and what to check before citing each.
The publisher and the author’s track record change what a post can support.

Use seven checks without turning them into a score

These checks are prompts for investigation. They are not a validated rating system, and passing six does not cancel a serious problem in the seventh. Missing evidence for the central claim matters more than a polished author biography or attractive design.

CheckWhat to inspectWhat it helps you decide
1. Relevant author and publisherWho wrote it, their role, relevant experience and who controls the site. Verify important credentials outside the page. For an expert’s own blog, look for earlier work in the same field published by reliable, independent publishers.Whether the author is positioned to know this particular subject.
2. Original evidenceOpen the cited study, document, dataset or reporting. Find the passage supporting the claim.Whether the evidence exists and actually supports the statement.
3. Context and limitsDefinitions, methods, sample, units, geography, uncertainty and what the evidence did not examine.Whether the conclusion goes further than its evidence.
4. Timing and versionPublication/update dates, dates of underlying evidence and any newer official information.Whether the information fits the period and version you need.
5. Interests and transparencyFunding, sponsorship, affiliate relationships, ownership and stated purpose. In the US, the FTC expects bloggers to disclose payments, free products given for promotion and affiliate commissions clearly and close to the recommendation.What incentives or omissions may affect the presentation.
6. Editorial accountabilityNamed review where relevant, correction process, contact route and response to known errors.How errors can be identified and corrected; a policy alone is not proof of accuracy.
7. Independent corroborationOther reliable coverage that has not simply copied the same claim or press release.Whether separate evidence supports, disputes or qualifies the claim.

An old post can be appropriate evidence about a historical event. A newly updated post can still repeat an old price, law or study incorrectly. Look at what changed and when the underlying information was produced; the date in the page header is only the start of that check.

Relevant expertise matters more than a generic title. A professional can be knowledgeable in one area and outside their competence in another. An anonymous author is harder to verify, but anonymity alone does not prove a claim false. Where the evidence cannot be checked, narrow what you use the source for or find a better source.

CRAAP, SIFT or claim-level checks?

Many library guides teach the CRAAP test, which librarian Sarah Blakeslee created at California State University, Chico’s Meriam Library in 2004. It is a useful memory aid, and it works best alongside checks that take you off the page.

MethodWhat it asksWhere it helpsWhere it falls short
CRAAP test (Sarah Blakeslee, Meriam Library, CSU Chico, 2004)Currency, relevance, authority, accuracy and purposeA memorable checklist for one sourceApplied only on the page itself, it can reward a polished site
SIFT (Mike Caulfield, 2019)Stop, investigate the source, find better coverage, trace claims to the original contextFast checks off the page before you share or citeNeeds good searching and access to the originals
The seven checks aboveAuthor, evidence, limits, timing, interests, accountability and corroboration for one claimDeciding what a particular claim can supportSlower, and deliberately not a score

Leave the page and trace the evidence

Do some checking outside the article itself. Search for the author or organization, open better coverage and follow the claim to its original context. Mike Caulfield’s SIFT framework describes four moves: stop, investigate the source, find better coverage and trace claims, quotes or media to their original context. The point is to recover context before accepting or sharing a claim.

Leaving the page is not just a tidy theory. In a study published in Teachers College Record in 2019, Sam Wineburg and Sarah McGrew asked 10 professional fact checkers, 10 historians with PhDs and 25 Stanford undergraduates to evaluate live websites. The fact checkers left unfamiliar sites early, opened new tabs to see what others said about the source and reached better-supported conclusions in less time, while the historians and students were more often swayed by features such as logos and domain names (Wineburg and McGrew, 2019). That was a small, purposely chosen group, not a population estimate. A separate 2019 Stanford History Education Group assessment of 3,446 US high school students points the same way: 52% judged an anonymous video to be strong evidence of US voter fraud, and only three students found where the footage came from.

Tracing a claim usually involves these steps:

  • Locate the source: Follow a citation to the actual report, not merely another blog repeating the same sentence. If the link is broken, search the cited title and publisher; do not pretend you read an unavailable document.
  • Match the wording: Find what the original says. Check whether “associated with” became “causes,” a small trial became a universal result or an estimate became a measured fact.
  • Check the scope: Record the population, country, product version, date range or conditions. An introductory offer is not necessarily a renewal price.
  • Look for independent coverage: Several sites copying one press release are several repetitions of one source, not several independent confirmations.
  • Keep the limit visible: If the original is preliminary, disputed, unavailable or narrower than the claim, preserve that qualification or leave the claim out.
Blog claim check path: name the claim, open the original evidence, check scope and date, compare independent coverage, then use it with limits, mapped to SIFT.
Evaluate the evidence for the particular statement; a site-wide trust label cannot replace this check.

