ETD Guide/Universities/Measuring Production and Use of ETDs: Useful Models

In addition to assessing some of the programmatic goals of the ETD program, institutions will want to have some basic assessment measures in place to document the production and use of their ETDs. The work done at Virginia Tech’s Electronic Thesis and Dissertation Initiative can provide a model for other institutions. Using web statistics reporting software, Virginia Tech monitors a number of measures for their ETDs, including:

  • the availability of campus ETDs
  • multimedia in ETDs
  • which domains within the United States and abroad are requesting ETDs
  • requests for PDF and HTML files
  • distinct files requested
  • distinct hosts served
  • average data transferred daily

In compiling its counts, Virginia Tech eliminates everyone working on ETDs at the institution, including the Graduate School. They try to eliminate repetitive activity from robots and other sources of that type as well. Virginia Tech is also working on the compilation of an international count of ETDs produced in universities.

Institutions must decide whether they will report their ETD collections and usage separately, in conjunction with other campus web site usage, or in conjunction with other electronic resources managed by the library. Now, when practices and standards for gathering these statistics are evolving, institutions may need to collect the information and report it in conjunction with more than one type of related collection. In any case, institutions should keep abreast of national and international initiatives that are seeking to define and standardize statistical reporting of the number and use of electronic information resources.

Several projects currently focus on collection of statistics related to information resources:

  • The Association of Research Libraries (ARL) collects statistics from libraries of large research universities in North America and has been working on several initiatives that explore data collection related to digital information. In particular, one of its New Measures initiatives, the E-Metrics Project, is recommending a particular set of measures and defining their collection.
  • The International Coalition of Library Consortia’s (ICOLC) has created Guidelines for Statistical Measures of Usage of Web-Based Indexed, Abstracted, and Full-Text Resources. They encourage vendors of electronic information products to build into their software statistical report generation that will meet the ICOLC Guidelines, promoting comparability of products.
  • In Europe, the EQUINOX Project, funded under the Telematics for Libraries Programme of the European Commission, addresses the need to develop methods for measuring performance in the electronic environment within a framework of quality management. Their products include a consolidated list of data sets and definitions of terms.
  • Library Statistics and Measures, a web site maintained by Joe Ryan of Syracuse University, also provides a useful set of links to resources on library statistics and measures.

Another important type of post-processing is the extraction of statistical information from metadata sets. For administrative purposes, institutions may be interested in the number of ETDs supervised by each professor, the keywords most used , the month(s) in which more ETDs are submitted, etc. Usually, relevant metadata are extracted from the ETD database and processed using specialized tools like Microsoft Excel. The access to the database can be done using either ODBC drivers or specialized middleware utilities.


Next Section: Statistics and Usage