My Courses
syllabus Discussion E-mail resources previous forward

On the Horizon: Large-Scale, High-Stakes Online Testing

Educators should be aware of state efforts to develop online environments for large-scale, high-stakes testing. Even though states may take the lead in developing testing systems, they will dramatically impact the technology infrastructure and personnel of every school and district within that state. Use of a statewide system will still require networking, hardware, software, training, maintenance, and support. Individual districts should be aware that many of these functions and their associated costs may impact their own technology budgets.

Some certification and achievement tests have already been adopted for online presentation, such as the TOEFL (Test of English as a Foreign Language), the GMAT (Graduate Management Admission Test), and parts of the GRE (Graduate Record Examinations) and National Board for Professional Teaching Standards Assessments (NBTS). Several states have taken an early lead in the development of online systems for administering annual high-stakes tests. Much like the integration patterns of individual teachers, these early efforts focus on replicating existing practices and rely on questioning formats most often applied to paper-and-pencil tests.

The National Online Assessment Conference in Washington, DC, on November 20, 2000, was organized by AEL for the U.S. Department of Education. It featured panel discussions and presentations by key representatives in assessment, policy, and technology, and included online follow-up support. States face many issues concerning the deployment of online testing, which are presented below.

  • Access. While many schools have reduced their student to computer average to close to 5:1, this ratio is inadequate for large-scale testing. This ratio also includes many computers that may not meet the performance requirements of new testing systems.
  • Equity. Not only must tests be designed to measure curriculum standards, but variations in presentation may affect the validity and reliability of results. Factors such as display size and resolution impact how much information students see and may create disparate testing environments. Differences in computer platforms and processor speeds also impact how students may view testing information. If pilot tests are presented to select, small populations, will these results be equitable and generalize to all students that would evenutally be required to use the system?
  • Infrastructure. While many schools are connected to the Internet, there are no standards governing how many computers are connected and access speeds. If entire schools or districts access a network at the same time, connection speed will slowed down, which can erode test reliability. Some states are investigating server farms that download test items to a local server, but this poses threats to security.
  • Security. Not only must test items and student scores be kept secure, but schools must be able to identify students taking the tests. Finance and business experiences contribute to developing the most secure environments possible, while several biometric approaches are being explored to supplement visual identification of test subjects.

Did You Know?

The Educational Testing Service Network (ETS) is a major force in the research and development of large-scale, high-stakes testing for numerous purposes. Randy Bennett from ETS attended the National Online Assessment Conference and shared some of his research in the development of online test presentations. Bennett's paper, How the Internet Will Help Large-Scale Assessment Reinvent Itself, discusses many of the issues surrounding this topic in greater detail.