Cornell PhD students Philip Davis and Matthew Connolly in their D-Lib article "Institutional Repositories: Evaluating the Reasons for Non-use of Cornell University's Installation of DSpace" describe a picture of faculty apathy towards the institutional repository (IR) at Cornell.
After a balanced description of the camps advocating IRs (not easy to do in this day and age), the authors describe some of the reasons faculty don't deposit research in the DSpace IR at Cornell:
1. Faculty prefer to be associated with their research field rather than their institution.
2. Faculty may be concerned about how the practice of depositing articles in an IR will affect their chances to be published in an archival journal, which is still very important to them.
3. Faculty may work in a field where early dissemination of results is not desirable, either for reasons of competition or of completenss, accuracy and quality.
4. Faculty may find the IR technology more difficult or less convenient than posting to a personal or departmental web page.
The authors found that the culture of researcher's discipline was an important component to the use of the IR. For example, the library collection is one area of Cornell's IR that is seeing steady growth. Presumably, the librarians are the most familiar and supportive of the IR, and so use it for their materials.
This study presents actual data (how refreshing) on how it's really going with IRs. Proponents can use the information here to make IRs more successful. Detractors can use it to say they are failing.
The authors don't talk about it, but I think what's really missing here is marketing. People (researchers included) will generally take advantage of a tool, service, or system that benefits them if they know about it. Until IRs have a sense of importance or imprimateur, or unless they provide a place to put stuff that has no other home (teaching materials, grey literature, datasets), it doesn't suprise me that they are underutilized.
Marketing? Isn't that something publishers are pretty good at? But it takes resources.
Once again, we see that as libraries take on the traditional roles of publishers, they need to become more like publishers in order to succeed.