Abstract
This study explores early francization models for a linguistic minority currently struggling to preserve its language. The French Acadians of New Brunswick, Canada, represent 30% of the province's total population, yet their numbers and their linguistic vitality are decreasing. New Brunswick has two public school systems: the English language schools cater for the majority while the French language schools have been created for the minority, with the goal of this separate, homogeneous school system being to preserve the French language. However, for some children who are members of this linguistic minority, the language is no longer spoken at home, and thus, when parents in such situations opt for the French school, their children are immersed. Data were collected from nine schools in order to examine the particular Early Francization Model they operate, and how they manage their mixed intake. Four different models of early francization (sheltered class, mainstream class with pull-out, mainstream class with pull-out plus in-class support and integrated model) for this revitalization effort are described and group interviews were carried out to assess how the schools organise their francization programme. In addition, French language outcomes were measured for students enrolled in the programme in the kindergarten year (five years old). The quantitative data indicate that children who participated in the integrated model fared better in terms of French vocabulary outcomes.
Acknowledgements
We would like to acknowledge the financial contribution of the Réseau d'appui à l'intégration des enfants des ayants droit au système éducatif francophone for this study. We would also like to thank Natasha Prévost who participated in the data collection. We are grateful and wish to thank the reviewers and editors for their insightful comments. A special thank you and special thoughts are expressed for Anne Lowe who was a main collaborator in this study. Anne Lowe passed away in February 2012 after a courageous battle against brain cancer.
Notes
1. Simple comparisons with Sidak's correction show that Model 4 is significantly superior to Model 1 (Δm = 16.433, ε = 3.839, p = 0.000), 2 (Δm = 14.495, ε = 3.624, p = 0.001) and 3 (Δm = 21.988, ε = 4.449, p = 0.000). The differences (Δm) between the other models vary from 1.938 and 7.494 and none are significant.