Predatory publishers are distinguished by intentional acts that are meant to deceive or harm. They should not to be confused with legitimate publishers who offer honest services. The 'types' presented on this page are not exclusionary: it's possible for more than one to apply to a predatory operation.
Poses as a well-established journal or as a publication associated with a well-known brand or society.
Lures you in with promises then charges large fees after your paper has been accepted. Persistent phishers may demand payment even though no paperwork has been signed and no promises made.
Mass production of shoddy work made to order, often through machine-learning or plagiarism. Unlike the other types of predators, papermills are meant to deceive readers and editors, not authors.
Has a legitimate and impressive looking website but upon closer inspection nothing is what it seems. The journals are empty shells or worse, populated by stolen, plagiarized, or gibberish articles
Too good to be true! Unicorns claim to offer services -- fast peer review, indexing in databases, impact factors, etc. -- but don't deliver.
As of 2022-11-09 "Warning signs" that reinforce global and academic inequities have been removed from this guide. You can read more about why this was done on the Resources, Disclaimers, Etc. page.
The list of "warning signs" on this page was assembled by the author from multiple sources and her personal experience working with authors and open access licenses over the past nine years. Teixeira de Silva et al. (2022) was especially helpful in identifying problematic criteria often found on 'white' and 'black' lists.
This guide is just that, a guide. Ultimately it is up to each author to make the final decision on where to publish and to decide what they expect from their publisher.
Megan N. O'Donnell authored this guide. Last updated on 2023-01-04. First published in 2016, this work is licensed under a Creative Commons Attribution 4.0 International License. The graphic icons from icons8 and can be reused as long as attribution is retained.