4

Here is a format for an exam style. I would like to know if this has a name in EDU circles and if there's a reference. I've asked a few people in the field and they aren't familiar with it exactly.

First there is a C-Level section These questions are fairly straightforward. Let's say there are fifteen of them and if there are at least ten correct answers (call this "passing the C-section") then the student is guaranteed a C.

Next there is a B-Level section Let's say there are 6 questions here. If the student has "passed the C-section" and has provided 4 correct answers, (call this "passing the B-section") then the student is guaranteed a B.

Finally there is an A-Level section Let's say there are 2 questions here. If the student has "passed the B-section" and has provided 1 correct answer, then the student is guaranteed an A.

All other scenarios would have to be communicated somehow but any version of this is good for my purposes.

I'm happy to share the pros and cons in my experience if anyone is interested but I would like to know if the experts have a name for this type of exam.

1 Answer 1

1

Three terms that spring to mind are "tiered testing", "adaptive testing" and "sequential testing". Each of those terms are used in the education literature to describe examination formats that have some features in common with what you describe but also (potentially) some differences. Your testing arrangement is obviously, at a purely descriptive level, both sequential and tiered although the terms are sometimes used in the literature to describe a sequence of tests that are interspersed with additional learning ... and that doesn't seem to be the case with your description.

For that reason, I think you might find it most useful to consider "adaptive testing" — a term that is usually used in the context of computer-administered examinations, as in the Wikipedia link I have given, but is clearly applicable to your own case, as you will see.

The main difference between what you describe and computerized adaptive testing (CAT) is simply the involvement of a computer. In the case of CAT, the examinee doesn't get to see all the examination questions at once. Instead, they would be presented with the questions from section C in sequence, section C would be scored, and if the student has done well enough, they will then be presented with the questions from section B. Frequently, if the student gains insufficient points to be allowed to progress to section B, they will be presented with additional C-type questions. The purpose of the additional questions is typically not for the purpose of allowing them an additional chance of progressing to section B, but rather to give a better estimate (see, for example Rasch testing) of their position within tier C.

If the examinee is allowed to progress to section B, they will then go through a similar process before being considered for section A.

The process you describe seems to differ from this approach in one minor aspect (i.e., the scoring process being applied only at the end after the student has attempted whatever questions they choose), and one more significant aspect (the lack of a Rasch type refinement of score having failed to progress from one tier to the next).

It's worth adding, that if you actually want to construct such a test and to do so rationally, rather than on a simple ad-hoc basis, then you will need to learn something about item response theory, and or Rasch theory.

For further reading, I suggest looking at "tiered testing" and "sequential testing" (primarily to confirm that these terms are not helpful), and then to look at CAT, and item response theory. You might also look at some of the papers (such as this), which discuss related, though not absolutely identical, ideas.

Added after reading the OP's comment that says: "I personally apply a type of refinement for “in between” scores but it feels more ad-hoc than the C/B/A part so I was hoping to learn more." ...

It is item response theory that will show you how to do that in a way that is not ad-hoc ... but you will need more information that you might think in order to do it properly. Again, have a look at the reading material I've indicated. This paper also provides a nice introduction.

2
  • ‘The main difference between what you describe and computerized adaptive testing (CAT) is simply the involvement of a computer.’ I probably’ll get myself booted, c’est la vie!, but this doesn’t pass my Turing challenge. I’ll be happy to be educated otherwise. Jan 5 at 4:00
  • Thank you @crimsondark I’ll look into those and wait for other responses before immediately awarding an answer. For what it’s worth, I personally apply a type of refinement for “in between” scores but it feels more ad-hoc than the C/B/A part so I was hoping to learn more.
    – cheyne
    Jan 5 at 4:14

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .