Standard EF tasks can still have predictive validity within diverse cultural contexts

Kroupin et al.(1) administered standard executive function (EF) tasks to children in the United Kingdom, Bolivian children attending "limited" schooling, and Kunene children in Namibia attending or not attending school.Children in the United Kingdom performed best on these assessments, whereas unschooled Kunene children performed worst, often never providing expected responses.The authors conclude that standard EF tasks, created and normed in industrialized and schooled countries, capture cognitive skills that develop in those contexts (see also refs.2-5 for discussion).I agree with Kroupin et al. that unmeasured cultural factors likely drive the observed mean differences in EF task performance; different cultural affordances lead to specialized forms of EF engagement, which in turn influence scores on standard EF assessments across diverse samples (6-8).However, additional analysis of Kroupin et al. data indicates that standard EF tasks capture theoretically and practically meaningful variation in children's EF, even in these less industrialized samples.Kroupin et al.(1) dataset includes measures of reading and math for the Bolivian and Kunene children attending school, in addition to age and years of schooling.Additional analysis of these data reveals two key patterns beyond those described in the article.First, EF task performance increased with age in the Bolivian and both Kunene samples, across EF assessments; age remained a significant predictor of performance for many EF assessments in the Bolivian and schooled Kunene samples even when controlling for years of schooling (Table 1).Second, EF task performance correlated with math and reading scores in each of the schooled samples; many of these correlations persisted even when controlling for years of schooling (and 17/18 coefficients remained positive; Table 2).If EF task performance primarily reflects school exposure, then these correlations should disappear when controlling for years in school.Instead, agerelated increases were observed in all samples, and