The questions used in the survey were developed by the author and piloted in four adult literacy CASPs in New Brunswick. The questions were drawn from other studies in the field of adult literacy (Gabor-Katz, 1991; Horsman, 1990; MacKeracher, 1993) and from my personal experience of teaching reading and language arts courses to Bachelor of Education students for over ten years. Each year I have designed a literacy profile survey and administered it to these students. I wanted to explore their literacy history and their current literacy activities. These informal survey results led to more development and revisions, until the survey became useful for planning and teaching, and finally became a tool for research.

The survey questions were modified following the pilot study. In consultation with the CASP learners and their facilitators, I developed the questions most appropriate for meeting the research objectives of this study. The resulting questionnaires were sent to the facilitator at 16 CASPs in the central region of New Brunswick. Permission to do this was obtained from the Provincial and Regional Literacy Coordinators who were responsible for the CASPs in the region.

Precautions were taken to prevent the possibility of identifying any individual respondents. Because CASP learners had an established relationship with their respective facilitators, the questionnaires were administered by these individuals. This approach was deemed to be potentially less threatening to the learners than if I, a complete stranger, had come into the setting. Completed surveys were received from 175 anglophone CASP learners, 50 men and 125 women. Not all respondents answered all questions.

The data were coded numerically and then prepared for analysis using SPSSX. Frequency distributions were run for all questions and crosstabs calculated for all questions by age and by gender. Chi squares were then calculated for the more interesting crosstab results. Only those results which seemed interesting or which were of significance at a confidence level of 0.01 or better are reported here.