Читать книгу Everyday Courage - Niobe Way - Страница 28

After narrative summaries were created for each topic in the seventy-one interviews, I created conceptually clustered matrices in order to detect themes across and within narrative summaries. Miles and Huberman suggest this matrix technique as a way to consolidate and present the data.¹⁴ The conceptually clustered matrix has a simple respondent by topic (or variable) format. All the topics in a matrix are conceptually related to each other. For example, I created a matrix for the general conceptual category “relationships with peers”; within this matrix the topics were relationship with best or close friend(s), and with boyfriend or girlfriend. Beneath each topic, I placed the three narrative summaries for that particular topic (representing each year of the interview) to the right of each adolescent’s pseudonym. For example, next to Marie’s name and under the topic “relationship with best friend,” one would find three entries summarizing her relationship with her best friend for each year she was interviewed (see Appendix B for an example). I created separate matrices for each of the five general conceptual categories: (1) self-perspectives; (2) relationships with family members; (3) relationships with peers; (4) perspectives on school; and (5) views on the larger society.

From these matrices, I detected themes across narrative summaries. Themes were repeated phrases, terms, or concepts that I heard within and across narrative summaries. For instance, when the adolescents spoke each year about their fears of betrayal by close friends, I considered betrayal a theme within the topic “relationships with close or best friends.” When the girls spoke about being outspoken in their relationships in and out of school, I considered this a theme across the topics of relationships with friends, family members, and school. In addition, I also went back to my notes taken during the initial two readings of each interview (the first two readings in the analysis based on “The Listening Guide”). I used these notes along with the conceptually clustered matrices to determine the central themes in the data. Those themes that were evident in half or more of the interviews in any one year were considered common themes. I then searched for evidence of each common theme in the original interview transcript of each participant.