Discussion 1. Are self-ratings an adequate measure of L2 proficiency? Figure 1 • Significant positive correlation between self-ratings and cloze test scores. • Self-ratings are also an adequate measure of L2 proficiency. • Self-ratings include data on a number of factors, such as reading, writing, speech production and comprehension. • Self-ratings are a more detailed measure than cloze tests • Cloze tests only evaluate reading and grammar skills. 2. Does the LBQ yield reliable and systematic results? Figure 2 • Participants grouped consistently in the same way. • Four different stable grouping classifications • LBQ does yield reliable and systematic results. 3. Can the LBQ successfully assign participants to different groups based on Ao1E, AoI, MoA and proficiency? Figures 3-6 • Groupings are not the same across variables. • LBQ can tease apart complex, often confounded variables Figures 7-9 • Expected pattern of correlations and predictors. • AoI and Ao1E are similarly correlated • AoI and MoA serve as a predictor of proficiency • MoA and AoI/Ao1E have different groupings and R values Myriam Lapierre, Laura Sabourin, Michèle Burkholder, Christie Brien and Jean-Christophe Leclerc University of Ottawa References [1] Sabourin, L., Brien, C., & Burkholder, M. (2014a). The effect of age of L2 acquisition on the organization of the bilingual lexicon: Evidence from masked priming. Bilingualism: Language and Cognition, 19, 1–14. [2] Sabourin, L., Leclerc, J.-C., Burkholder, M. & Brien, C. (2014b). Bilingual Lexical Organization: Is There a Sensitive Period? Paper presented at the Conference of the Canadian Linguistics’ Association, St. Catharines, Canada. [3] Marian, V., Blumenfeld, H., Kaushanskaya, M. (2007). The Language Experience and Proficiency Questionnaire (LEAP-Q): Assessing Language Profiles in Bilinguals and Multilinguals. Journal of Speech, Language, and Hearing Research, 50, 940-967. [4] Brown, J. D. (1996). Testing in language programs. Upper Saddle River, NJ: Prentice Hall. [5] Tremblay, A. (2011). Proficiency assessment standards in second language acquisition research: “Clozing” the gap. Studies in Second Language Acquisition, 33, 339-372. Our LBQ • Aims to gather information regarding schooling method, language proficiency self-ratings, MoA, Ao1E and AoI. • Two parts, which exist in both English and French: Part 1. The Short LBQ • Provides a general overview of the participant’ s language background, which includes the following information: • Biographical data • Key language information • Self-ratings of overall proficiency • Serves to identify functional monolinguals, exclusions and the possible need for additional information. Part 2. The Long LBQ • Only participants for which additional information is required must fill out the Long LBQ. • E.g.: bilinguals and multilinguals • Provides complementary information on numerous factors influencing L2 acquisition in order properly operationalize them: • Ao1E and AoI: language exposure during early childhood (e.g., language(s) of caregivers) or later L2 exposure • Proficiency: detailed self-ratings • MoA: evolution of language use from birth to present Introduction • Language background questionnaires (LBQs) are a fundamental tool used in studies targeting bilingual and multilingual populations. • Their success depends on their ability to assess complex, often-confounded variables such as age of first exposure (Ao1E), age of immersion (AoI) manner of acquisition (MoA) and language proficiency. • Other LBQs [3] are available. • They lack the ability to fully and systematically tease apart some of these variables • Less practical for testing bilinguals in Canada, where Ao1E, AoI and MoA play an important role in L2 acquisition [1,2] Goal • Develop a LBQ that assesses multiple factors affecting bilingual processing, e.g., Ao1E, AoI, MoA and proficiency. Research questions 1. Are self-ratings an adequate measure of L2 proficiency? 2. Does the LBQ yield reliable and systematic results? 3. Can the LBQ successfully assign participants to different groups based on Ao1E, AoI, MoA and proficiency? Results Question 1 Figure 1 • Significant correlation between self-ratings and cloze test scores (r=0.667, p<.001). • If participants had an L1 that was neither French nor English, both English and French scores were included. Question 2 • Short LBQ reliably identifies monolinguals. • Confirmed by looking at long LBQ’s data Figure 2 • Significant correlation between test and retest (r=0.882, p<.001) conditions. • Stable groups across two rounds • Each participant was assigned to a group according to four different variables (MoA, AoI, Ao1E and proficiency) under two conditions (test and retest). • Each participant yielded a pair of values based on their level of the corresponding grouping variable Results (continued) Question 3 Figures 3-6 • Grouping of participants, when AoI is used as a baseline. • Large amount of variability in the grouping of participants, depending of the chosen variable • Multiple regression analysis using self-rated proficiency as a DV and MoA, AoI and Ao1E as IV. • Only MoA (R 2 =.390, p<.001) and AoI (R 2 =.216, p<.001) are significant predictors of proficiency • Ao1E was not a significant predictor (p=.153) Figures 7-9 • MoA (r=0.749, p<.001), AoI (r=-.529, p<.001) and Ao1E (r=-0.372, p=.001) all significantly correlate with self-rated overall proficiency • MoA correlates most with self-rated overall proficiency • AoI (r=-.536,p<.001) and Ao1E (r=-.395,p <.001) AoI (r=-.536,p<.001) with correlate most with self- Ao1E (r=-.395,p with with self- rated pronunciation. The Language Background Questionnaire in the Canadian Context Methodology Participants • Round 1: 81 participants (n female=70). • L1 French=29, L1 English=42, L1 other=10 • Round 2: 48 of the same participants tested in round 1 (n female=43). • L1 French=19, L1 English=21, L1 other=8 Materials • Short and Long LBQs. • English and French cloze tests to assess proficiency [4,5]. Procedure • Round 1: Both versions of the LBQ and both cloze tests. • Round 2: Both versions of the LBQ. Figure 1: Correlation between L2 self-ratings and cloze test scores Figure 2: Correlation between test and retest results of participant groupings Figure 7-9: Correlations between self-rated L2 proficiency and MoA, AoI and Ao1E respectively n=75, r=0.749, p<.001 n=78, r=-0.372, p=.001 n=65, r=-.529, p<.001 points are jittered, n=185, r=0.882, p<.001 n=85, r=0.667, p<.001 Future directions • Adapting the LBQ to serve different populations. • Translating into other languages • Investigating MoA as the best predictor of proficiency, a newer variable in L2 acquisition research. Figure 3-6: Grouping of participants according to AoI, Ao1E, MoA and proficiency, where AoI is used as a baseline