V. Discussion

\label{v.-discussion}
The deposit rate of articles written by OSU authors is around 45% as revealed in an early study (citation to our 2015 OA paper), which indicates a space of improvement in OA promotion. Previous studies have shown that time and effort required for self-deposit has been identified as a significant barrier for recruiting article from faculty (citations). Two main issues related to this barrier are poorly designed submission interface that is not intuitive (citation) and the time required to create or extract content metadata (citation). These findings motivate the development of the OA form, which is designed to be simple to use and served as the de-facto interface for article submission to ScholarsArchive@OSU. Since it was launched in March 2014, there is a slow but clear trend that the OA form is adopting by more authors. Although majority of the authors still submit their manuscripts as email attachment, we observe a scenario that a dozen authors voluntarily deposit articles that are not requested by the library using the link to OA form included in the library’s article recruiting email in year 2015 alone. This scenario is encouraging because authors all prefered to send additional articles in email to library before the OA form is available, and none of them ever used the submission interface of ScholarsArchive@OSU. In result, the OSU Library recently adds one sentence in the article recruiting email to explicitly promote the OA form as the preferred deposit mechanism over email attachment.
The development of OA form is library’s continued effort of creating, harvesting, and enhancing existing metadata for contents in IR (citation 1, 2). The metadata created for records in ScholarsArchive@OSU is mixed with copyright related fields generated by library staff (e.g., license types and embargo) and bibliographic fields primarily generated by external sources (e.g., publishers or authors). When an article manuscript is submitted as email attachment, the library staff has to copy and paste the bibliographic metadata from its publisher version into ScholarsArchive@OSU, which is not only labor intensive but also open for human errors. One such example is that an article in Physics has more than 100 people as co-authors and requires that the order of these authors in IR should be the same as its publisher version preciously. The OA form is designed to solve this metadata harvesting issue by extracting record from CrossRef’s database by DOI and automatically populate the corresponding Dublin Core qualifiers in IR.
Just getting articles from faculty remains the biggest challenge, but also getting the version that we are allowed to deposit to federal agencies. Its always a challenge promoting our services successfully. I don’t think email is terribly effective communication tool any more and think that faculty ignore them generally. So getting the word out is a big challenge.
libraries go to more trouble than we’d like checking to make sure that the articles we receive from faculty contain the required supplementary data and are the accepted manuscript versions required by the agencies rather than final copy-edited versions of record.