Based on an analysis of about 19,000 journal self-archiving policies in RoMEO, SHERPA estimates that only 5% of these journals prohibit self-archiving.
Here's an excerpt from the announcement:
This chart shows that a remarkable 94% of journals allow archiving of peer-reviewed articles after any embargo period has expired and any additional restrictions have been complied with. Indeed, for nearly a quarter of journals, the publisher's version/PDF itself can be archived. Just 1% of journals only permit the pre-peer review submitted version to be archived. This leaves only 5% of journals that do not permit self-archiving of some form or another. . . .
Unfortunately, assigning journals to policies is not an exact process, due to the vagueness of some publishers' policies and the fact that some publishing houses do work for societies and other third parties whose own open access policies may take precedence. It is therefore difficult to gauge the precision of these figures, but we guestimate that they are accurate to within 2%. The charts do not take into account journals that are not covered by Romeo's own database, but we expect that the relative proportions would be similar.
| Transforming Scholarly Publishing through Open Access: A Bibliography | Digital Scholarship |