Welcome to the discussion!
I'd like to ask if anyone has any views to support the notion that perhaps
it's not just maths that allows for the randomisation of branch elements,
but any subject where the data randomised is completely objective and
Thinking of my own area, biomedicine, if I asked what was the normal blood
pressure in mmHg for a healthy adult (120/80), is it valid to get the
computer to come up with valid distracters e.g. 80/120, 100,80, 140,100,
etc, you get the idea.
Presently, extended matching questions are generating some interest here.
These are basically (forgive the simplistic definition) a subject, a list or
set of responses and any number of stems. the advantage of extended matching
questions is that the list or set of responses may be very large, typically
20-40 items. Thus:
With regards to the capital cities of countries:
London Paris New York Moscow
Madrid Brussels Stockholm Canberra etc
What is the capital of France?
What is the capital of Spain?
It would seem to me that these kinds of questions are ideally suited to
computer assisted assessment, and particularly computer-generated lists and
Again, in my own area, the response lists could a long list of enzymes, or
muscle, or infections agents, whatever. Single stems can then be rapidly
generated that uses this extended matching response lists.
Is anyone using extended matching questions?
Extended matching questions are a lot more than simple MCQs with 40
branches, and there are sound principles and rules for designing response
sets, but for the purposes of this discussion this analogy will suffice.