Abstract:
This research study investigated the extent to which the language of discussion influenced Samoan students' performance on academic tasks. The hypothesis, based on Saville-Troike's (1984) work, was that bilingual students (Samoan and limited English proficiency) who completed an academic learning task in Samoan would do better on an English medium test than a matched group of bilingual students who did the task using English.
Over a period of four weeks, sixteen Form four and fourteen Form six students who were native speakers of Samoan completed tasks on science topics at an English medium high school for girls. Within each level, students were matched and randomly assigned to either an experimental or comparison group. The experimental group completed their tasks using Samoan as the language of group discussion while the comparison group used English. The students' performance was then measured in English in terms of vocabulary learning, knowledge development, and the content and organisation of their writing.
The results of the study corroborate evidence published elsewhere that students with limited proficiency in English appear to benefit from the opportunity to discuss school work in their primary language, even though the instructions, resources, tasks and criterion tests are in English. The students in the experimental group not only showed a better grasp of key vocabulary and a higher level of propositional knowledge, but they also produced more elaborated written answers to content questions.
The results of the study are discussed in terms of the nature of the discourse produced under two treatment conditions by representative students. When Samoan was used as the base language, English technical words and other units of meaning were frequently embedded in the discourse.
