Millisecond Forums

I am coding an experiment in which participants learn about two categories of objects then are given an old/new memory task. For the first part of the experiment, participants see half of the total pool of objects and learn that half of those (so a quarter of the total) belong to Category A and the other half (another quarter of the total) belong to Category B. For the second part of the experiment, participants see all of the previously seen objects again, interspersed with all of the remaining objects from the original pool of objects. (25% Category A, 25% Category B, 50% Remaining)

While it looks pretty straightforward to choose which objects are in Category A/B/remaining up front and just randomize their presentation, what I would like to do is randomly assign the objects from the pool of objects to the categories and then randomize the order of presentation. What is a good way to go about doing this?

Hypothetical code:

<item pool_of_objects>
/1 = truck
/2 = banana
/3 = sweater
/ 4= umbrella
</item>

*magical category assignment randomization such that I end up with ...*

<item pool_CategoryA>
/1 = banana
</item>

<item pool_CategoryB>
/1 = sweater
</item>

<item pool_Remaining>
/1 = truck
/2 = umbrella
</item>


<trial CategoryA>
/stimulustimes = [0 = CategoryA]
</trial>

<trial CategoryB>
/stimulustimes = [0 = CategoryB]
</trial>

<trial Category_Remain>
/stimulustimes = [0 = Remaining]
</trial>