The sort of bible reference, I believe, is the 1960 Boulder Conference on Clays (ASCE). Might try to find a copy of it and read the paper by, again, I believe, Bishop.
One caution on compacted clay fills - need to keep your lifts proper. I know one job that had major problems in tha the upper part of the lift was compacted and the lower part not nearly enough. After construction distress, they found this clearly evident by moisture content profile and undrained shear strength profile.
I would also surmise that the methods of compaction are very important as well. Use of sheepsfoot roller or PTR rather than vibratory. In the lab, things always are "ideal" in that the sample before compaction would be of similar gross gradation (i.e., clump gradation) but on the job site, clay borrow usually comes in much bigger clumps. If there is no care in breaking down the clumps before compaction, you might only pack the clumps together in a non-homogeneous mixture. The interconnected voids would have effects on the permeability of the fill mass as well on the strength of the mass.
Just some initial thoughts.