Is bilingualism associated with better working memory capacity? A meta-analysis

ABSTRACT

Due to their experience of managing two languages, it has been suggested that bilinguals could receive more practice in the domain of working memory (WM) leading to a WM bilingual advantage. Although some studies have shown that bilinguals can outperform monolinguals in WM tasks, the studies investigating WM capacity in bilinguals provide inconsistent findings regarding the existence of a bilingual advantage. We therefore conducted a meta-analysis on the association between bilingualism and WM capacity. Data from 116 studies (involving 177 pairs of participants and 444 effect sizes) were extracted. Based on previous findings, we examined age, characteristics of the WM tasks –i.e. complexity (simple span vs transformation vs complex span tasks) and domain (verbal vs nonverbal) – age of first exposure to L2, and L2 proficiency as potential moderating variables. Results indicated a small bilingual advantage in WM (g = .12, p = .054), which was moderated by the language used in the verbal WM task. The bilingual advantage was stronger when the verbal WM task was performed in L2 compared to L1. Thus, the bilingual experience is associated with slightly higher WM capacity. Nevertheless, future studies would benefit from greater consideration of individual differences within groups of bilingual individuals.

KEYWORDS:

• Bilingualism
• working memory
• meta-analysis

Acknowledgements

The authors would like to thank Dr Guillaume Chevance for his support regarding the data analysis strategy and Raluca Pierrot for its guidance in the bibliographic research. We are very grateful to all the authors who kindly responded to our queries and who provided additional data to the meta-analysis. Thanks to their help, this study was as complete as possible.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 For the interpretation of effect sizes, we refer to the guidelines suggested by Cohen (1988): >.0 = very small difference; .2 = small difference, .5 = medium difference, .8 = large difference.

2 In the literature, the term age of first bilingual exposure is often interchangeably expressed through the expression ‘age of acquisition’, which has been commonly employed to denote the age at which a monolingual individual first starts learning a new language.

3 SES was operationalized as a compound variable that included family income, parental education, occupational status and place of residence.

4 Among the 177 effect sizes available for the ‘young’ category, only 15 came from samples of adolescents (>12 y.o.) while all the others (162) were observed in children <12 y.o. This weak proportion of adolescents led us to regroup all those results into a single category.

5 A funnel plot is a scatterplot of study-specific effect sizes (on the x axis) against the measures of study precision (here, standard errors) on the y axis. Funnel plots are used to explore the presence of small-study effects often associated with publication bias. In the absence of publication bias and heterogeneity, the scatter resembles a symmetrical funnel and a large proportion of effects sizes appear in the triangle area formed by connecting the centered summary estimate with its 2.5% and 97.5% quantiles on either side, indicating that they estimate the same effect. In our case, the presence of dots on the bottom, right of the plot, suggests that some small-power studies are included which generated the strongest effect sizes.

Additional information

Notes on contributors

Catherine Monnier

Catherine Monnier is associate professor of developmental psychology at the Paul-Valéry Montpellier 3 University, Montpellier, France. Her research focuses on children's working memory and particularly on the strategies children use during verbal or visuo-spatial short-term memory task performance. More recently, her interest has also focused on the development of prosocial lying in children.

Julie Boiché

Julie Boiché is associate professor at the Sport and Exercise Sciences Department of University of Montpellier, France. Her research focuses on implicit and explicit psychosocial processes implied in physical activity and sedentary behavior lifestyles. During the last decades, she developed an expertise in the field of systematic review and meta-analysis.

Pauline Armandon

Pauline Armandon completed a Master degree in developmental psychology at the Paul-Valéry Montpellier 3 University, Montpellier, France in 2019.

Sophie Baudoin

Sophie Baudoin completed a Master degree in developmental psychology at the Paul-Valéry Montpellier 3 University, Montpellier, France in 2019.

Stéphanie Bellocchi

Stéphanie Bellocchi (Ph.D. 2008, Bologna University, Italy) is an Associate Professor of Developmental Psychology at the Department of Psychology and at the Epsylon Research lab of the Paul-Valéry Montpellier 3 University, Montpellier, France. Her research focuses on reading development in monolingual and bilingual children. Particularly, she is interested on the differences and similarities between monolinguals and second-language learners while learning to read. She also studies the impact of visual-perception skills on typical and atypical reading development. Mostly, she explores the nature of the link between visuo-attentionnal deficits and developmental dyslexia with or without developmental coordination disorder.