Manifestations of Metadata Structures in Research Datasets and Their Ontic Implications

ABSTRACT The proliferation of research datasets and their availability in various repositories require metadata that provides sufficient context and organizational clarity to enable their reuse. However, datasets come in myriad forms, structures, and relationships. As characteristics of datasets vary across disciplines, it is reasonable to suggest that the methods by which they are discoverable by metadata should be informed by the considerations unique to differing research areas. As such, for their internal unity to be accurately represented it is necessary that their ontological characteristics be documented reflectively along the same ontic spectrums. This article will explore the relationship between metadata and dataset structures in order to illuminate ontic alignments between them and how this impacts contextualization. We will examine common data structures for dataset metadata and implications each contains. A survey of the types of datasets from the Life Sciences, Social Sciences, Humanities, Geospatial Information Systems, and Research Software will be examined for their ability to be accurately documented by common standards, particularly in discipline-specific repositories. Similarly, an investigation into how metadata is being used in different contexts and how various systems help or hinder its fullest functioning may point towards challenges operating on a deep level.