Assessment presents special problems in the implementation of Web-Based Instruction. In his chapter 'Testing Learner Outcomes in We-Based Instruction', Hudspeth (1997) identifies some of these problems and suggests possible solutions.
The dynamic nature of Web learning reaources and the use of learner-centred instructional design can cause confusion about what is to be learnt and how it is to be assessed. The virtue of continually updated, authentic information, leading to redefinition of learning goals, creates problems unless there is explicit agreement on what is being tested. Similar problems may arise from changes in the goals and backgrounds of learners on a course, particularly if the learning strategy emphasizes student ownership and direction of learning. Even on traditionally - delivered courses, authenticity of work submitted for assessment is an issue, specially if the work was produced remotely.
Hudspeth's approach to the problem of authenticity is to minimise cheating by:
Such steps can motivate learners to cooperate with online testing as a valid measure of what they have achieved. While this may be true, it seems an inadequate safeguard against cheating in a climate where accreditation rather than understanding is the prime motivation of some learners.
As Hudspeth points out, traditional multi-choice and essay forms are often inappropriate to Web-based courses, partly because of the dynamic nature of the learning resources. He suggests that immediate feedback is important for lower-level skill assessment, but less so for skills such as problem-solving and evaluation. He distinguishes between the requirements for formative assessments, used to provide feedback on enabling objectives and progress, and a 'culminating test' designed to take account of variations in learners' backgrounds and purposes in taking the test.
One approach Hudspeth recommends is the use of a Learning Contract defining an agreement between learner and instructor on what each undertakes to do. Such a contract would specify learning outcomes explicitly, but would not prescribe the means by which they would be attained, enabling the learner to exploit the dynamic information base of the Web.
While Hudspeth's chapter is useful in identifying reasons why traditional assessment methods may not work for online learning, it is clear that much work is needed to develop new methods which will provide a valid measure of learning outcomes in Web environment.
| Guide Home Page | Implementation Guidelines | Specific Applications | Return to Top |