Rachel A. Mandell has self-archived her thesis, "Researchers' Attitudes towards Data Discovery: Implications for a UCLA Data Registry," in SSRN.
Here's an excerpt:
The UCLA Data Registry is a tool designed to serve the greater UCLA research community by collecting and making available surrogate records of research datasets. To figure out how to build this system in accordance with the needs of the community, a total of 20 researchers from disparate disciplines were interviewed about their data and metadata practices. The results indicate that researchers' attitudes and behaviors towards making their work discoverable depend on their concept and definition of data. Given that the UCLA Library will build the UCLA Data Registry, it is important to consider the other possible tools that researchers could use in conjunction with the registry to enhance the discoverability of their data. The Data Registry will be built utilizing a basic metadata schema rather than very specific descriptive fields. The interviews also demonstrated that the culture of publishing and venues for data dissemination are shifting away from the traditional journal article publication, especially in emerging areas such as the digital humanities.