next up previous contents
Next: Colorado Component of Up: USER INTERFACE EVALUATION Previous: USER INTERFACE EVALUATION

UCSB Component of the Evaluation Team

Membership: Hill (leader), Carver, Dolin, Frew, Larsgaard, Nideffer, Rae, Simpson

Mission Statement of Team: The goal of the UCSB component of the Interface Evaluation Team is to evaluate the effectiveness of the ADL system from the perspective of potential users of the system. Knowledge gained from these evaluation activities is used (1) to inform the design and implementation of the ADL system on the Web, and (2) to document in detail the effectiveness of the ADL and areas calling for improvement, both in the interface design and in the underlying system functionality and content.

Research activities over the last year:

  1. Results of Beta Tester Evaluation of the Web Interface
    The team wrote a report on the findings of the user evaluation of the Web Interface. During the study period, user feedback was collected through various formal and informal methods. These included online surveys, beta tester registration, ethnographic studies of ADL users and potential users, target user group focus sessions, and user feedback comments while using the interfaces. The paper briefly describes the evaluation studies conducted and what was learned about user characteristics and about the study approaches themselves. User reactions to the ADL interface and to the functionality and content of ADL are summarized. Finally, the value of these findings to design and implementation decisions is considered. The citation to the report and to an electronic version is:

    Hill, Linda L., Dolin, Ron, Frew, James, Kemp, Randall B., Larsgaard, Mary, Montello, Daniel R., Rae, Mary-Anna, and Jason Simpson. (1997). User evaluation: Summary of the methodologies and results for the Alexandria Digital Library, University of California at Santa Barbara. Proceedings of the 60th Annual Meeting of the American Society for Information Science, November 1-6, 1997, Washington, D.C. http://www.asis.org/annual-97/alexia.htm

  2. ADL Design Review Panel Meeting, February 19-21, 1997: Final Report
    A report of the Second Alexandria Design Review Panel on February 19-21, 1997 was written and posted on the ADL internal web page. The workshop was to review user requirements for the Alexandria Digital Library and to discuss plans for testbed development and research progress in the coming eighteen months. The panel recognized that the challenge to build an operational digital library testbed is improved by the research accomplishments of the ADL team and that it is important to recognize that some research questions cannot be addressed in the absence of an operational testbed. This balance of emphasis will continue to be a concern through the rest of the project.

    Thirty-two representatives from the public sector, private sector, and academia came from the U.S. and one from Canada for two-and-a-half days to participate in plenary and small group discussions, software demonstrations, and informal discourse.

    In preparation for the meeting, the Steering Committee agreed upon a set of objectives to guide the content of the agenda. A three-part focus was established to address issues of search, browse, and retrieval; content and processing; and interface and navigation. These three areas encompass a number of the more difficult problems to be resolved immediately. Their selection was based primarily upon feedback gained in focus groups, survey questionnaires, and videotaped or recorded interactions with actual and potential digital library patrons over the first two years of the Alexandria Project. Many specific topics were discussed including ways for users to view the contents of the library (an overview of the holdings), giving the user a ``road map'' of the services of the library, visualization (presentation) of search results, support for iterative search processes, and customization of the user interface.

    From the wide range of comments and recommendations contributed by the Panelists that are detailed in the report, some strong themes can be identified. First, the Panel is supportive of the efforts of the research teams working at Santa Barbara and at Colorado. Panelists and funding agency representatives agreed that keeping a careful balance between developing an operational digital library and contributing to theoretical and applied research is necessary. The importance of the content of the library was strongly supported, as were the enabling technologies for distributed ingesting of data with attention to the representation of data provenance and verification. Collection development criteria were discussed at length, with emphasis on focusing on particular geographic areas and collecting all forms of georeferenced information about that place. Panelists recommended that ADL develop enabling technologies and specifications for connecting local collections to ADL, with distributed searching capability.

