How to Evaluate Sources for Academic Research
Evaluating a source is a procedure, not a vibe. Ask the official CRAAP questions from Meriam Library at California State University, Chico — Currency, Relevance, Authority, Accuracy, Purpose — then open a new tab and check what other people say about the author or publisher before you lock the answers. This URL is the letter-by-letter method. For the accept / caution / reject verdict those answers feed, use How to tell if a source is credible.
By CitationEasy Editorial Team24 min read
On this page
How to Evaluate Sources: Quick Answer
Stay on the item long enough to read a date, a name, a methods note, and a purpose. Then search the author or organization minus the site you are on. A checklist you run only inside the page is how students bless a professional-looking front group. A search you never bring back to the methods is how students cite a genuine article that is too old, off-topic, or retracted.
Fill from the source in front of you
Keep for this claim:
I can name the accountable person or office, a publication or revision date that fits this claim, and a method, dataset, or locator I can open. I will cite this source for [this claim], not for a claim it does not support.
Keep only as background:
The publisher is real, but the item is a summary, a superseded wave, a consumer page, or off-audience for this paper. I will name that limit, or I will open the underlying table or article and cite that instead.
Do not cite for this claim:
I cannot name who is accountable, the purpose is to sell or recruit, the item has been retracted, or a sixty-second search shows the publisher is not what the About page pretends.
Those three lines are a verdict after the questions, not a substitute for them. Run the official CRAAP list below. Take the verdict to How to tell if a source is credible if you still need the three-way decision in one place.
What this page is for
CitationEasy already has a page for the decision (Is this source credible enough to cite?), a page for claim-checking (How to fact-check sources), and a page for source types (Scholarly vs popular sources). This URL is the evaluation method: the full CRAAP question lists, the failure modes each letter catches, and five real sources run all the way through. If you only need a yes / no / not-for-this-claim verdict, start on the credibility page and come back here when you need the questions.
The Association of College and Research Libraries treats evaluation as part of inquiry, not as a one-time stamp. Filed by the ACRL Board on 2 February 2015 and adopted on 11 January 2016, the Framework still names six frames. The one this walkthrough starts from is Research as Inquiry:
ACRL Framework for Information Literacy for Higher Education (2016)
Research as Inquiry
Research is iterative and depends upon asking increasingly complex or new questions whose answers in turn develop additional questions or lines of inquiry in any field.
That is why a source can pass four letters and still fail the fifth, and why you re-evaluate when the research question changes. A Bureau of Labor Statistics monthly release is excellent for last month's U.S. payroll picture and useless for a literary reading of Moby-Dick.
The official CRAAP test (Meriam Library)
Librarian Sarah Blakeslee published “The CRAAP Test” in LOEX Quarterly 31, no. 3 (Fall 2004). The later library handout is the wording this page treats as official. Evaluating Information — Applying the CRAAP Test (dated 9/17/10, CC BY 4.0) is the Meriam Library, California State University, Chico, list. Cornell University Library restates CRAAP for health pages — check a last-updated date, try the links, name the owner — but the questions below are Chico's, not a campus rewrite. The handout opens with the problem, not a score:
Meriam Library, California State University, Chico
When you search for information, you're going to find lots of it . . . but is it good information? You will have to determine that for yourself, and the CRAAP Test can help. The CRAAP Test is a list of questions to help you evaluate the information you find. Different criteria will be more or less important depending on your situation or need.
CRAAP is a list of questions, not a points system. The handout never tells you to add a score and keep anything above 15. A source can ace Currency and still be the wrong audience, or be written by a real expert selling a product. The five official glosses below are the library's words. The question lists under each heading are the library's questions. Two of them — functional links, and whether the URL reveals anything (.com, .edu, .gov, .org, .net) — are marked on the handout as especially useful for the web.
| Letter | Official gloss |
|---|---|
| Currency | The timeliness of the information. |
| Relevance | The importance of the information for your needs. |
| Authority | The source of the information. |
| Accuracy | The reliability, truthfulness and correctness of the content. |
| Purpose | The reason the information exists. |
Currency: is this recent enough for this claim?
Currency is not “newer is always better.” The official question is whether your topic needs current information. An August 2026 labor release is stale the month after the next one lands. Thomas S. Kuhn's The Structure of Scientific Revolutions (University of Chicago Press, 1962) is still the book a history-of-science seminar assigns. Currency also includes revisions: the first-print BLS July 2026 payroll change was later revised, so a paper that froze the first headline is already citing a superseded figure.
