Provenance and credibility in scientific data repositories

Abstract

Despite a long history of rich theoretical work on provenance, empirical research regarding users’ interactions with and judgments based upon provenance information in archives with scientific data is extremely limited. This article focuses on the relationship between provenance and credibility (i.e., trustworthiness and expertise) for scientists. Toward this end, the authors conducted semi-structured interviews with seventeen proteomics researchers who interact with data from ProteomeCommons.org, a large online repository. To analyze the resulting interview data, the authors apply Brian Hilligoss and Soo Young Rieh’s empirically tested theoretical framework for user credibility assessment. Findings from this study suggest that together with other information provided in ProteomeCommons.org and subjects’ own experiences and prior knowledge, provenance allows users to determine the credibility of datasets. Implications of this study stress the importance of the archival perspective of provenance and archival bond for aiding scientists in their credibility assessments of data housed in scientific data repositories.