I have a rather general question : How to figure out what to test ? And how to design a test ?
Let say you have a theory and/or implementation of some aspect of human behavior.
The question then become practical : How to make a test ?
The problem is I have zero experience in psychological testing. I can see some examples
mentioned in papers and books, but it is very hard to extrapolate from them.
For example I want to test concepts/category formation. I have some ideas and I can do the coding
i.e. I can use Nengo SP or code symbolically the ideas or theories I may have.
The question then arises how to design a test ?
Or if you go the other way around i.e. from test =to=> theory, is there a such a place as
psychological test REPOSITORY (like we have) where I can read about testing specific human capabilities.
F.e. I often read about 7+/-2 chunk idea or the limits on humans remembering list of items or … or …
but can’t find synthesized test suites which I can reproduce.
In two words WHERE IS THE REPOSITORY 
I got the idea for this question reading : Concepts as Semantic Pointers: A Framework and
Computational Model, which is an example of typical paper, where you almost can figure out how to
make a test … almost …