Dissertations, Theses, and Capstone Projects

Date of Degree

2-2016

Document Type

Dissertation

Degree Name

Ph.D.

Program

Educational Psychology

Advisor

Bruce Homer

Advisor

Patricia Brooks

Committee Members

Patricia Brooks

Jay Verkuilen

Subject Categories

Cognitive Psychology | First and Second Language Acquisition | Psycholinguistics and Neurolinguistics

Keywords

bilingualism, executive function, meta analysis

Abstract

A much-debated topic in psycholinguistics is whether lifelong bilingualism enhances executive functions (EF), the set of higher-order cognitive processes involved in the control of thought and action. Several researchers have predicted bilingual advantages on various EF tasks, especially interference-control and task-switching tasks. Many studies have tested these predictions, but results have proven unreliable. As a complementary approach to recent quantitative syntheses on this topic, the present dissertation tests whether the bilingual advantage is moderated by a number of theoretically significant variables: dependent variable (DV), task, age, age of L2 acquisition and lab.

Two meta-analyses were conducted. Study 1 considered interference-control tasks. It synthesized 168 effect sizes from 43 studies. There was a significant interaction between age and dependent variable: the bilingual advantage was larger for children than young adults on global reaction times (global RTs), and larger for older adults than younger adults on both dependent variables. There was also a significant interaction between age of acquisition and dependent variable: samples with bilinguals that learned their second language early exhibited larger effect sizes on global RTs than samples with bilinguals that learned their second language later. However, both of these interactions could be explained in terms of differential publication bias. There was also a strong lab effect. Study 2 considered task-switching tasks. It synthesized 30 effect sizes from 10 studies. However, it yielded a non-significant overall effect size that was not moderated by DV or lab. Overall, the two studies yielded relatively inconclusive evidence for the bilingual advantage. While Study 1 revealed some coherent patterns of moderation, all of these effects could be due to publication bias. Furthermore, Study 2 revealed no evidence for an advantage on task-switching tasks. Various limitations of the present analysis and the literature more broadly may have obscured bilingual advantages if they do exist.

Download

Share

COinS