Funders in many countries require the accepted manuscript to be deposited in a public repository, usually after an embargo.
Funders in many countries require the accepted manuscript to be deposited in a public repository, usually after an embargo. These versions are the peer-reviewed text with different typesetting, and they are legitimate copies posted by the authors or their institutions.
National and institutional repositories hold enormous collections and are indexed by ordinary search engines. Searching the exact title in quotation marks surfaces a deposited copy more often than most readers expect.
Every formally published paper carries a persistent identifier. Searching that string rather than the title avoids confusion between similarly named papers and between versions, and it is what repositories index against.
Where a preprint exists, it is freely available permanently. It is not identical to the final paper, and comparing them shows what review changed, which is occasionally more interesting than either version alone.
Corresponding authors are listed on every paper and are generally willing to send a copy. It is a normal part of academic practice rather than an imposition, and it is explicitly permitted by most publishing agreements.
Public and university libraries frequently provide access to large collections, including remotely for members. For readers outside academia this is the most under-used route to the literature.
Accepted manuscripts, published versions and preprints can differ in numbers as well as wording. Quoting a figure means knowing which version it came from, and the version is usually stated on the first page.
A claim traced only to an abstract cannot be checked. Most of the time the full text is a minute away and free, which removes the main practical excuse for not verifying a cited figure.
An accepted manuscript and the published article can differ in numbers as well as wording, because corrections and reviewer-requested analyses land between them. Quoting a figure means knowing which version supplied it, and the version is stated on the first page of most deposited copies.
Most citation errors are found by reading the source, and most people do not read the source because they assume it is paywalled. Since a free legitimate copy usually exists, the practical barrier to checking is smaller than it appears.
The claims on this page about most paywalled papers have a legal free version, and finding it takes a minute can be checked against sources outside this site: