I'm making a body of questions for my English students, and I want a dynamic rating system to help kids get the right questions. My idea is that each question, and each student, will have a rating (like chess rating). Every time a student gets a question correct, his/her rating will go up a little bit, and the rating of the question will go down.
The point of this is that instead of having teachers guesstimate what level different questions are, the system will be able to collect statistics about each question, and give only those questions which are within an appropriate range from the student's.
Here's my problem: the questions may be multiple choice, which kind of breaks the zero-sum game effect of say a player-vs-player chess game. If a very low-rated player luckily chooses the answer to a very hard question, or if a high-level student decides to deliberately answer all the questions on an easy test wrongly, it might screw things up.
Any ideas on how to make elo work in this case?
The point of this is that instead of having teachers guesstimate what level different questions are, the system will be able to collect statistics about each question, and give only those questions which are within an appropriate range from the student's.
Here's my problem: the questions may be multiple choice, which kind of breaks the zero-sum game effect of say a player-vs-player chess game. If a very low-rated player luckily chooses the answer to a very hard question, or if a high-level student decides to deliberately answer all the questions on an easy test wrongly, it might screw things up.
Any ideas on how to make elo work in this case?