Official Currency questions (Meriam Library)
Official gloss
Currency: The timeliness of the information.
- When was the information published or posted?
- Has the information been revised or updated?
- Does your topic require current information, or will older sources work as well?
- Are the links functional?
Typical assignment sheets — this is instructor habit, not a CRAAP score — want recent work for clinical or technology claims, and they still want foundational texts in philosophy, history, and theory. If the page you opened has already been superseded, the page itself often says so. Americans' Social Media Use 2025 (Pew Research Center, 20 November 2025) reports a 5 February–18 June 2025 survey. The 31 January 2024 report is the previous wave, and that page now points readers to the 2025 release. That pairing is the Currency answer: use the 2025 release for “what share of adults use TikTok now,” and keep the 2024 report only if you are writing about that specific survey.
Relevance: would you cite this in the paper you are writing?
A source can be famous, recent, and still useless for your question. Pew's 2025 platform-use tables are the right evidence for “which sites U.S. adults report using.” They are the wrong evidence for teen use (Pew publishes a separate teen survey) and the wrong evidence for whether Instagram causes loneliness. Relevance is fit, not prestige.
Official Relevance questions (Meriam Library)
Official gloss
Relevance: The importance of the information for your needs.
- Does the information relate to your topic or answer your question?
- Who is the intended audience?
- Is the information at an appropriate level (i.e. not too elementary or advanced for your needs)?
- Have you looked at a variety of sources before determining this is one you will use?
- Would you be comfortable citing this source in your research paper?
That last official question is the one students skip. If you would be embarrassed to put the URL in a footnote, that is data. The fourth question is the one confirmation bias skips: look at a variety of sources before you lock this one in. See Bias in sources.
Authority: who is accountable, and are they qualified here?
Authority is a person or organization you can name, plus a reason they are the right kind of expert for this claim. A cardiologist is not automatically authoritative on medieval liturgy. A survivor is highly authoritative on what they experienced and a weak source for national prevalence. The ACRL Framework calls that constructed and contextual:
ACRL Framework, Authority Is Constructed and Contextual (2016)
Authority is constructed in that various communities may recognize different types of authority. It is contextual in that the information need may help to determine the level of authority required.
The credibility page quotes the longer paragraph. Here the job is operational: find a name, then leave the page.
Official Authority questions (Meriam Library)
Official gloss
Authority: The source of the information.
- Who is the author/publisher/source/sponsor?
- What are the author's credentials or organizational affiliations?
- Is the author qualified to write on the topic?
- Is there contact information, such as a publisher or email address?
- Does the URL reveal anything about the author or source? examples: .com .edu .gov .org .net
The URL question is a clue, not a seal. The .org top-level domain is unrestricted. A .edu host can serve a student paper or a 2009 lab page. A .gov host is a stronger signal for official U.S. public information, and you still check the office and the date. MedlinePlus: Vitamin D is a National Library of Medicine consumer page that attributes recommended amounts to the NIH Office of Dietary Supplements. A lateral search for “MedlinePlus NIH” returns the same federal library, not a supplement shop. That is how Authority is answered — off the page — not by admiring the layout.
Accuracy: can someone else check the claim?
Accuracy is evidence you can follow: a methods section, a questionnaire, a dataset, a cited table, a DOI that resolves. Spelling quality is a weak signal. Peer review is a stronger one and still not a lifetime guarantee — see How the peer-review process works and check Retraction Watch when a finding seems too tidy.
Official Accuracy questions (Meriam Library)
Official gloss
Accuracy: The reliability, truthfulness and correctness of the content.
- Where does the information come from?
- Is the information supported by evidence?
- Has the information been reviewed or refereed?
- Can you verify any of the information in another source or from personal knowledge?
- Does the language or tone seem unbiased and free of emotion?
- Are there spelling, grammar or typographical errors?
John P. A. Ioannidis's essay “Why Most Published Research Findings Are False” (PLoS Medicine 2, no. 8, e124, 30 August 2005, DOI 10.1371/journal.pmed.0020124) is a real, open-access, peer-reviewed argument with a 25 August 2022 correction. It is also a model, not a census of every paper. The title is doing rhetorical work. Accuracy here means: read the assumptions (study power, bias, how many teams chase the same question), do not treat the headline as a finding you can drop into a methods section, and do not use the essay as a license to ignore journals. Cross-check one claim — for Pew, open the methodology note; for BLS, open the Technical Note that explains the household survey versus the establishment survey.
