It all depends, in my view. Some clayey soils have already been sheared in the past - or through creep have undergone large deformations. In these cases the residual strength should be considered. (i.e., there is a case history in that a slope at 11 deg failed when the toe was undercut by one metre for widening a road). You may use peak strength when such previous deformations have not occurred (as in clay fill). The factor of safety would ensure that you don't go over the peak. However, seldom, again in my view, would you use clayey backfill or but a retaining wall up against a cut clay. There is also another case history of which I am aware for a slope holding up a railroad took 80 years to creep - and residual being reached.
Many papers on this - I think that Hutchinson (1969 Mexico State of Art Papers for ISSMFE) was a good one. Can also check out (you'll have google):
Residual Strength of Colluvium and Stability Analysis of Farmland Slope
B.Georges Kakou , Hideyoshi Shimizu ????, Shinichi Nishimura
E-mail: F6102005@guedu.cc.gifu-u.ac.jp
Rigo (2006) in Canadian geo Journal "Residual strength of Tropical soils"
Stark (1994) in ASCE GT Journal "Drained Residual Strength of Cohesive Soils".
Sorry I can't give out the URL's on the later.