This document is a work in progress, but we felt it was important to make the information availalble while we are working on the completed report. If you have any questions or comments please email Ron Stewart

Accessibility of Online Databases
A Usability Study of Research Databases

Ron Stewart, Director
Northwest Center for Technology Access
Oregon State University



For every student a crucial component to effective participation in the modern post secondary education system is the ability to retrieve, use and apply research based information in their chosen field of study. In today's technologically rich education environment this content is most typically found in a variety of "Web" based research databases that are highly proprietary in nature, and presented in a very graphical environment that can be challenging to use and interact with. For users with print related disabilities this environment presents an additional challenge and may be impossible to use if the adaptive technology that they must have to "surf the web" does not prove to be effective in accessing the content of these database sites.

For the last several years the Technology Access Program (TAP) has conducted an informal evaluation of the online research databases that the students of Oregon State University normally encounter as a part of their program of study. In the first year of examination 1997, approximately 60% of the databases had some issues with accessibility, but could be used by most students with print disabilities with appropriate adaptive technologies due to a predominately text based delivery medium provided via TELNET from the OSU Library. Most of the database systems were provided from a CD Server housed in the library and if the student could not interact with the database effectively over a network connection then we did have the option of using the database CD on an adaptive workstation in the library. While this system was cumbersome and limited in its flexibility, access for the individual with a printer related disability could be provided.

By 1999 things had significantly changed for the worse, along with an almost total movement to Web based delivery the research databases had also started using script based transaction protocols that presented an insurmountable obstacle to the adaptive technologies that were currently available to students. In this third evaluation, 95% of the databases failed the accessibility evaluation, those that did not fail could be accessed directly on CD-Rom or through a text based TELNET connection. The testing was done with an audio web browser PWWebSpeak and text based browser called LYNX. The primary areas of failure could be directly attributed to an inability of the adaptive technologies to interact effectively with the JAVA and CGI scripts that were used to login to the site, query the database, and structure the returned content. This state of affairs prompted us to write (the letter) to all of the major database providers none of which replied to our inquiry or a second letter that was sent out by our University Librarian. It is important to note that one of the most significant issues was the ability to even login into the database, a problem that has been effectively eliminate by the common use of gateway sites by most major schools.

Following the disclosure of these finding to the Adaptive Technology Community we were asked by many of our peers to conduct a formal evaluation of the state of accessibility of these online databases. The purpose of this paper is to present these findings as well as the methodology that was used to conduct what we feel is a very comprehensive overview of the major database systems and their accessibility with a variety of adaptive software products in common use by persons with print related disabilities. It was important to us to insure that this evaluation was conducted in as realistic a situation as possible, using real world examples commonly encountered by our students in their normal course of study at OSU. This project was the cooperative effort of the Technology Access Program and the Subject Librarians of the OSU Library.

This report is the first of what we hope will be many generated on the the interface between disability, technology and the applications of the concepts of "Universal Design" as it applies to the technological systems used in post secondary education. The current movement to intergrate the issues of "Accessibility" into the mainstream of our modern educational system is a very positive step forward, but as this results of this paper will reflect, the gap between "Functional Usability" and basic "Comformity" with a standardized set of benchmarks can be a large one.


All of the major research databases found on the Oregon State University Library website were evaluated. These databases represent 120 content specific resources produced by 37 different vendors. Review Appendix1, for a comprehensive listing and specific comments by database vendor in regards to the database products they produce, their overall accessibility, and any atypical procedures that were required to use the database or conduct the evaluations. In most instances OSU subscibes to the full text version of these databases and the results are based on this configuration. Other more limited configurations of the databases are available and in post instances these products deliver image based PDF document files which present an increased level of inaccessiblity to the user. These documents may require secondary intervention or a more complex conversion process before the products of the database output can be fully utalized by the researcher.

The evaluation was conducted on a Pentium 4 computer system running Windows 2000 Pro, and is reflective of the Windows based workstations present in most of OSU's campus computer labs. The databases were evaluated from a network based connection, with the screen readers being tested using Internet Explorer v6.0. A second test was also done using a remote connection through the OSU proxy server. All of the latest service packs have been installed on the workstation and all the adaptive software had been updated to the most current version at the time of the final evaluation. A follow up evaluation will most likely be conducted using other operating systems and browsers, but it is our assumption based on extensive experience in this area of investigation that accessiblity will decrease as we move away from the test environment used in this study.

This database evaluation was conducted using the following adaptive technology products. This listing represents what we feel to be a comparative presentation of the leading Windows based adaptive software suited to the purposes of this evaluation. The focus of this evaluation was on adaptive software products designed primarily for blindness since it is this area of disability that presents the greatest challenge for the database vendors, and for our users with disabilities. A second evaluation using alternative input devices will most likely follow.

Audio Browsers:

Screen Readers:

These products were used in a default configuration without any modification. This method was selected to reflect the typical situation in our campus computer labs which do not allow for custom configuration or user modification of the installed software products. We understand that increased functionality could be obtained in most of these products by the use of scripting or custom settings. It was our belief that unmodified use would present the most objective evaluation, and provides a default baseline which is most usable for the database vendors in correcting the faults in their product presentation. The evaluators of the research web sites were users with a basic level of familiarity with the various adaptive software products and their use and functionality in accessing web based resources. This type of user is also most reflective of the typical student with a disability that we encounter at Oregon State.

Each database evaluation was conducted using preestablished criteria based on a task analysis of the function steps involved in conducting a database query and in both basic and advanced search modalaties. The criteria and a brief explanation of what was the desired measurements are found in Appendix3.

Each of the research databases was evaluated with the adaptive software products using a typical research task reflective of those commonly undertaken by undergraduate students at Oregon State University. These research activities were provided to us by the Subject Librarians of the OSU library from sample activities commonly used in content based courses that include an introduction to research component.

It was not the purpose of this evaluation to validate compliance with Section 508, but to test for functional usability of the various research web site by students with print related disabilities using common adaptive software. However we did do a subsequent 508 based review of the five most used databases using IBM HPR and the accessibility criteria found in Appendix4.