As an instructor I have have had little to no experience with computer testing. As a student I have only taken computerized versions of the GRE. At the time that I took the SAT/ACT there was only a paper version available. When I took the GRE in Germany the adaptive model was a novel approach to test taking, but adding the computerized version did allow me to take the test while living overseas at any time which was convenient for me.
The adaptive format was brand new and I had to re-learn how to approach a standardized test. Previous instruction had taught me to work through the pencil and paper version in a certain way: answering questions that I knew and then circling back to the ones that I wasn't sure about before the time ran out. We were also thoroughly versed in the tactics of guessing vs. not guessing depending on the scoring model for each test. Neither of these strategies was possible with the new format where questions must be answered before the next one is presented. While both formats have a timing element to them, the computerized version seemed more nerve-wracking and gave the illusion that I had less control over how my scores turned out. When I could work through the questions and choose which ones I knew, it gave me good overview of how well (or poorly) I was doing. When the computer was choosing which question I answered next, it took away some of that feeling of control and, as several others have mentioned, made me constantly question how well I was doing.
The element of instant scoring is very nice and is certainly a cost saver and, as Chapelle mentioned, since technology is becoming such a salient aspect of communication and ability to function in the society, it makes sense that it be included as an aspect of language testing. I know some other instructors who have used computer-aided testing in their courses although I have not. Their general assessment seems to be that it is helpful when ther is a large number of students to test, but varying levels of computer savvyness on the part of their students makes for a constant stream of trouble shooting in order to get the same uniform results that can be obtained with face-to-face testing.
2 comments:
I guess I've still not been added as a contributor, either, so here goes my comment.
I've had very little experience with computer-aided testing, except for a few tests here and there that were run through ANGEL.
I agree that the time savings are nice in some instances, but the parts that can be easily graded by a computer (vocabulary, grammar, multiple choice and the like) are generally also the things that are quick for me to correct when my students take paper tests/quizzes, too. It's the sections of assessments that have grey areas that take the longest for me (short answer, essay), and these, as we read in the article, are not yet able to be perfectly assessed by a computer.
I'm not sure what I think of computer-aided testing for language learners. If students are able to pick a multiple-choice option for a vocabulary word, do they really know the word, or did they guess correctly? If the test isn't multiple choice and students have to write vocab words out, as an example...does this save anyone any time in terms of assessment? I do like the paper savings, however...
I understand your concerns, but I really like the idea of computer-adaptive testing. Granted we need to know more to make them much better. Yet, they have great potential. With them I could see where people are actually at without wasting anyone's time. If you are at a high proficiency you do not need to answer all the lower level questions, if you are at a low proficiency, the test will stop after your ceiling has been reached. That makes sense and I would love to have materials that way, too.
Post a Comment