Individual Submission Summary
Share...

Direct link:

Poster #116 - Moderators of the Relationship Between Age of Acquisition and Second Language Proficiency: A Meta-Analysis

Thu, March 21, 2:15 to 3:30pm, Baltimore Convention Center, Floor: Level 1, Exhibit Hall B

Integrative Statement

The “younger is better phenomenon” of second language learning suggests that adults will have a more difficult time learning a second language (L2) when compared with children learning that same language (DeKeyser & Larson-Hall, 2005). This notion is pervasive not only in research, but also in popular culture and education, as both laypeople and academics attempt to explain adults’ L2 shortcomings in terms of an age effect. A large body of literature describes the negative relationship between age of acquisition (AoA) and L2 proficiency using evidence from linguistics, psychology, neuroscience, and biology. Relying on this interdisciplinary approach, Oyama (1978) concluded that L2 acquisition concerns “interactions between the developing phenotype and the environment.” Research suggests that the nature of these interactions is likely to change during the course of development, and therefore limit the potential for native-like L2 attainment in adulthood.

Researchers have formally studied the possibility of a critical or sensitive period for language acquisition since the mid-20th century (Penfield & Roberts, 1959). This body of research demonstrates differences in grammar, accent, fluency, vocabulary etc. between those who began learning an L2 as children, and those who acquired an L2 in adulthood (DeKeyser & Larson-Hall, 2005). The purpose of the current, ongoing study employs meta-analytic methodology in order to estimate the relationship between age of acquisition (AoA) and L2 acquisition, and to determine the empirical basis of arguments related to the critical period hypothesis. Specifically, the goal of the current, ongoing study is to identify the extent to which AoA matters in attaining native-like L2 proficiency, and the moderators that influence the relationship.

The current meta-analysis includes published and unpublished works utilizing all study designs that can determine a relationship between AoA and L2 proficiency, as well as all populations learning an L2 in either an immersive or formal context. While coding and analysis is ongoing, the following moderators will be included in final analyses: gender ratio of the sample, type of L1, type of L2, type of outcome, age at testing, age of first exposure to L2, age of immersion in L2, and length of exposure to L2. Currently, the authors have identified and coded information from approximately 50 studies (which include approximately 120 effect sizes) for inclusion in these analyses. The results indicate an overall negative correlation between AoA and L2 proficiency [rho = -.40; 95% CI (-.58, -.20)], which lends support for the existence of a sensitive period for learning an L2. Additionally, the negative correlation was stronger when both L1 and L2 were European/English [rho = -.50; 95% CI (-.70, -.31)], compared to that when L1 was Asian and L2 was European/English [rho = -.38; 95% CI (-.58, -.13)]. The negative correlation was especially strong for L2 grammar [rho = -.52; 95% CI (-.68, -.23)], compared to that of L2 pronunciation [rho = -.43; 95% CI (-.62, -.19)] and that of global L2 measurements [rho = -.31; 95% CI (-.47, -.17)]. Results indicate that a variety of factors may moderate relationships between AoA and L2 proficiency.

Authors