Can intelligence be increased by training on a task of working memory?

1 I am grateful to D. Detterman for bringing this issue to my attention. The May 13, 2008 issue of Proceedings of the National Academy of Sciences featured a cover article that purported to demonstrate increases influid intelligence following training on a task ofworkingmemory (Jaeggi, Buschkuehl, Jonides, & Perrig, 2008). The authors described their own findings as a “landmark result”. Their studywas the subject of an introductory comment by Robert Sternberg (2008), as well as articles in the mainstreammedia, including a lengthy column in a recent edition of The New York Times (Wang & Aamodt, 2009). In view of the potential significance of the study and the quantity of attention it has received, the results have been subjected to remarkably little critical analysis. A close examination of the evidence reported by Jaeggi et al. shows that it is not in fact sufficient to support the authors' conclusion of any increase in their subjects' fluid intelligence. What Jaeggi et al. reported were modest increases in performance on a test of fluid intelligence following several days of training on a task of working memory. The reported increases in performance are not in question here. But the manner in which the test was administered severely undermines the authors' interpretation that their subjects' intelligence itself was increased. The subjects were divided into four groups, differing in the number of days of training they received on the task of working memory. The group that received the least training (8 days) was tested on Raven's Advanced Progressive Matrices (Raven, 1990), a widely used and wellestablished test of fluid intelligence. This group, however, demonstrated negligible improvement between preand post-test performance. The other three groups were not tested using Raven's Matrices, but rather on an alternative test of muchmore recent origin. TheBochumerMatrices Test (BOMAT) (Hossiep, Turck,& Hasella, 1999) is similar to Raven's in that it consists of visual analogies. Inboth tests, a series of geometric andotherfigures is presented in amatrix format and the subject is required to infer a pattern in order to predict the next figure in the series. The authors provide no reason for switching from Raven's to the BOMAT. TheBOMATdiffers fromRaven's in some important respects, but is similar in one crucial attribute: both tests are progressive in nature, which means that test items are sequentially arranged in order of increasing difficulty. A high score on the