A concurrent working memory load task does not impair visual selective attention: A meta-analysis

Many studies have investigated how selective attention can be modulated by information maintained in Working Memory (WM) within a dual-task paradigm. In those studies, an attentional task must be performed while a concurrent working memory load task must be done at the same time. While some studies have found that memory loads can impair attentional performance, others have not found such effects, and even a few of them have shown an improvement of attentional selection when memory is loaded. De Fockert (2013) made a first attempt to integrate the results of most of those studies. The main conclusion he made from the results of his study was that loading WM does in fact impair attention. The objective of the present study was to re-analyse most of those studies (and a few others meeting inclusion criteria) improving his work by using the meta-analytical methodology on the interaction between WM and attention. Interestingly, the main finding of the present study contradicts De Fockert's: We have found a non-significant combined effect size, that is, there is no interaction between WM and attentional processes in those dual-task studies. However, there is in fact an important heterogeneity among effect sizes (I2 > 90%) worth to explore for an explanation. We discuss how several variables may have an impact modulating those effects. Finally, we have also detected an important publication bias: there is a tendency to publish significant results, leaving unpublished those studies finding no significant results. This publication bias strengthens our results, which are also in accordance with theories explaining at least part of the heterogeneous results among those studies (Olivers et al., 2011). Meeting abstract presented at VSS 2017