Validating Quality in Large-Scale Digitization

Project Summary

IN LESS THAN A DECADE the large-scale digitization of books has begun transforming the way we read and learn and changing how research libraries manage and preserve their collections.

DIGITIZED BOOKS made by third-party vendors are being preserved in online repositories. In this new preservation environment, the quality of what is preserved becomes an important factor in inspiring trust that digitized books are fit for the purposes envisioned for them.

INNOVATIVE RESEARCH presented at this website is developing and testing methods for measuring the severity of detectable errors in digitized books and validating the impact of error on the end-user. Here you will find information on the project, selected findings, and links to the project’s reports, presentations, publications, and products.

HATHITRUST DIGITAL LIBRARY serves as a testbed of digitized books and serials for the project, which has three overlapping phases.

  • Phase 1 (2011) - Define a model of digitization error and a severity scale for recording observed error consistently and accurately.
  • Phase 2 (2011-12) - Apply the research methodology to representative samples of digitized volumes.
  • Phase 3 (2012-13) – Validate the results of the error analysis for specific use-case scenarios.

THE SCHOOL OF INFORMATION at the University of Michigan is leading a multi-year collaboration with the University of Michigan Libraries, the HathiTrust Digital Library, and the University of Minnesota Libraries. The project is generously supported by the Institute of Museum and Library Services and builds on a planning effort sponsored by the Andrew W. Mellon Foundation.

Return to top