Week 13 Reflection: Computer Assisted Language Testing

The theme this week was computer assisted language testing, a topic I'm glad we covered seeing as I will be taking Assessment in the fall. From the readings this week, as well as the questions posed in the slideshare, I feel like computerized testing is more of an advantage than a disadvantage--at least in most contexts.

I like computerized testing for several reasons. It can be practical (especially for scoring purposes), and computerized graphics can make images look cleaner, which can be calming for test takers. Computers also make adaptive testing possible, which can present a more precise score.

After this week, I've come to realize that validity, reliability, and practicality are of most importance when testing. Validity, I feel, comes down to how the questions are written and less so on the actual mode of presentation (i.e., computer-based or paper-based). Technology, if inefficient or faulty, can greatly affect reliability, so having backup plans for technological failure is a must. Practicality, on the other hands, seems to come down to context. That is, a computer-based test might be great for digital natives, but someone of my mom's generation, for instance, might have a harder time. To combat this potential issue, tests should include tutorials including a practice question or two to ensure learners can properly use the technology.

Previously I mentioned computer-adaptive testing. When I first heard of this type of testing, it was in relation to the GRE, which one of my friends had just taken. At that time, the thought terrified me in the sense that if the test taker is doing really well, they will be getting more and more difficult questions. Now that I am looking into taking the GRE for my PhD applications (and at the same time I am receiving this intro to assessment), I feel that I understand the concept a little better. Though I am sure taking the GRE will not rank anywhere near my top hundred things to do, I can appreciate the intent behind it--to be an accurate representation of where I stand in regard to the items being tested, and this accuracy should be the goal for any test.