A growing trend in memory rehabilitation is to combine learning principles to enhance treatment effects. While this makes intuitive sense, little is known about the added value of incorporating each method. A further complication is that some interventions, although primarily based on one learning principle, actually incorporate several, which again adds to the difficulty in differentiating the individual contribution of techniques. In this paper we report results of two experiments comparing the effectiveness of combining principles of errorless learning (EL) with vanishing cues (VC) relative to each in isolation. Healthy controls (N = 60), learning under standard and dual-task conditions, and patients with probable Alzheimer's disease (N = 22) took part in Experiments 1 and 2, respectively. In each experiment, participants were asked to learn greeble- or face-name associations, and memory was tested after interference or brief delay. For healthy controls, EL alone and EL + VC produced the best results, but there was no difference between these conditions. For the Alzheimer's patients, all treatment conditions (EL, VC, EL + VC) were significantly better than trial and error, and, in this case, we found that the combined method was significantly better than VC in isolation. Importantly, in both experiments there was little support for use of combined over individual learning principles.