Absolutely! I based my dissertation on analyzing how students selected different lower level CS exercises (typing exercises, Parson Puzzles, output prediction, etc.) based on ICAP. What I observed was lower Active exercises benefit all students, the general order of exercise selection follows a sawtooth wave (work upwards to assessment, then reset for the next week), completers and non-completers in a MOOC selected 'next exercises' similarly, and no one likes pop ups recommending you to 'downgrade' if you're struggling on a problem. There are limitations to my work, like the order of how exercises were presented primed their selections, but the overarching theme I tried to convey was that lower-level drilling is an absolute necessity for learning.