What’s in the Box? Assessing the potential usability of four decades of thesis and dissertation supplementary files

Steven Van Tuyl

doi:10.7191/jeslib.2019.1142

Abstract

Objectives: The objective of this study is to evaluate the quality and usability of supplementary data files deposited, between 1971 and 2015, to our university institutional repository. Understanding the extent to which content historically deposited in digital repositories is usable by today’s researchers can help inform digital preservation and documentation practices for researchers today. Methods: I identified all graduate level theses and dissertations (GTDs) in the institutional repository with multiple files as a first pass at identifying documents that included supplementary data files. These GTDs were then individually examined, removing supplementary files that were artifacts of either the upload or digitization process. The remaining “true” supplementary files were then individually opened and evaluated following elements of the DATA rubric of Van Tuyl and Whitmire (2016). Results: Supplementary files were discovered in the repository dating back to 1971 in 116 GTD submissions totalling more than 25,000 files. Most GTD submissions included fewer than 30 files, though some submissions included thousands of individual data files. The most common file types submitted include imagery, tabular data, and databases, with a very large number of unknown file types. Overall, levels of documentation were poor while actionability of datasets was generally middling. Conclusions: The results presented in this study suggest that legacy data submitted to our institutional repository with GTDs is generally in poor shape with respect to Transparency and somewhat less so for Actionability. It is clear from this study and others that researchers have a long road ahead when it comes to sharing data in a way that makes it potentially useable by other researchers.

Highlights

Over the past decade, researchers in academia and government have been increasingly called on to better manage and share the results of their research, including publications and datasets
Since the mid-1990s, ScholarsArchive@Oregon State University (OSU) has seen about 4-8 supplementary file deposits per year - a rate that remains to the end of the study period
It should not be much of a surprise that much of the data associated with graduate-level theses and dissertations (GTDs) submissions does not live up to emerging standards for data sharing in repositories, given the decades of difference between these submissions and our current thinking about data sharing

Summary

Introduction

Researchers in academia and government have been increasingly called on to better manage and share the results of their research, including publications and datasets. While not explicitly offering guidance for best practices, a number of researchers have explored the current behaviors and attitudes of researchers with respect to data management and sharing including a host of largely domain-agnostic institutional surveys (e.g. Van Tuyl and Michalek 2015, Akers and Doty 2013, Whitmire et al 2016, and Rolando et al 2013, among many others), and a number of domain-specific studies to the same effect in areas such as Neuroimaging (Borghi and Van Gulick 2018), Structural Engineering (Johnston and Jeffryes 2013), Atmospheric Science (Wiley and Mischo 2016), Biomedical Clinical Research (Federer et al 2015), and Water Quality (Carlson and StowellBracke 2013)

Objectives

Methods

Results

Discussion

Conclusion