Purpose: why does this page exist?
Purpose is the letter students confuse with “has a point of view.” Every source has a point of view. The official questions ask whether the page is there to inform, teach, sell, entertain, or persuade, and whether that intention is clear. A BLS news release exists to publish survey estimates on a schedule. A vitamin-shop landing page exists to sell bottles. Both can contain a true number. Only one is doing the job you think it is doing.
Official Purpose questions (Meriam Library)
Official gloss
Purpose: The reason the information exists.
- What is the purpose of the information? Is it to inform, teach, sell, entertain or persuade?
- Do the authors/sponsors make their intentions or purpose clear?
- Is the information fact, opinion or propaganda?
- Does the point of view appear objective and impartial?
- Are there political, ideological, cultural, religious, institutional or personal biases?
Lateral reading: how you actually answer Authority and Purpose
CRAAP still asks you to inspect the source. Professional fact-checkers do not finish that inspection from inside the source. Sam Wineburg and Sarah McGrew compared 10 Ph.D. historians, 10 professional fact checkers, and 25 Stanford University undergraduates on live websites. Their 2019 Teachers College Record paper, Lateral Reading and the Nature of Expertise, is the named study. Stanford also hosts the 28 July 2018 working copy (“in press, Teachers College Record”). The structured abstract is the wording this page quotes:
Wineburg and McGrew, Teachers College Record (2019) / Stanford working copy (28 July 2018)
Historians and students often fell victim to easily manipulated features of websites, such as official-looking logos and domain names. They read vertically, staying within a website to evaluate its reliability. In contrast, fact checkers read laterally, leaving a site after a quick scan and opening up new browser tabs in order to judge the credibility of the original site.
Vertical reading is what the About page and the padlock icon invite. Lateral reading is the first move of Mike Caulfield's SIFT method (Stop, Investigate the source, find better coverage, Trace the original context). The claim-checking half — find better coverage, then trace a quote or figure back — lives on How to fact-check sources. The investigation half is the Authority/Purpose method for this page:
Mike Caulfield, “SIFT (The Four Moves),” 19 June 2019
Taking sixty seconds to figure out where media is from before reading will help you decide if it is worth your time, and if it is, help you to better understand its significance and trustworthiness.
- Open a new tab before you finish the About page.
- Search the organization or author's name, then add a word such as “funding,” “retraction,” or “review.”
- Prefer Wikipedia, newsrooms, and university pages about the source over the source's self-description.
- If a statistic matters, trace it to the table or paper it came from and evaluate that.
- Come back to CRAAP. Date, audience, and methods still matter after you know who is talking.
When CRAAP is the wrong first tool
The Meriam Library handout assumes you already have an item in hand. That is the right tool for a PDF you pulled from the library, a BLS release, or a book on your desk. It is the wrong first tool for an unfamiliar URL that arrived in a group chat. On a stranger's page, start with SIFT: stop, then investigate the source in another tab. If you cannot name the publisher after a minute, you do not yet have a source to run CRAAP on.
Start with SIFT
- The URL is new to you.
- The layout looks official and you have not left the page.
- You care more about the claim than about this particular article.
Start with CRAAP
- You already know the publisher is real.
- You need to decide whether this dated item belongs in this paper.
- You are comparing two real candidates for the same footnote.
The ACRL frame Information Creation as a Process is the other reason a checklist run on the homepage fails. Format is not the same as the process that produced the item:
ACRL Framework, Information Creation as a Process (2016)
Information in any format is produced to convey a message and is shared via a selected delivery method. The iterative processes of researching, creating, revising, and disseminating information vary, and the resulting product reflects these differences.
A peer-reviewed PDF, a BLS HTML release, and a MedlinePlus topic page can all be “webpages” in your browser. The creation process — refereeing, a locked embargo, or a living consumer summary — is what Accuracy and Purpose are asking about. Do not score the chrome.
Worked evaluations: five real sources
The five items below are identifiable. Each one is run through all five letters. None of them is “always cite” or “never cite.” The question is what job they can do.