    In connection with ADL user interface design, the implementation of user profiles was discussed with ideas for multiple profiles per person and for the way in which the profiles could influence the interface presentation and conduct of the search. Support for a method of providing ``something with a heartbeat'' to be available to assist in the use of ADL was expressed in several ways, with the related recommendation that ADL aim to create a professional tool for information specialists as well as for other user communities. A discussion of measures of success for ADL focused on grounding those measures in the user's work environment and information needs.

    ADL Staff identified four areas in which the Design Review has had an immediate impact: (1) collection development strategy, (2) user profile strategy, (3) development of a strategy to provide human support to digital library users, and (4) the design of the new interface.

    The overall thread that runs through the recommendations is that ADL needs to focus better on what can be accomplished in the short time left in the current phase of the project. Most panelists reported that the format of the Workshop worked well, but they would like for it to have been more focused and they expect to see specific responses from ADL on the recommendations made.

  3. ADL System Requirements for Fall Release (first draft of May 20, 1997)
    The results of the beta tester study and the Design Review were used to develop an initial set of user requirements for the design of the new ADL Java-based interface. The initial set of requirements was further refined by the Interface Development Team, the Implementation Team, and the Evaluation Team. The following list of requirements came directly out of the evaluation studies and is included to illustrate the application of evaluation results. This should not be interpreted as the final requirements list.

    1. Search Functions
      1. The system shall present a unified search screen to the user (integration of gazetteer, catalog and map-based searching)

      2. Reduced, simplified set of search types for metadata attributes (search buckets); map to ADL and MARC schema

      3. The user shall be able to select search areas on map (rather than use whole map window as search area)

      4. User shall be able to search non-contiguous spatial areas
    2. Session Management
      1. A summary view of the session log shall be viewable by the user during the session. (query history and corresponding results sets)

      2. Session log shall be available to ADL help desk personnel.

      3. User shall be able to stop a query and issue a new one.

      4. User shall be able to return to a previous point in the session.

      5. User shall be able to modify and reissue a query.

    3. Result Display
      1. The total number of items in the results set shall be displayed to the user.

      2. User shall be able to sort result set by genre, date and ?

      3. System shall display footprint distribution of a result set.

      4. Set of display formats for metadata defined and implemented.

      5. Low resolution browse images that can be enlarged for evaluation.

    4. User Workspace

      1. The system shall provide a user workspace where selected items can be collected in a set of user-named collections.

      2. User-saved queries may be re-issued to the system during a later session.

    5. Holdings Visualization
      1. The system shall present the user with an overview of the contents of collections by geographic coverage, genre and date.

    6. Gazetteer

      1. The system shall present the user with an overview of the contents of the gazetteer by geographic coverage and feature type.

      2. The system shall allow storage of user-named geographic areas in a 'personal gazetteer'.

    7. User Help Functions

      1. Search examples on the interface itself to aid user in creating a query.

      2. Context-sensitive help

      3. Tutorial with FAQ

      4. ``What's new in the library'' for announcement of new collections in ADL.

    8. User Registration

      1. Optional user registration form

    9. Usability Features

  4. Status of process indicator

    1. Data Ingest

    2. Data Distribution

      1. Screen on user interface for data ordering

      2. Slice and dice large files for delivery

  5. User Scenarios

    The Evaluation Teams continued to develop a list of the ways in which users would want to use a system such as the Alexandria Digital Library. This list of 58 scenarios, derived from ADL Target Group Members, ADL documents, Global Change Data and Information System user queries, GIPSY report (Woodruff), Petroleum Abstracts indexing staff, and analysis of the ADL Rapid Prototype questionnaire replies by the University of Colorado Team is available in the ADL User Scenarios at http://www.alexandria.ucsb.edu/ lhill/scenarios.html.

  6. Design of Components of the New ADL Interface

    The UCSB Evaluation Team designed the components for the User Registration form and the Exit Poll form for the new interface. These designs are based on prior experience with the beta web interface and on the evaluation research questions identified by the Team. The registration form collects 16 pieces of information about each user, most of these are selections from lists of possible values. The Evaluation Team will be able to categorize user groups based on these values. The Exit Poll appears each time that a user exits the ADL interface and asks some simple questions: what was your purpose in using ADL; usability of the interface; usability of the content; and a text box for comments, suggestion and questions.

