U S Bureau of Labor Statistics ResearchBased Metadata

U. S. Bureau of Labor Statistics Research-Based Metadata Requirements for a BLS Reports Archive Scott Berridge John Bosley Daniel W. Gillman US Bureau of Labor Statistics

Current Historical Archive Situation • Publications date to 1886 • For many publications – 1 copy exists • Many publications irreplaceable – no disaster recovery 14 September 2006 / UN Dissemination 2

Relevance to BLS • Unique historical archives – 120 years of publications – 10 years systematically available online • User expectations are rising • Standards are evolving • Mandates are under discussion 14 September 2006 / UN Dissemination 3

Meeting Emerging Standards • File Format – Adobe Acrobat PDF/A • Labeling – Adobe’s Extensible Metadata Platform (XMP) • Metadata Schema – Data Documentation Initiative (DDI) 14 September 2006 / UN Dissemination 4

Archive on Bureau’s Website • • www. bls. gov Servers inside and outside firewall Public domain Accessible thru BLS Home Page 14 September 2006 / UN Dissemination 5

Choosing Metadata Elements • Schema Choice – DDI Subset • Schema Useful? – Dissemination and Preservation – Perform user studies – 2 phases • Phase 1 -- Initial studies (3) – What users want / need – Relatively open-ended, exploratory • Phase 2 -- Focused studies (2) – Expose users to DDI subset – Obtain feedback 14 September 2006 / UN Dissemination 6

Reasons for DDI • Want success @ BLS • Minimize capture burden • Unknowable metadata – Old documents (back to 1886) – DDI allows level of granularity • Conformance – Easy with DDI 14 September 2006 / UN Dissemination 7

Potential Problem • • Dissemination and Preservation? Is DDI suitable for preservation? Preservation elements Looking at other standards 14 September 2006 / UN Dissemination 8

User Studies--Overview • Adopt techniques of – focus groups – group interviews • Recruited members of general public – Screened for familiarity w/ Gov’t Stats • Five groups, 2 -6 members per group – 3 groups -- initial, exploratory – 2 groups – review DDI elements 14 September 2006 / UN Dissemination 9

User Studies Results • Initial exploratory groups (3) – Descriptors consistently identified • title, date, and geographic coverage • Descriptors essential and informative – Members – generalize in abstract - hard • Difficulty imagining other descriptors • If personal need does not exist a priori, then – “What can I do with this information? ” – “How is it useful to me? ” 14 September 2006 / UN Dissemination 10

User Studies Results • DDI-focused groups (2) – Descriptors verified • Keywords also very important • However, very large sets of keywords – Counterproductive or Confusing – “TMI” – information overload 14 September 2006 / UN Dissemination 11

User Studies Results • DDI-focused groups (2) – Some want title or subtitle to answer • “How is this report useful to me? ” • “How can I put it to use? ” • Example – Title includes “A guide” – More useful than the formal title/name – Criterion for acquiring a document • Action, use of information • Not subject matter 14 September 2006 / UN Dissemination 12

Conclusion • Questions • Contact – Scott Berridge • Berridge. Scott@BLS. Gov – John Bosley • Bosley. John@BLS. Gov – Dan Gillman • Gillman. Daniel@BLS. Gov 14 September 2006 / UN Dissemination 13