Working Memory Training Does Not Improve Performance on Measures of Intelligence or Other Measures of Far Transfer

Contrary to previous findings, a recent study has shown that there is no good evidence that working memory training improves intelligence test scores or other measures of ”real-world” cognitive skills.

A recent publication by Melby-Lervag et al., “Working Memory Training Does Not Improve Performance on Measures of Intelligence or Other Measures of ‘Far Transfer’,” provides a meta-analysis of studies that had claimed working memory training effectively benefits the participants. This meta-analysis review focused on common issues and misleading claims about far transfer measurements found in working memory training studies.

87 publications were identified as claiming to have supportive finds for working memory training, with a combined 145 different experimental comparisons. At least 50% of all of these studies’ foci were on working memory training. Melby-Lervag et al. organized the various measurements of cognitive abilities measured in these studies into three categories-- near transfer measures (similar to the tasks performed during the study), intermediate transfer measures (related working memory tasks), and far transfer measures (tasks clearly separate from the in-study tasks). Far transfer measures in this study refer to the necessary, long-term skills needed for academic success; verbal ability, nonverbal ability, reading comprehension, word decoding, and arithmetic. These studies all had different (or missing) untreated control groups, immediate and follow-up training/check-in, or real world post-test review.

This meta-analysis review of these publications used the Comprehensive Meta-analysis Program, which divides the difference between pretest and posttest results. Studies which showed a positive pretest-posttest gain were considered to have effective results that support working memory training.

The results of this meta-analysis show that immediate effects of far transfer measurements on nonverbal ability, verbal ability, word decoding, and arithmetic were close to zero and not significant, and there was no follow up effect of verbal ability, nonverbal ability, word decoding, or reading comprehension. However, there was a small positive effect on reading comprehension that was only observed immediately after training, and on arithmetic that was observed as a follow-up with no immediate-posttest review, but these two improvements lacked the proper, additional immediate/follow up test that would ensure correct retention.

Studies which conducted intermediate and near transfer measures, in contrast, did show significant immediate and follow up effects, which Melby-Lervag et al. suggested means that far transfer measures do not effectively produce immediate or long-term improved performance. These near and intermediate transfers have strong evidence for improved verbal and visuospatial working memory abilities, while recorded far transfers of these improvements were close to zero.

In their conclusion, Melby-Lervag et al. state that far transfer measurements during working memory training do not correspond to improved long-term effects. The studies used in this meta-analysis all had inconsistent methodological factors including age, training dose, design type, learner status, and perhaps most importantly the lack of untreated control groups in some of the studies, which Melby-Lervag et al. claim contributed to the perceived successful results of these studies.

This meta-analysis review was published in the Perspectives on Psychological Science on July 29, 2016.