Testing, testing. Is this a test? Yes, IT is

In his short but concise paper, Alderson talks about an emerging trend in education - using computer and information technology (i.e. the Internet) for testing. Computers are beginning to deliver language tests in many settings, from high stakes ones like TOEFL to low stakes ones like in-class quizzes and like all IT tools, there are lots of pros and just as many cons.

First, the disadvantages. While multiple-choice and gap-filling assessments are ubiquitous on computers, open-ended items and speaking and writing tests are still a challenge for machines. Also, there might be those who are still too IT-illiterate to take a test online on a computer. In fact, Educational Testing Service provides computer literacy training to TOEFL-takers via sampler CDs or real-life training. Sounds unbelievable but true.

The advantages, on the other hand, are numerous. One obvious advantage is that computer-based testing eliminates the need to fix a date and location as per pencil-and-paper assessments. Results can be immediately available afterwards and Internet-based tests are even more flexible as items and the difficult level can be readily amended and calibrated. These tests can also be saved onto a database of items for future use and test security can be enhanced by randomly placing these items on a paper. While there's always a security risk of running a high-stakes test like TOEFL online, teachers can use IT liberally for low-stakes ones.

Alderson then proceeds to the pedagogical advantages of computer-based testing (CBT), which coincidentally echos the Gaskell and Cobb article on feedback from the previous week. Alderson says that CBT allows test takers to receive immediate feedback that may be conducive to learning, even if they are only concerned about the final grade. Another advantage in pedagogy is that real-time instructions and tools (i.e. online dictionaries) can be given to minimize confusion. Students could also be asked to self-assess and re-take a test with different levels of difficulty as to improve the final score.

While there have been sparse updatees on recent innovations in CBT, things are looking promising with pictures and graphics being inserted into the computer version of TOEFL as well as keys that allow learners to allow how soon the next question appears. Better yet, an EU-funded program called DIALANG provides tests in different languages and assess learners' productive skills like the all-important speaking. Okay, so real-life recording is still not possible but users can choose their preferred response to simulated scenarios. Answers will then be rated by human judges.

The article then goes on to talk about other new programs which I will not list here. Bottom line is, there are numerous advantages and disadvantages to CBT and the key is to do more research on the validity of computer tests. What strategies do test takers use? What is the most appropriate way to measure a learner's performance? Which feedback method is the best (Cobb again) and what is the most effective integration method? While it is not difficult to run a CBT (MCs-Scantron being the most common), educators should take into account the amount of learning generated by CBT as well as test takers' attitude. Will their behavior affect the reliability of the test in any way? Are we really testing what we are supposed to test (validity)? Are we using machines for our own convenience or for the sake of our students? Can computer testing help relate assessment more closely to learning? These issues and more should be addressed through research.

As for my school, only the Science and Math teachers use Scantron, which, I know, is a very primitive version of CBT. But according to Alderson, it is making a comeback and my colleagues have been using it for a few years. Feedback has been good so far, mainly due to convenience, and students have gotten used to it. To them, a Scantron test is just as good as a pen-and-paper one. Again, the only thing they care about is the final score. Here's where self-assessment tools could help but I don't think my school's going to implement any of them due to a lack of funds.