For an image or quotation, recover the surrounding context too. A real photograph may have an incorrect date or caption. A correctly quoted sentence may omit the author’s next sentence qualifying it. Cite the source you actually examined and avoid presenting a repost as if it were the original.

A worked example: a persuasive percentage

The following is a hypothetical teaching example, not a real study or a claim about a named publisher. A blog headline says: “80% of people succeed with Method X.” Its reference turns out to be a voluntary feedback survey of 50 customers who finished a paid workshop; 40 said they found one exercise useful.

QuestionWhat you findHow that changes your use
What was measured?Finding an exercise useful, rather than achieving the promised outcome.Do not describe the survey as evidence of overall success.
Who was counted?50 responding workshop customers, not a representative sample of all users.Do not generalize the result to everyone trying the method.
Where did the number come from?40 / 50 = 80%.The arithmetic can be correct while the headline’s interpretation is unsupported.
Who produced it?The workshop seller collected the feedback.Disclose the commercial context and look for independent evidence.
What can you honestly write?The seller reports that 40 of 50 respondents found one exercise useful, subject to the survey’s limits.Use that narrower statement only if it is relevant and you can inspect the actual source.

The decision is not “the whole blog is fake” or “the statistic has a citation, so it is proven.” You have found a specific mismatch between evidence and conclusion. You can exclude the claim, investigate further or describe the limited finding accurately. If the survey itself is inaccessible, say that rather than treating this hypothetical source as verified.

Apply stronger scrutiny to consequential decisions

For health, legal, financial or safety decisions, a blog should not be your only basis for acting. Seek relevant current authoritative evidence and qualified help for your circumstances. A clear general explanation can be useful without being suitable advice for a particular patient, jurisdiction or investment decision.

For health information specifically, the National Library of Medicine’s evaluation checklist asks about the provider, funding, information quality and privacy. It includes checking where information comes from, whether experts review it and whether the site makes unbelievable claims. Those are useful questions; a medical-looking design is not a substitute for the answers.

If a post summarizes research, try to find the actual paper and understand what kind of work it is. Do not assume that every linked paper was peer reviewed, that peer review eliminates errors or that one study settles a complex question. If you cannot assess the evidence yourself, use an appropriate expert explanation and retain its stated uncertainty.

Also worth readingHow to Start a Science Blog in 2026: 8 Steps for Beginners

What does not prove that a blog is credible

  • A .com, .org, .edu or .gov label alone: The publisher and context matter, but a suffix does not establish that every statement is correct or suitable for your purpose.
  • Domain age, traffic, search position or an SEO score: These do not verify factual accuracy. An SEO spam score is not a fact-checking result or a measure of academic reliability.
  • Professional design, HTTPS or fluent writing: These can make a page easier or safer to use in particular ways, but do not establish the truth of its content.
  • Fluent, generic text with no named author: Many sites now publish machine-written posts with little human oversight, and smooth prose says nothing about whether anyone checked them. Look for a named, verifiable author and original sourcing rather than polish.
  • A long reference list: Citations may be irrelevant, outdated or misrepresented. Check the reference supporting the claim you intend to use.
  • A recognizable brand or expert title: Inspect the actual evidence and relevant competence. Company ownership can be relevant even when its documentation is accurate.
  • A numerical trust checklist: Counting boxes cannot turn an unsupported claim into supported evidence. State the reason for your decision instead.

Bias also needs a precise meaning. A viewpoint or commercial interest deserves disclosure and scrutiny; it does not automatically invalidate every factual statement. Conversely, a neutral tone can hide selective evidence. Separate the author’s interpretation from observable facts, and ask what important contrary evidence is missing.

Record your decision and cite what you actually used

Keep a short source record while researching so you can explain your choice and find the evidence again. This is especially useful when a live page changes, a price expires or a post revises its wording.