1. Pew Research Center, Americans' Social Media Use 2025 (20 November 2025)
Open the report. Authors Jeffrey Gottfried and Eugenie Park. Pew is a nonpartisan fact tank. SSRS ran the National Public Opinion Reference Survey for the Center: 5,022 U.S. adults, 5 February–18 June 2025, address-based sampling, web, mail, and phone. The page reports that 84 percent of U.S. adults have used YouTube, 71 percent Facebook, 50 percent Instagram, and 37 percent TikTok (up from 21 percent in 2021). A separate American Trends Panel wave (5,123 adults, 24 February–2 March 2025) asked how often people visit four platforms. The 31 January 2024 report (5,733 adults, 19 May–5 September 2023; 83 percent YouTube, 68 percent Facebook, 47 percent Instagram) now tells readers to use the 2025 release for the latest adult figures.
CRAAP on the 2025 Pew report
- Currency. Dated 20 November 2025 from a 2025 survey. Use this wave for current adult levels; keep the 2024 report only for that survey year.
- Relevance. Right for U.S. adult platform use. Wrong for teens, for one campus, or for causal claims about well-being.
- Authority. Named research center and named authors; methods and questionnaire are on the same site. A lateral search confirms Pew, not a lookalike pollster.
- Accuracy. Sample size, dates, mode, and weighting are published. Do not quote a percentage without the group (adults, not “Americans of every age”).
- Purpose. To inform. Pew states its funding and methods. It is still a survey of self-reported use, not a traffic log.
2. BLS, The Employment Situation — August 2026
Open the Employment Situation Summary. The August 2026 release (USDL-26-1435, 8:30 a.m. ET Friday, 4 September 2026) reported that total nonfarm payroll employment increased by 162,000 and the unemployment rate was unchanged at 4.1 percent (7.0 million unemployed). The same page scheduled the September 2026 release for Friday, 2 October 2026, at 8:30 a.m. ET. The July 2026 first-print payroll change of −23,000 was later revised to +21,000. Monthly headlines move; cite the release month, and expect later revisions.
CRAAP on a BLS monthly release
- Currency. Monthly by design. Cite the release month, and expect revisions in later releases — July 2026 is the worked case.
- Relevance. Right for recent U.S. labor-market levels. Wrong for one firm's headcount or for a historical chapter on the 1930s (use historical BLS series).
- Authority. A federal statistical agency. Search “BLS Employment Situation” and you land on bls.gov, not a newsletter summarizing BLS.
- Accuracy. Two surveys (household / establishment) with a Technical Note. A headline payroll number can move by tens of thousands when more reports arrive.
- Purpose. To inform on a locked schedule. Embargo language and the next-release date are printed on the page.
3. MedlinePlus, Vitamin D (consumer health page)
Vitamin D is a MedlinePlus topic page from the U.S. National Library of Medicine. The live page was last updated 10 May 2024. It summarizes why the vitamin matters and lists recommended daily amounts from the NIH Office of Dietary Supplements: 400 IU from birth to 12 months; 600 IU for children 1–13, teens 14–18, and adults 19–70; 800 IU at 71 and older; 600 IU if you are pregnant or breastfeeding. The page tells you to check with a health-care provider before you treat those numbers as a personal dose.
CRAAP on MedlinePlus Vitamin D
- Currency. A living consumer page last updated 10 May 2024. Check the date on any specific fact you quote, and follow the ODS link for the nutrient table.
- Relevance. Right for background in a nursing or public-health paper. Wrong as your only source for a clinical recommendation — follow the journal articles MedlinePlus lists.
- Authority. NLM / NIH. Lateral search: this is the federal consumer-health encyclopedia, not a clinic or a brand.
- Accuracy. Figures are attributed. Still a summary. For a methods-level claim, open the ODS fact sheet or a systematic review.
- Purpose. To teach patients and students. The “check with your provider” line is the purpose statement.
4. Ioannidis, “Why Most Published Research Findings Are False” (2005)
The PLOS Medicine essay (DOI 10.1371/journal.pmed.0020124) is one of the most-cited pieces in the replication conversation. Ioannidis writes, as his own claim:
John P. A. Ioannidis, PLoS Medicine 2, no. 8 (2005)
It can be proven that most claimed research findings are false.
That sentence is a conclusion from a model of power, bias, and how many teams chase significance — not a count of retracted papers. PLOS issued a correction on 25 August 2022 (10.1371/journal.pmed.1004085) for a missing set of parentheses in Table 2. Peer review got the essay into the journal; it did not make the title a fact you can cite without the argument.
CRAAP on Ioannidis 2005
- Currency. 2005, with a 2022 correction. Still cited because the argument is about study design, not last week's trial.
- Relevance. Right when you are discussing replication, p-values, or why a single small study is a thin foundation. Wrong as a reason to skip the literature in your field.
