Thanks! One main difference is that you'd run all the answers from a section through the system at once with a section adaptive scheme. So with two modules there's only one point at which the material is really being adapted to the test taker (after the first module), and the overall ratings would only be calculated twice (once in the middle, once at the end).
Thanks! There are a lot of ways you could do it. Just having the difficulty rating for a question is very powerful – that combined with a good tagging system can get you pretty far. For example, you could say, "The only 600-level questions you miss are 'wordy' ones." And you can make certain inferences about why a question was missed – the more data you have, the more accurate the inferences are. But it's quite difficult to come up with a tagging system that's intuitive and useful for everyone. I might write about tagging in a future post.
Yes, I think lots of people have. The problem is that the systems often make sense to the creator but not to outside users. But one common approach is word count.
Excellent article. How does IRT differ between section adaptive and question adaptive testing?
Thanks! One main difference is that you'd run all the answers from a section through the system at once with a section adaptive scheme. So with two modules there's only one point at which the material is really being adapted to the test taker (after the first module), and the overall ratings would only be calculated twice (once in the middle, once at the end).
Thanks! There are a lot of ways you could do it. Just having the difficulty rating for a question is very powerful – that combined with a good tagging system can get you pretty far. For example, you could say, "The only 600-level questions you miss are 'wordy' ones." And you can make certain inferences about why a question was missed – the more data you have, the more accurate the inferences are. But it's quite difficult to come up with a tagging system that's intuitive and useful for everyone. I might write about tagging in a future post.
Yes, I think lots of people have. The problem is that the systems often make sense to the creator but not to outside users. But one common approach is word count.