RecordExample of what to save
Claim and intended useThe exact statement you want to support and the question it answers.
Source detailsAuthor, article title, publisher, URL and the dates shown on the page.
Evidence locationThe supporting section, original document and relevant page/table or passage.
Checks performedWhat you opened, who produced it and whether other coverage was independent.
Decision and limitsUse for a firsthand account only; use with a stated qualification; investigate further; or exclude.
Access and versionWhen you checked it and any archived version you actually consulted, where appropriate.

The archived version matters because cited pages disappear. In a report published in May 2024, Pew Research Center found that when it checked a large sample of web pages in October 2023, 38% of the pages that existed in 2013 were no longer accessible. Across all the pages it had collected from 2013 to 2023, 25% were gone. It also found that 54% of Wikipedia pages had at least one broken link in their references section, and that 23% of news webpages and 21% of government webpages contained at least one broken link. Pew counted a page as gone only when it returned a definitive error code, so these figures are conservative. Academic citations decay too: Zittrain, Albert and Lessig (2014) found that more than 70% of the URLs in the Harvard Law Review and other journals, and 50% of the URLs in US Supreme Court opinions, no longer led to the cited material. When you rely on a blog post, save a copy with the Internet Archive’s Save Page Now or a library-supported service such as Perma.cc if your institution offers it, and record the date you checked the page.

Link rot statistics: 38% of 2013 pages gone by 2023, 54% of Wikipedia pages with a dead reference link, 50% of Supreme Court URLs no longer showing the cited material.
Save an archived copy and the date you checked any blog post you cite.

How to cite a blog post in APA, MLA and Chicago

StyleReference list entry or noteIn the text
APA 7th editionAuthor, A. A. (Year, Month Day). Title of post. Blog Name. URL(Author, Year)
MLA 9th editionLast, First. “Title of Post.” Blog Name, Publisher if different, Day Mon. Year, URL.(Last), or name the author in your sentence
Chicago 18th edition, notes and bibliographyNote: First Last, “Title of Post,” Blog Name (blog), Month Day, Year, URL.A numbered note. Bibliography: Last, First. “Title of Post.” Blog Name (blog). Month Day, Year. URL.

These templates follow the APA Style blog post examples, the University of Lethbridge’s MLA 9 blog guide and the University of Queensland’s Chicago 18 blog guide, which notes that “(blog)” may be added after the blog name. Check the edition your assignment names. If the page shows no author or date, do not invent one; follow your style’s rule for the missing element. Cite the underlying sources you use in the same style. If the blog helped you discover a report, read the report before citing it as evidence. If you quote or analyze the blog’s own wording, cite the blog itself.

Ask your instructor or librarian about ambiguous source requirements. “No blogs” is not an invitation to relabel a blog as a journal article; a peer-reviewed-source requirement still applies. An instructor may allow a blog as an object of analysis while requiring scholarly research to support the rest of the argument.

If you publish a blog, make it easy to verify

Readers, instructors and editors can only check what you make visible. Whether you write a class, research, journalism or expert blog, these habits make each post easier to verify:

  • Name the author: Give a byline and the experience that is relevant to the topic, not just a job title.
  • Link each important claim to its original source: Put the link beside the claim rather than in a list at the end.
  • Show the dates: Display the published date and, after a real change, the updated date and what changed. Article markup from the free article schema generator can state the author and dates for search engines, but it describes the page and does not make its claims true.
  • Disclose free products, sponsorships and affiliate commissions: The FTC’s endorsement guidance for US bloggers asks for disclosures that are hard to miss and placed close to the recommendation.
  • Label opinion, predictions and hypothetical examples: Readers should be able to tell them apart from reported facts.
  • Correct errors visibly: Note what changed and keep a contact route open so readers can report problems.
  • Recheck old posts for dead source links: The free broken link checker finds cited pages that no longer load, and the outbound link checker lists each outbound link with its status code.

The distinction between a personal blog and microblog describes format and use, not an automatic credibility grade. If you are setting up a class, research or journalism blog, the guides to creating a blog for students, starting a science blog and starting a journalism blog cover sourcing in more detail.

If you publish your own blog, Grigora’s AI website builder drafts the site from a short description, and you approve visitor comments before they appear on a post. You can try it with a 14-day free trial, no card needed.

Blog source credibility FAQs

Get the next guide by email

Practical writing on building a site, growing an audience and sending email people open. No more than one a week.

Karan Singh Bhakuni

120 articles on the Grigora blog

More on Guides