- Authority. John P. A. Ioannidis, then at the University of Ioannina / later Stanford; a named, highly cited methodologist. Confirm the author, not just the shocking title.
- Accuracy. Open-access, peer-reviewed, DOI resolves, correction exists. Read past the abstract. Do not convert the title into “most papers are junk.”
- Purpose. To argue for better design and interpretation. It is an essay, not a systematic review.
5. Kuhn, The Structure of Scientific Revolutions (1962)
Thomas S. Kuhn, The Structure of Scientific Revolutions (Chicago: University of Chicago Press, 1962) is a university-press book, not a webpage. Students throw it out because “it is old,” or they cite it for a CRISPR protocol because “it is a classic.” Both mistakes are Currency and Relevance failures, not Authority failures.
CRAAP on Kuhn 1962
- Currency. 1962 is current for the history of the idea of paradigm shift. It is not current for how a 2020s lab runs a genome edit.
- Relevance. Right for a philosophy or history paper about scientific change. Wrong as a methods citation in molecular biology.
- Authority. A named historian and philosopher of science; a university press. Later editions exist — cite the one you used.
- Accuracy. A book-length argument, not a dataset. Later scholars have criticized and refined it; that is scholarship as conversation, not a reason to hide the book.
- Purpose. To interpret how sciences change. It is not a lab manual and not a news report.
How the same questions change by source type
CRAAP does not change letters when the format changes. The evidence you can expect does. For the taxonomy of scholarly versus popular, use Scholarly vs popular sources. For unpublished firsthand material, use Primary vs secondary sources.
| Type | Usually strong on | Still check |
|---|---|---|
| Peer-reviewed article | Named authors, methods, references, a DOI | Funding, sample, whether it was retracted, whether you read past the abstract |
| University-press book | Editorial process, depth | Edition, date, whether later work replaced the claim |
| Government statistical release | Accountable agency, documented methods, a timestamp | Which survey, which revision, whether you need the historical series |
| News report | A dated public record of an event | Whether it is news or opinion; whether the study it summarizes is the thing you should cite |
| Organization or company site | A public statement of that group's position | Purpose and funding; do not treat .org as a quality mark |
| Wikipedia | A map of terms and footnotes | Do not cite it as the scholarly source — follow the notes. See Using Wikipedia for research. |
| Social post or blog | A primary source of what that person said | Almost never a substitute for the underlying paper or dataset |
Is this journal trusted? Think. Check. Submit.
Accuracy asks whether the item was reviewed or refereed. A PDF footer that says “peer-reviewed” is not enough. Think. Check. Submit. is the cross-publisher checklist for whether a journal is a trusted venue. The journals list was last updated in March 2023, with further amendments in June 2023. The closer for authors is also the closer for students deciding whether to treat a download as refereed:
Think. Check. Submit., Journals checklist (June 2023)
Is the journal clear about the type of peer review it uses?
Complete the checklist and submit your manuscript only if you can answer 'yes' to most or all of the questions above.
For a paper you are citing, the same questions still apply: can you identify and contact the publisher, is the journal name unique, and is the title in DOAJ if it claims to be open access? Then confirm the article was not later retracted. See predatory journal and How the peer-review process works.
Red flags that should stop the citation
None of these is official CRAAP scoring. They are the failure modes the five letters keep catching in student papers.
- No accountable author or organization. Anonymous is not the same as a corporate author you can name.
- A journal you cannot verify. Use Think. Check. Submit. and the Directory of Open Access Journals before you treat an open-access PDF as refereed. See predatory journal.
- A finding that has been retracted. Search the title plus “retraction,” or look up the DOI. The credibility page walks through a retracted Lancet paper; the rule is the same for any field.
- A first-print statistic that has already been revised. The BLS July 2026 payroll change moved from −23,000 to +21,000. Cite the release month and check the latest revision.
- Correlation sold as causation. Common in press releases and in papers that over-claim their design.
- Undisclosed funding on a product, a drug, or a policy the sponsor sells.
- You have only read the abstract. Abstracts omit limits. If you cannot see methods, you cannot finish Accuracy.
Mistakes the checklist does not prevent by itself
- Treating .edu or .gov as automatic credit. Student work and outdated agency pages live on those hosts.
- Judging by layout. Wineburg and McGrew's historians were caught by logos and official-looking domains.
- Confirmation bias. The official Relevance question already asks whether you looked at a variety of sources. If every citation agrees with you, you stopped early.
