administration times (Shermis et al., 1996), and make possible relatively low-cost scaling to large numbers of test takers. They could also be designed to meet the needs of special populations, including people with physical disabilities and people from diverse cultural or linguistic backgrounds (Naglieri et al., 2004).
However, there are legitimate concerns about using computers in educational testing. A potential limitation, of course, is the lack of computer literacy of the test population. Test takers—children or adults— who do not have at least a basic familiarity with computers and computer keyboarding may not perform as well as those who have at least basic computer skills (Russell, 1999). In addition, requirements for computer memory and processing speeds, graphics quality, and bandwidth— for applications using the Internet—may pose significant cost and resource barriers.
There are legitimate concerns about using computers in educational testing.
Computer-based tests would be just as susceptible to cheating as traditional paper-and-pencil assessments, although the types of cheating and strategies for countering them may differ. For example, someone other than the registered examinee could take the test or help answer questions on an assessment administered remotely (online). To preclude this kind of cheating, authentication could be attempted using a biometric measure (e.g., a fingerprint or retina scan), or the test taker could be required to take a short, proctored confirmatory test (Segall, 2001).
It is important to keep in mind that although computer technology could potentially increase testing flexibility, authenticity, efficiency, and accuracy, computer-based assessments must still be subject to the same defensible standards as paper-and-pencil assessments, particularly if the results are used to make important decisions. The reference of choice is Standards for Educational and Psychological Testing (AERA et al., 1999).
The following discussion focuses on aspects of computer-based testing that offer significant potential benefits for the assessment of technological literacy.
Computer-based, flexi-level, branching, and stratified adaptive testing have been investigated for more than 30 years (Baker, 1989; Bunderson et al., 1989; Lord, 1971a,b,c; van der Linden, 1995; Weiss, 1983). Research has been focused mostly on using interactive (computer) technology to select, in real time, specific items to present to individual