MLA News Selected Articles

September 1999

Results of 1999 Bylaws
Amendments Ballot Announced

An Elections Committee comprised of Ramune K. Kubilius, AHIP, Galter Health Sciences Library, Northwestern University, Chicago; Neera Kukreja, Academic Center Library, Cook County Hospital, Chicago; and Patricia E. Pinkowski, AHIP, Green-Field Library, Alzheimer’s Association, Chicago; met at MLA headquarters on August 12, 1999, to count ballots for proposed amendments to the MLA Bylaws. Adoption of the amendments to the bylaws requires that at least 25% of the total ballots mailed be returned for counting and two-thirds of these ballots must be in the affirmative for each of the amendments.

Members attending the business meeting during MLA ’99 in Chicago, IL, recommended that the Bylaws’ amendments be sent to all voting members by mail ballot. On June 28, 4,306 ballots were mailed to all MLA voting members. By the July 28 deadline, 1,344 ballots were returned—exceeding the 1,077 that were needed to comply with the Bylaws.

• Amendments one and two, which proposed changing the slate for the office of president-elect from dual slate to a single slate, did not pass due to the lack of a two-thirds approval.
• Amendments three through seven, which relate to the restructuring of the Section Council, passed.

Detailed information on each of the amendments can be found in the March 1999 issue of the MLA News. Amendments three through seven become effective January 1, 2000. The table on page 16 provides the wording and vote count for each of the motions contained on the ballot.

Bylaws Amendment Election Results

Research Spotlight:
Search Efficiency Studies

Submitted by Kristin Stoklosa, National Institutes of Health Library, Bethesda, MD, and Research Resources Committee, MLA Research Section

Editor's Noted: This series features MLA members’ research projects published outside the library literature in scientific and biomedical publications. The Research Resources Committee of the MLA Research Section shares this series to promote awareness of information research, to encourage research in library practice, to stimulate interest in a variety of publications, and to inspire further MLA research.

Efficiency is the watchword for both studies in this Research Spotlight in the context of online searching and information access for the fulfillment of information needs. Finding an optimal search formula to generate single-screen online public access catalog (OPAC) display is the goal of a highly focused study by Frederick G. Kilgour, AHIP, School of Information and Library Science, University of North Carolina–Chapel Hill. In turn, an article by Alexandra Dimitroff, Ph.D., AHIP, School of Library and Information Science, University of Wisconsin–Milwaukee, contributes new ideas to the integration of both hypertext and Boolean features into bibliographic database systems.

"Retrieval Effectiveness of Surname-Title-Word Searches for Known Items by Academic Library Users"

MLA Member and Project Leader
Frederick G. Kilgour, AHIP, School of Information and Library Science, University of North Carolina–Chapel Hill

Source
Kilgour FG, Moran BB, Barden JR. Retrieval effectiveness of surname-title-word searches for known items by academic library users. J Am Soc Inf Sci 1999 Mar;50(3):265–70.

Project Description
Asserting that the best screen design required the least user manipulation, the practical goal of this research was to develop an OPAC search formula that would retrieve single-screen displays 98% to 99% of the time. The study used a systematic sample of titles from University of North Carolina–Chapel Hill circulation records with 917 title pages representing known items. In the University of Michigan’s NOTIS system, the researchers searched each title using between one and three successive search formulas: starting with a strategy of author surname and first significant title word, each search added one word until a single-screen display was generated for the title. The single-screen was conservatively defined as a twenty-line display with a separate line for each author followed by lines for each title. Of the 806 titles with hits, one word produced a single screen for 78% and two words for 19% of the titles; the authors had to resort to three words for 1.5% of the titles. Thus, their recommended strategy, which generated single-screen displays for 98.5%, was surname plus three title words. The only study limitation was that twenty of the positive results were achieved by fielded searching using MARC numbers as search suffixes, a strategy likely to be too abstruse for many users. Encouraged by their findings, the authors forecasted automated cataloging in an electronic book environment with OPACs populated by digital title pages.

