A content analysis involves dividing the data into meaningful segments according to emerging themes and concepts (Miles & Huberman, 1994) while listening for "the meaning of what is said" (H. J. Rubin & I. S. Rubin, 1995, p. 226). An experienced qualitative researcher from a New Brunswick university was engaged to review the plausibility of the categories proposed in the content analysis. A reliability check on the measures, as recommended by Appleton (1995), was performed by having a colleague independent of the research and familiar with parent-child interactions review a random sample of four of the videotapes and perform the same measurements to seek congruence in theme recognition. Any discrepancies were discussed and resolved to mutual satisfaction. This was a way to safeguard for reliability (Kvale, 1996).

Material that related to a theme was sorted and grouped by category with the aid of a word processing program to manipulate the data into category files and to help in the retrieval of segments. All units were coded with a sentence number, date and pseudonym. The frequency of codes was then counted and comparisons were made across categories as per procedures described by H. J. Rubin and I. S. Rubin (1995). The transcripts were read repeatedly for evidence to confirm or refute the themes. Additional sets of data were likewise read, with the named categories in mind and expanded as required.

Questionnaire items were analyzed first by using frequency distributions and then qualitatively. Responses to the open-ended questions were qualitatively analyzed and, where indicated, further probing was done to ensure understanding of parental beliefs and a link to emergent themes. The compiled data were described comprehensively in a case study narrative. Quotations were used for "drawing the reader into the case" (Creswell, 1998, p. 186). Once the within case analysis was completed, a cross case analysis was done to build an explanation that fit all codes. After each case was presented, this process involved seeking "converging lines of inquiry" (G. Anderson, 1990, p. 163).

In the next section I will highlight the specific means employed to ensure the trustworthiness of the study and give credence to the research.