Abstract
Purpose - Researchers at the University of Tennessee Libraries experimented with crowdsourcing to determine if contributions by members of the public could be utilized to add citations and subject tags to a new online bibliography, Database of the Smokies (DOTS: dots.lib.utk.edu). Design/methodology/approach - The database is built in Drupal, an open source platform that provides a crowdsourcing mechanism. The public was offered the opportunity to create accounts and add content to DOTS. After three months, the project team performed a transaction log analysis of user submissions in order to determine whether an editorial process was necessary.Findings - This analysis revealed that 18% of database content was the result of crowdsourcing and that much of the content submitted by participants was either obscure or difficult to locate. The analysis also showed that while contributors added valuable citations, an editorial review process was necessary to ensure this crowdsourced content could be published in the database. In addition, contributor supplied subject tags were not of significant uniqueness or quantity to substantially influence the existing taxonomy. Finally, the publicity of the crowdsourcing feature allowed other institutions to contribute to the project and add rare material.Originality/value - This paper offers a model for utilizing crowdsourcing to entice a sophisticated user group to help build a bibliographic database.
Purpose - Researchers at the University of Tennessee Libraries experimented with crowdsourcing to determine if contributions by members of the public could be utilized to add citations and subject tags to a new online bibliography, Database of the Smokies (DOTS: dots.lib.utk.edu). Design/methodology/approach - The database is built in Drupal, an open source platform that provides a crowdsourcing mechanism. The public was offered the opportunity to create accounts and add content to DOTS. After three months, the project team performed a transaction log analysis of user submissions in order to determine whether an editorial process was necessary.Findings - This analysis revealed that 18% of database content was the result of crowdsourcing and that much of the content submitted by participants was either obscure or difficult to locate. The analysis also showed that while contributors added valuable citations, an editorial review process was necessary to ensure this crowdsourced content could be published in the database. In addition, contributor supplied subject tags were not of significant uniqueness or quantity to substantially influence the existing taxonomy. Finally, the publicity of the crowdsourcing feature allowed other institutions to contribute to the project and add rare material.Originality/value - This paper offers a model for utilizing crowdsourcing to entice a sophisticated user group to help build a bibliographic database.