Retrospective verbal-report data were collected from twenty-four University of Minnesota Spanish, French, and German immersion program students, directly after they performed a classroom task (e.g., engaging in process writing, listening to a lecture, watching a video, or discussing an article). A third of the students also performed listening, reading, writing, and speaking activities around a central topic outside of class, providing verbal report data. For listening, reading, and writing tasks, verbal report took place during the activity itself. The listening portion involved stopping the tape to report strategies for dealing with challenging sections of a pre-recorded oral recitation. During the speaking task, students were interrupted twice - usually at hesitation points - and asked to provide a verbal report concerning their language production strategies. Directly after completing the task, their performance was replayed to them, and they provided retrospective data on the processing and production strategies they had employed. Similarities and differences in bilingual processing strategies were analyzed: 1) according to individual student across modality, task, and program, and 2) comparatively among students within the same reporting context (e.g. class activity or out-of-class tasks). Furthermore, over a dozen instructors in non-immersion classrooms completed a questionnaire dealing with obstacles to target-language processing by their students and strategies employed to overcome these obstacles.

The immersion program students reported using less mental translation and more cognitive processing directly through the immersion language than did non-immersion program students who were also studying in the immersion classes. Further, students reported on the contribution of the unique social context of immersion on their language learning. These findings would suggest that the immersion context - due to its special linguistic and social dimensions - promoted cognitive processing directly through the immersion language, which was not so much the case for non-immersion students in those same classrooms. Furthermore, the questionnaire data from instructors in non-immersion classrooms would suggest that in that context, the challenge of stimulating continuity in target-language cognitive processing is even greater.