Working memory training: An updated meta-analysis of transfer to fluid intelligence measures
S.S. Wootan, P.J. Reber & D.K. Mroczek
Annual Meeting of the Cognitive Neuroscience Society
Working memory (WM) training research has sparked both excitement and skepticism, particularly around the claim that these interventions can result in improved fluid intelligence. Robust far-transfer effects following training would have significant implications for theories of working memory, neuroplasticity, and fluid intelligence in addition to applications of these theories in education, training, and cognitive rehabilitation. Among the challenges to addressing the reliability of effects is the fact that many interventions studies are underpowered to detect potentially subtle far-transfer effects. Melby-Lervåg and Hulme (2012; MLH) reported a meta-analysis of WM training effects that only included studies with substantial training and a comparison control group. Their analysis suggested that these WM training paradigms have not produced consistent fluid intelligence gains across studies. Here we report an updated meta-analysis in which additional recent studies are included. Critically, the addition of new findings allows for separate examination of participant populations that were combined in the MLH report: young healthy adults, older adults and children (some with learning or other disabilities). In 13 studies with young healthy adults, the meta-analytic effect size of working memory training interventions was d=0.238, p<.001, indicating consistent gains in measures of fluid intelligence and cross-study reliable far transfer. Thus it appears that for adults, working memory training does produce consistent increases in performance on tests of fluid intelligence and research efforts should be directed to better understanding of the interventions and conditions that lead to the most effective gains.