Readers’ attitudes to self-archiving in the UK

The online self-archiving by authors of their scholarly articles has been proposed as an alternative to author-pays open access publication, but has the potential to undermine journal publisher income if the ready availability of self-archived articles leads to a drop in subscriptions. This study investigated the awareness of self-archiving and use of self-archived articles in a survey of a mainly academic population including both authors and non-authors, and looks at their attitudes to self-archived papers and whether they view them as an authoritative alternative to subscription access. In total, 70% of respondents had heard of self-archiving, though only 15% knew a lot about it, and 71% had used self-archived papers. These proportions are higher than in previous studies, suggesting that awareness has grown. Most self-archived papers used came from websites rather than repositories, particularly among those whose awareness of self-archiving was low. Use of self-archived articles was greater amongst those who had published more papers and also depended strongly on subject field – use and awareness were both particularly low in the field of medicine. People who were more aware of self-archiving were less likely to view the publisher’s official version as the only authoritative version and more likely not to care about the online location of articles. Moreover, authors who had self-archived tended to archive the publisher’s official version regardless of whether they were permitted to. These results suggest that the awareness of self-archiving is currently mostly limited to academic authors and is unlikely to grow beyond this in the short term. However, in the long term, the combination of high rates of self-archiving of the publisher’s official version, coupled with the devaluation of the journal as the authoritative source of material together with increased convenience of access to self-archived material, could result in fewer people accessing articles through subscription-based methods