ITEM METADATA RECORD
Title: Levels of aggregation: Identification of sociolinguistic determinants of L1 proficiency in higher education
Authors: Tummers, José
Deveneyns, Annelies
Issue Date: 3-Jun-2015
Host Document: Languages, Literatures & Literacy. 10th IAIMTE Conference 2015 pages:257-258
Conference: International Association for the Improvement of Mother Tongue Education edition:10 location:Odense date:3-5 June 2015
Abstract: We will investigate sociolinguistic determinants of written L1 proficiency of students in higher education in Flanders. Firstly, higher education institutions set up various support initiatives to improve the academic writing skills of incoming students (Peters & Van Houtven 2010). Secondly, research identified a correlation between first year students’ language proficiency and their study progress (De Wacher et al. 2013).
To support evidence-based language policy, the following research question will be addressed: to what extent is the academic native writing proficiency constrained by sociolinguistic determinants?
At University College Leuven, 348 first year bachelor students wrote an academic argumentative text of 500 words in Dutch (L1) at the end of the first year. The participants were invited to convince government officials of their opinion on social media. They had one hour to complete that task on a computer and were allowed to use all sources deemed useful. The participants were sampled from 13 bachelor programs, ranging from laboratory analysis and business management over nursery to teacher training and social work.
The texts were analytically rated on four symmetrical 4-point scales for language, structure, argumentation and persuasion (Hawkey & Baker 2004). The formal requirements of each scale level were explicitly described (Knoch 2011). An analytical evaluation criterion is considered sufficiently realized, when it corresponds to level B2 of the Common European Framework of Languages, which is generally considered the minimal proficiency level to enter higher education (CNaVT). The impact of the following sociolinguistic variables was tested: the student’s socio-economical profile, his/her study choice in higher education and his/her secondary education history.
First a Multiple Correspondence Analysis (Greenacre 1984) was carried out reducing the four analytical criteria to two latent dimensions identifying a cline from (very) well written to (very) poorly written texts. Next, a hierarchical linear model was fitted (Gelman & Hill 2007), with student properties modeled at level-1 and school/program properties modeled at level-2.
A major impact of the study program emerged, which can largely – but not entirely – be related to the embedment of L1 language modules in the curriculum. Furthermore, secondary education predetermines to a certain extent the level of writing proficiency in higher education.
Publication status: published
KU Leuven publication type: IMa
Appears in Collections:Department of Business and Management Studies - UC Leuven
Quantitative Lexicology and Variational Linguistics (QLVL), Leuven
Department of Teacher Training - UC Leuven

Files in This Item:
File Description Status SizeFormat
IAIMT2015_TummersDeveneyns.pdf Published 191KbAdobe PDFView/Open

 


All items in Lirias are protected by copyright, with all rights reserved.