    The UCSB Evaluation Team has also contributed to the design of the help components for the user interface, in particular the tutorial, the glossary, the Frequently Asked Questions, and the system to receive, categorize and respond to user comments and questions through email. Also recommended several approaches to letting users know that their comments and questions have been received, including an immediate reply giving an idea of the expected wait interval and giving a number to call for follow-up, if necessary.

  7. Policy on Permitting/Supporting Evaluation Studies of the Alexandria Digital Library

    The following policy was developed by the UCSB Evaluation Team and approved by the ADL Executive Committee, 11/24/97:

    Requests are received to use the Alexandria Digital Library (ADL) interface/system for evaluation studies from time to time. It is in the interest of ADL to encourage and support formal evaluation studies, especially if the results are shared with ADL. However, such evaluation studies may also require support from the ADL Team that will have an impact on staff resources. The purpose of this policy is to encourage good evaluation research studies based on the use of ADL while controlling the impact on ADL staff.

    Procedure:

    1. Applications to evaluate the ADL interface should be made to the Evaluation Team by email to alex-uie@alexandria.ucsb.edu.
    2. Applicants will need to provide descriptions of their studies that include a. the name of the principal researcher, organization, purpose of the study, and contact information b. the number and type of participants c. the time frame of the study d. the equipment and network environment for the study, and e. the research questions, data collection, and analysis plans.
    3. ADL may propose revisions to the study plan if the plan does not take into account a. the purpose of ADL b. the status of the interface c. the collections and services available, and d. the support needed from ADL staff to support the study.
    4. The ADL Evaluation Team will accept or reject evaluation requests based on the quality of the research proposal and the ability of the ADL staff to provide the support required in the time frame requested.
    5. The ADL Executive Committee will give the final approval based on the Evaluation Team recommendation.
    6. To protect the privacy of user session logs, evaluators must comply with all relevant legal requirements for informed consent before ADL will provide access to the session logs of study participants.
    7. A copy of the final report and the study data in electronic form for all completed evaluation studies must be submitted to the ADL Evaluation Team; these will be made available through the ADL internal web.

  8. Internal Evaluation of New ADL Interface

    The Evaluation Team held the first internal evaluation of the new ADL interface design on August 19-21, using volunteers from the Project and from the UCSB campus. Twenty-three testers participated by using the interface for 30-60 minutes while they were observed and videotaped. They were given the following instructions:

    1. Find what the library holds for the Arizona / Mexico border and put the most recent 3 aerial photographs in your workspace in a folder named ``Arizona/Mexico Border''

      Redo the search for the same area, but limit the search to aerial photographs taken from 1990 to date.

    2. Find remote sensing images for the San Francisco Bay area. Display a browse image from the retrieved set on the map and zoom in to see it in more detail.

    3. Find what the library has for the Santa Cruz Island. Display the browse image for the Digital Elevation Model (DEM) on map and zoom in for a better view.

    4. Free time: Spend the rest of your time exploring and using the interface.

    A report was written of the results of the evaluation and discussed with the Interface Evaluation Team. They used the report to guide further interface development.

    The second internal evaluation period was scheduled for three days, October 20-22. Thirty-one people were scheduled to test the interface. After two days of testing with 21 people participating, the evaluations scheduled for Wednesday were cancelled. We had, at that point, collected a great deal of information from the testing that had been done so far. Another factor in the decision to cut the evaluation short was that the system had been unstable during the evaluation period.

    The registration process and the exit poll were tested as well as the use of the interface for searching and retrieval. The process was similar to the first round, except that each person (or group of persons) was given an hour to use the system. They were videotaped and observed. Each tester was given the following set of scenarios to use, but were also allowed to do their own explorations if they preferred.

    Scenario 1:

    A. Find waterfall locations in Northern California (search the gazetteer using 'falls' as the type). Set limit for the retrieval to 30 on the Submit Search page.