- Impact factor as a quality score for one article. Impact factor describes a journal's citation rate, not this paper. See Citation metrics.
- Hunting for a source with zero perspective. Account for purpose; do not wait for a view from nowhere.
- Stopping at the checklist. A site designed to look scholarly will pass a vertical CRAAP pass. Leave the page.
Tools that answer a CRAAP question
- [Google Scholar](https://scholar.google.com/) — other work by the author, and who cited this item (Accuracy, Authority).
- Ulrichsweb (through a campus library) — whether a title is refereed (Accuracy).
- [Retraction Watch](https://retractionwatch.com/) — whether the item was withdrawn (Accuracy).
- [Think. Check. Submit.](https://thinkchecksubmit.org/) — whether a journal is a trusted venue before you treat it as peer-reviewed.
- [DOAJ](https://doaj.org/) — a directory of legitimate open-access journals, not a list of every PDF on the web.
- [DOI validator](/tools/doi-validator) — whether the identifier you copied actually resolves.
A browser extension that paints a site “trustworthy” is not a substitute for the questions. Neither is a citation count. Those tools can support one letter. They cannot finish the evaluation.
Frequently asked questions
- Is the CRAAP test still taught, or did SIFT replace it?
- Both are still taught, for different jobs. CRAAP is the list of questions you run on a source you already intend to use. SIFT (Caulfield, 2019) is the set of moves you make in the first minute on an unfamiliar URL. Use SIFT to find out who is talking; use CRAAP to decide whether the dated, on-topic, checkable item in front of you belongs in this paper. This page owns CRAAP. How to fact-check sources owns the claim-checking half of SIFT.
- Do I have to answer every official CRAAP question for every source?
- No. The Meriam Library handout says different criteria will be more or less important depending on your situation. A 1962 university-press book does not need “are the links functional?” A monthly BLS release does. Skip a question only when you can say why it does not apply — not because you are in a hurry.
- How do I know if a journal is peer-reviewed?
- Do not trust the PDF's footer alone. Check Ulrichsweb if your library subscribes, the journal's own editorial policy, and Think. Check. Submit.. Then confirm the article was not later retracted. How the peer-review process works explains what refereeing does and does not guarantee.
- Are older sources automatically weak?
- No. The official Currency question is whether your topic requires current information. Kuhn (1962) is still assigned in history and philosophy of science. A 2024 social-media survey is already the wrong wave if a 2025 release exists and you are describing the present. Age is a clue about the claim, not a score for the object.
- Can I use Wikipedia when I evaluate other sources?
- Yes — as a lateral-reading tab. Wikipedia is often the fastest way to see what other people call an organization. It is a weak thing to cite as the scholarly source for the claim. Follow the footnotes. The longer workflow is Using Wikipedia for research.
- What if I cannot find anything about the author?
- Treat that as an Authority failure until you can name an accountable organization instead. A federal statistical release can be cited by agency. An unsigned blog post with no About page and no methods cannot. Find a more transparent source rather than inventing credentials.
- Is a biased source unusable?
- Not automatically. Purpose asks you to see the bias, not to wait for a source with none. An advocacy group's fact sheet can be a primary source of that group's position. It is a poor substitute for the underlying government series or journal article. Name the limitation in your prose. More in Bias in sources.
- How many sources should I evaluate before I pick one?
- There is no official quota, and “evaluate 25 to keep 12” is not a rule. The official Relevance question is whether you have looked at a variety of sources before deciding this is the one. For a short paper that may be a handful of candidates; for a thesis it is many more. Stop when you can explain why the ones you kept beat the ones you opened and rejected.
- Does a high journal impact factor mean the article is accurate?
- No. Impact factor is a journal-level citation rate. Ioannidis (2005) is the reminder that study power, bias, and how many teams chase the same question still matter inside famous venues. Evaluate the article: methods, sample, funding, and whether a correction or retraction exists.
- The BLS number I copied last month already changed. What now?
- That is Currency and Accuracy doing their job. The August 2026 Employment Situation (USDL-26-1435) is the current monthly release as of this rewrite; the July 2026 first-print payroll change of −23,000 was later revised to +21,000. Cite the release month, note the revision if the figure moved, and do not treat a first headline as a frozen fact. The September 2026 release is scheduled for 2 October 2026.
Finished evaluating? Cite the source you actually used
Once a source survives CRAAP and a lateral search, cite the version you read — the Pew wave, the BLS release month, the book edition, the DOI that resolves. The credibility page is the decision; this page is the method.