In most studies on plasticity using behavioural data, cognitive plasticity has been studied in the context of short-term interventions such as memory training. In order to systematically review the literature on memory training and summarize its effects for old healthy people and people with mild cognitive impairment on multiple functional domains, we conducted a meta-analysis of all published randomized controlled trials (RCT) between 1970 and 2007. Overall, 24 studies examining memory training effects in healthy and in mildly cognitively impaired old adults were identified and included in the analysis. Only memory trainings (e.g. memory skills training, imagery, method of loci) with cognitive outcome measures, duration of intervention with up to one year with at least a baseline and a post-intervention assessment reported, were included. Results demonstrate significant training effects for paired associate learning and immediate and delayed recall in healthy old adults and for immediate recall in mildly cognitively impaired old adults. However, training effects were no larger than those found for active control conditions. Our results suggest that evidence on the effectiveness and specificity of training interventions is scarce. We discuss limitations of existing knowledge about the efficacy of memory training interventions and implications for future research to improve knowledge regarding effective cognitive interventions.