"Hypertext vs. Boolean-Based Searching in a Bibliographic Database Environment: A Direct Comparison of Searcher Performance

MLA Member and Project Leader
Alexandra Dimitroff, Ph.D., AHIP, School of Library and Information Science, University of Wisconsin–Milwaukee

Source
Wolfram D, Dimitroff A. Hypertext vs. Boolean-based searching in a bibliographic database environment: a direct comparison of searcher performance. Inform Process Manag 1998 Nov;34(6):669–79.

Funding
Supported in part by a grant from the Ronald McNair Research Internship Program.

Project Description
This study compared hypertext and Boolean bibliographic retrieval systems according to the performance of novice and experienced searchers. Developed by the researchers, both database systems used the same subset of the National Technical Information Service (NTIS) database. For the study, the researchers constructed one known-item search and three subject searches with small, medium, and large potential relevant-retrieval sets. All participants searched both systems using the same four strategies. Precision and recall measures compared system logs of users’ retrieval sets with the total relevant citation sets. A "reality check" measure of searcher confidence in recall was a unique study feature: comparing actual recall with user perception of success indicated whether users over- or underestimated recall, informing system usability assessments.

The results confirmed the strengths of hypertext for inclusive exploration and Boolean operators for exclusive searching, making the latter advantageous for precision in a large database. Higher Boolean-operator recall for the large subject search also highlighted the limitations of hypertext for large retrieval. However, differences in time performance measures between experience levels did not result in gains in recall or precision. The findings suggested a combination of Boolean operators and hypertext features in information retrieval systems. With hypertext links within Boolean result sets, users could create subsets of relevant items by meaningful rather than linear traversal of initial results.

Streaming Videos for Health Information

Edited by Melissa L. Just

Submitted by Howard Fuller, Health Library, Stanford Hospital, Palo Alto, CA

Working with the Stanford University Medical School IT group (MedIT), The Health Library at Stanford Hospital has been able to offer what might be the largest collection of digitized consumer health videos viewable through a Web interface. Although most of these programs were originally developed for the lay consumer, the videos have been used for educational purposes by medical and allied health education programs.

Over sixty videos are currently available for downloading. Many programs are provided in RealVideo from RealNetworks, and a small selection is available in high bandwidth streaming MPEG-1 videos from Optibase, Inc. The RealVideo collection is downloaded at one of six bandwidths. The Stanford server will automatically download the video stream at the fastest available speed as determined by the users’ telecommunication connection. Viewing the videos requires RealPlayer software of at least version 5.0 and preferably RealPlayerG2. RealPlayer offers trial software that may be downloaded at no charge from the RealNetworks Website. The MPEG-1 format requires a high-speed Internet connection of 1.5 megabits per second or higher, as well as a powerful computer such as one with a Pentium II chip or equivalent processor or a PowerPC Macintosh. The WebFORCE MediaBase Player must also be installed on the local computer.

The Health Library is a major source of health education programming that presents lectures, workshops, and classes by the staff of Stanford Hospital & Clinics, Stanford University School of Medicine, and the University of California–San Francisco. All programs are professionally videotaped, edited, and added to the collection for circulation. Select videotaped programs are digitized and made available on the Website for downloading. The videos are arranged into one of four series: Women and Health; Health and Society, a lecture format program; Health Matters, programs originally produced for PBS Television and later the local PBS station, KRON, as a community service of Stanford University Medical Center; and the Jonathan L. King Lectureship series, which encourages the compassionate and humane care of patients.

Metatags are currently being applied to all digitized videos, allowing users to search the content of the videos. When users searches the video collection they will be taken directly to the section within the video stream that contains the concept that they searched. For example, if a user posts a search for "acupuncture," the user will not just receive the videos that are cataloged with the descriptor, but will be linked to the specific sections within the videos that discuss acupuncture. This feature should be available later this year. To assist the consumer in accessing the digitized videos, links to the video collection will be added to the online public access catalog.

To date, The Health Library only has anecdotal measurement of the effectiveness of the streaming media resources, but all the feedback received has been positive. The number of Web server hits to the streamed media page continues to increase along with email requests regarding new titles and requests by media providers for The Health Library to develop new material. Clinicians and researchers now solicit The Health Library to sponsor their presentations with the knowledge that they will be digitized and streamed from the Website.

For more information or to access The Health Library video collection, go to healthlibrary.stanford.edu/catalog/realvideo.html. You may download RealPlayerG2 at www.realplayer.com.