We started with the multiple-choice format as it's a good 90-10 solution in terms of technology, but we now have free=response questions that are automatically evaluated and can handle fairly complex mathematical expressions - man, was that a lot of work!
But we've found the concern over the multiple-choice format to be overblown. People like to believe they can outguess a multiple-choice question by being clever, but that's not reality on our system or elsewhere such as the AP exams, the AMC exams, or the GRE Mathematics Subject Exam.
Later this fall we're going to be introducing a UI for constructing proofs that's looking really cool and should take things up a notch for the more abstract subjects like Abstract Algebra and Real Analysis. Teaching university-level proof techniques is extremely challenging and time-consuming process (most never really get it) even for undergraduate math majors at university, but I think our new tech will make it much less painful and with a much higher success rate.
> free=response questions that are automatically evaluated and can handle fairly complex mathematical expressions - man, was that a lot of work!
Wow, that's excellent. I can imagine that would have been a lot of work.
> People like to believe they can outguess a multiple-choice question by being clever
Personally I think it adds a bit more complexity and toughness if I can't see the answer in advance. But that is purely my individual style and opinion and I have not seen any research either proving or disproving my hypothesis.
But we've found the concern over the multiple-choice format to be overblown. People like to believe they can outguess a multiple-choice question by being clever, but that's not reality on our system or elsewhere such as the AP exams, the AMC exams, or the GRE Mathematics Subject Exam.
Later this fall we're going to be introducing a UI for constructing proofs that's looking really cool and should take things up a notch for the more abstract subjects like Abstract Algebra and Real Analysis. Teaching university-level proof techniques is extremely challenging and time-consuming process (most never really get it) even for undergraduate math majors at university, but I think our new tech will make it much less painful and with a much higher success rate.