Re: DSM: (Almost) errorless measurement

Marko Koskinen (
Wed, 08 Nov 2000 22:45:10 +0200

I agree. One of my fellow students is doing her graduate research about
her children. She is repeating the test by Piaget, but doing the test
using situations and materials that are common to her children so that
the children really understand what's going on and they really have a
problem that they can relate to. When Piaget did his tests they were
kind of random and didn't nescessarily have anything to do with the
world view of the child, so the children really couldn't understand the
problems and didn't pass the tests. But my fellow student has come to a
conclusion that the tests really aren't even very hard problems for her
kids when they are set in a common situation and have meaning to the

So I could assume that the key issue here is motivation. If the problem
is really important to you and you really want to solve it, you use much
more attention and effort in solving it, thus the probablility to pass a
test with "real" problems becomes much more probable than solving
"hypothetical" problems. This means that the "school test" actually have
no value at all, because they test only "school survival", not real life
skills and "global knowledge".

Also there is a lot of research in the field of transfer, meaning how
the learner can use the things learned in a new situation, stating that
learning is very much situated, and there is great difficulty in
transfering the knowledge in totally different kinds of environment.
There is also great difficulty in making instruction into transferable

This again means that if you learn something at school and know it in
the test, it is very likely that you have no means of using that
knowledge in real life situations, thus making the test results totally
useless and meaningless.

There is much discussion about these issues in Sudbury literature and
there is also a lot of research supporting the assumptions that are made
in the literature.

Marko Koskinen

> the main problem is with the tests themselves. They don't measure real
> knowledge. Real command of knowledge only shows itself in the presence of a
> true problem. A true problem is not one in which the answer is known and must
> be picked from a pile (That is just an exercise), but one in which the answer
> doesn't yet exist and must be discovered or created. You may eliminate guessing
> from computer-graded standardized tests, but the only way you can measure true
> knowledge is by how a person uses the knowns to find the unknowns. And thus the
> unknowns must be truly unknown to the person, and not just one of a pile of
> possibles.

This archive was generated by hypermail 2.0b3 on Thu Nov 09 2000 - 19:57:27 EST