Poster #136 - Evaluating Children’s Language Abilities in Côte d'Ivoire: Introducing the Ivorian Children’s Language Assessment Toolkit

• In Event: 2-002 - Special Poster Session 05 with Continental Breakfast Reception
  In Poster Session: PS 05 - International Section

Integrative Statement

INTRODUCTION: Introducing a child’s mother tongue as a language of instruction (LOI) in schools can improve academic outcomes (Adjibodou, 2011; Skattum et al., 2010). Hence, mother tongue LOI is part of educational policies in many developing and multilingual countries in sub-Saharan Africa. However, an evidence-base is required to understand how mother tongue instruction influences learning outcomes, and yet, there is often a lack of standardized scientific assessments adapted to local cultural and linguistic context. Here, we present the results of a new language assessment ​​for primary-school children: "The Ivorian Children’s Language Assessment Toolkit for ​​Abidji, Attié, Baoulé and Bété Languages". The toolkit is designed for educational and research activities and can be used to better assess children’s language learning and inform educational strategies.

METHODS: 830 children (age: 4-14y, M=9.36, SD=2.21; grades: 1st: n=206; 3rd: n=326; 5th: n=298) from 6 bilingual (mother tongue & French; n=338) and 8 monolingual (French) primary schools (n=492) completed assessments (adapted from EGRA) of phonological awareness (PA: identification, initial/final elision, segmentation), tonemic awareness (TA: tone matching), vocabulary (synonym, antonym generation) and oral comprehension in Abidji, Attié, Baoulé, and Bété languages.

RESULTS: Subtask scores and between-subtask correlations are shown in Fig1. PA was strongly correlated with vocabulary, whereas TA and oral comprehension were only modestly correlated with other tasks. We found overall significant effects of age (F(1,8)=41.96, p<.001, Wilk's Λ=0.635), grade (F(2,8)=15.71, p<.001, Wilk's Λ=0.678), language (F(3,24)=39.57, p<.001, Wilk's Λ=0.276), and school type (F(1,8)=6.45 p<.001, Wilk's Λ=0.919) on all subtasks. Variation in scores by language may reflect cross-linguistic differences (e.g., higher word-initial PA scores in Bete, which permits more complex-onset clusters), and/or regional SES differences (lowest for Baoule sample). Significant effects of school type (monolingual, bilingual) were observed specifically on initial phoneme identification, phoneme segmentation, and synonym generation scores: children at monolingual schools outperform their bilingual school peers. However, school type was also significantly associated with SES, with higher SES children attending more monolingual French-speaking schools (b=0.672, t(815)=2.162, p<.05).

DISCUSSION:
This study shows the first age and grade normative data describing rural Ivorian children’s performance on specific sub-tasks of a comprehensive language assessment. Presently, no references exist on the development of rural Ivorian school-children’s language abilities. The data from this toolkit provides descriptive information (mean, sd, range, distribution) for each grade, age, and gender. However, cross-linguistic differences in language structure and environmental factors (e.g., LOI, SES) also affect children’s scores on specific sub-tasks. Our study provides insights into the sources of variance in children’s language scores, and a guide that can be used by researchers and educators alike. For example, the ability to measure children’s PA, vocabulary and comprehension skills (robust predictors of literacy) in their mother tongue can yield useful information to support the child’s reading development.

Authors

• Hermann Akpe, Réseau Ouest et Centre Africain de Recherche en Education (ROCARE)
  Presenting Author

• Axel Seri, Centre de Recherche et d’action pour la Paix (CERAP)
  Non-Presenting Author

• Fabrice Tanoh, Université Félix Houphouët-Boigny
  Non-Presenting Author

• Kaja K Jasinska, University of Delaware
  Non-Presenting Author