Effects of script similarity on bilingual advantages in executive control are likely to be negligible or null

Coderre and van Heuven (2014) concluded that the orthographic overlap between a bilingual’s two languages is an important variable to consider in investigations of bilingual executive control (EC) abilities. Since greater overlap creates more cross-linguistic activation and (presumably) increases the daily demands of cognitive control Coderre and van Heuven hypothesize that similar-script bilinguals will have more effective domain-general EC. In their study German-English (GE) bilinguals were identified as a having high similarity, Polish-English (PE) as having moderate similarity, and Arabic-English (AE) as having low similarity; there was also a group of monolingual English speakers. All groups participated in both a Stroop and Simon task. The bilinguals did one Stroop session with printed words from L1 and a second session with L2. The interference effect (incongruent RT− neutral RT) was taken as a measure of general inhibitory control whereas global RT (mean RT across all trial types) was taken as a measure of “monitoring for conflict.” The hypothesis that orthographic overlap enhances EC efficiency (marked by smaller interference effects or faster global RTs) predicts this ordering: GE < PE < AE. Each measure (global RT, interference) affords three pairwise comparisons (viz., GE<PE, GE<AE, PE<AE) for each of the three tasks (Simon, L1 Stroop, and L2 Stroop) for a total of 18 pairwise tests. Using global RT as a measure of EC, the predictions are supported by only three of the nine comparisons. Even worse, the predictions for the interference effects were not supported by any of the nine comparisons: eight of the comparisons were null and the significant advantage of AEs over PEs in the L1 Stroop task contradicts the prediction. In total 3 of the 18 comparisons support the hypothesis that script similarity enhances EC. The assumption that the three significant differences in global RT that do occur between the bilingual groups should be interpreted as differences in EC is further eroded because none of the bilingual groups showed significantly faster global RTs compared to the monolinguals. With three different tasks and three different groups of bilinguals there were nine such opportunities. In summary, the entire case that orthographic overlap enhances EC is built on the three comparisons out of nine where lower similarity bilinguals were significantly slower in global RT compared to bilinguals with greater script similarity. But global RT is a very impure measure of EC (Paap and Greenberg, 2013) and, as discussed by Coderre and van Heuven, there could be many reasons why the AE bilinguals were slower than the other groups1. This spotty evidence