Before we dismiss this topic as having 'little or no solid research' I would, once again, take us back to the work of Hermann Ebbinghaus who was 'a' (probably 'the'

pioneer in the area of memory research. The 'curve of forgetting' is key to this (as his possibly his book 'A contribution to experimental psychology' - but I confess that I've not read it). I belive he also 'discovered' nonsense sysllables and serial learning. The fact that his work was conducted in the early 1900s shouldn't reduce its validity.
From the small amount that I've read about his work and experiments, it appears well researched and designed. I would also mention that his sequence of reviewing includes the 3month recall.
I will admit to be a full-on pragmatist and more interested in whether it works in practice, which for me it does, than the theory. I would however, also agree with Resli that the reality is that people will fit in these reviews 'as and when they can'. I do though believe our role in L&D is to help people to absorb and work with the material that we offer in a way that makes it easy for them to want to, and make time for regular reviews.