Start with the source, not the headline
Credibility lives in the details, and most people skip them. A confident tone, a professional-looking website, even a DOI — none of these alone guarantee you're looking at reliable information. What does work is a short, repeatable set of questions you can put to anything: a journal article, a news report, a preprint, a website, a dataset. Run through them in order and you'll catch most problems before they cause you trouble.
Who created this, and why should they know? Begin with authorship. A named author with a verifiable institutional affiliation and a publication record in the relevant field is a strong signal. Anonymous content, vague organisational by-lines, or credentials that don't survive a search are red flags. This doesn't mean unknown authors are always wrong — but it does mean you have less to go on, so other signals need to compensate.
Where was this published, and what is that publisher's reputation? For research, check whether the journal appears in the Directory of Open Access Journals (DOAJ) or a recognised index for your field. Peer review isn't a perfect filter, but it is a meaningful one: a paper that has been scrutinised by independent experts before publication sits on firmer ground than one that hasn't. Be alert to predatory journals — operations that charge fees, mimic legitimate venues, and skip genuine peer review. The look of legitimacy is easy to manufacture; the substance takes real checking.
When was this published, and does currency matter here? A 2009 paper on smartphone user behaviour is almost certainly outdated; a 2009 paper on the causes of the First World War may still be authoritative. The question isn't just age — it's whether the field moves fast enough for age to matter. Check whether newer work has superseded the claims, especially in fast-moving areas like public health, climate science, or technology.
Dig one layer deeper
What is the evidence, and where does it come from? Credible sources show their working. In academic writing, that means citations and a methodology you could scrutinise. In journalism, it means named sources, documents or data you could trace. Be suspicious of claims that are asserted rather than evidenced, statistics that appear without a source, and quotations that lack attribution. A reference list isn't decoration — it's an invitation to check.
Who funded or commissioned this work? Funding doesn't automatically bias findings, but undisclosed funding is a problem, and certain funding structures create incentives worth noting. Reputable journals require authors to declare conflicts of interest. If that disclosure is absent or hard to locate, treat the findings with extra caution — not dismissal, but careful scrutiny.
Has this been cited, critiqued or corrected? A paper that has been cited approvingly by subsequent researchers, especially those outside the original authors' own network, has survived a form of ongoing scrutiny. Equally important: check whether the work has been challenged, retracted or corrected. Retraction Watch maintains a public database of retracted papers; a quick search before you rely on a study is fast insurance. Preprints on arXiv, PubMed Central, or Zenodo can be excellent sources — many are the final published version in all but formatting — but they haven't been peer reviewed, so note that status.
Does this source agree with the broader body of evidence? A single study, however well-conducted, is not settled science. A finding that contradicts the established consensus deserves closer reading, not automatic rejection — genuine paradigm shifts do happen — but extraordinary claims require extraordinary support. If one paper tells you everything everyone else has concluded is wrong, that's not a reason to celebrate a scoop; it's a reason to read the methodology extremely carefully.
A checklist, not a tribunal
None of these questions is a trapdoor: one weak answer doesn't automatically disqualify a source, and strong answers don't guarantee perfection. Use them together. A source with a named expert author, peer-reviewed publication, transparent funding, clear evidence and broad corroboration is worth trusting. A source that fails most of those tests — but sounds compelling — is worth setting aside until you've found something better.