    B. Locate Burney Falls. Draw an area around this location and find out what is available in the ADL catalog about this area. Caution: Erase the first target area.

    C. Review the thumbnail images and the metadata. Select one or two to store in your workspace. Make a folder for the things you save and name it.

    D. Try (1) resorting the result sets, (2) pointing to a footprint and finding the Results listing for it, viewing different parts of the metadata, and (3) displaying the thumbnail images on the map itself.

    E. Resubmit the query but limit the results to tiff images only (search the catalog using 'image'- 'tiff' as the Available-as Type.

    F. Search the gazetteer directly for Burney to find Burney Falls.

    Scenario 2:

    You'd like to learn more about possible anchoring sites for the Channel Islands where you could moor your boat.

    A. Select the target area and then search the gazetteer and see what you can find.

    B. Select an anchorage area and search the catalog for information about that area. Limit the date to 1980 to current. Select some items that look good and put them in a fold named ``Channel Islands'' in your workspace.

    C. Store the contents of your workspace so that you can have access to it later.

    Scenario 3:

    For the Western United States, find what the library has that was published by the California Division of Mines and Geology.

    Scenario 4:

    You want to find a particular item in the catalog and you remember the record number. The number is 113 - set the target area to the world.

    Testers were asked to talk aloud as much as possible while the ``observers'' took notes. Several testers also provided comments through the exit poll. Our testers ranged from those who are very familiar with ADL to those who are only vaguely aware of what its purpose and current status is. Some testers also participated in the first round of testing.

    The report of the evaluation sessions contained comments on the registration process; general comments; comments on the tutorial/help sections; searching comments categorized by unpredictable results, free-text searching, setting limit on size of result set, display of search strategy, stopping query processing, search buckets; multiple windows management; using the map browser; sorting and understanding result sets; use of the Workspace Window; metadata display; and the exit poll. The Interface Design Team used the report for guidance in their continuing design process. The Evaluation Team also compared the functionality of the interface to the guiding user requirements.

  9. Classroom Evaluations of New ADL Interface
    Following the internal evaluations, the team invited several professors to incorporate ADL within their curricula in order to observe student use of ADL and evaluate the potential of ADL as a learning experience. One professor was able to include some ADL-related assignments within the laboratory sessions of his course. The data collection included videotapes of the lectures and the laboratory sessions; interviews of the professor, teacher assistants, and some of the students; and observations made while participating in some of the activities. The data are being analyzed from an ethnographic perspective, using discourse analysis to identify patterns of use and opportunities for learning. Immediate feedback was given to the interface and system designers to inform continuing development.

Abstracts of Published Papers

[1]
Hill, Linda L., Dolin, Ron, Frew, James, Kemp, Randall B., Larsgaard, Mary, Montello, Daniel R., Rae, Mary-Anna, and Jason Simpson. (1997). User evaluation: Summary of the methodologies and results for the Alexandria Digital Library, University of California at Santa Barbara. Proceedings of the 60th Annual Meeting of the American Society for Information Science, November 1-6, 1997, Washington, D.C. http://www.asis.org/annual-97/alexia.htm

Alexandria Digital Library (ADL) is one of the six digital library projects funded by NSF, DARPA, and NASA. ADL's collections and services focus on geospatial information: maps, images, georeferenced data sets and texts, and other information sources with links to geographic locations. Throughout the project, user feedback has been collected through various formal and informal methods. These include online surveys, beta tester registration, ethnographic studies of ADL users and potential users, target user group focus sessions, and user feedback comments while using the interfaces. This paper briefly describes the evaluation studies conducted and what was learned about user characteristics and about the study approaches themselves. User reactions to the ADL interface and to the functionality and content of ADL are summarized. Finally, the value of these findings to design and implementation decisions is considered.



next up previous contents
Next: Colorado Component of Up: USER INTERFACE EVALUATION Previous: USER INTERFACE EVALUATION



Terence R. Smith
Tue Jul 21 09:26:42 PDT 1998