The hypothesis that managing two languages enhances general executive functioning is examined. More than 80% of the tests for bilingual advantages conducted after 2011 yield null results and those resulting in significant bilingual advantages tend to have small sample sizes. Some published studies reporting significant bilingual advantages arguably produce no group differences if more appropriate tests of the critical interaction or more appropriate baselines are used. Some positive findings are likely to have been caused by failures to match on demographic factors and others have yielded significant differences only with a questionable use of the analysis-of-covariance to "control" for these factors. Although direct replications are under-utilized, when they are, the results of seminal studies cannot be reproduced. Furthermore, most studies testing for bilingual advantages use measures and tasks that do not have demonstrated convergent validity and any significant differences in performance may reflect task-specific mechanism and not domain-free executive functions (EF) abilities. Brain imaging studies have made only a modest contribution to evaluating the bilingual-advantage hypothesis, principally because the neural differences do not align with the behavioral differences and also because the neural measures are often ambiguous with respect to whether greater magnitudes should cause increases or decreases in performance. The cumulative effect of confirmation biases and common research practices has either created a belief in a phenomenon that does not exist or has inflated the frequency and effect size of a genuine phenomenon that is likely to emerge only infrequently and in restricted and undetermined circumstances.
Copyright © 2015 Elsevier Ltd. All rights reserved.