far more practicable but also let subsequent interviews benefit from prior ones. Once the
interviews were completed I coded the data from the first 6 interviews into 157 fine-grain
codes and then grouped the seemingly related codes into 31 level-2 interpretive codes
which I then used to code the remainder of the transcripts. Next I grouped the level-2
codes by commonalities to create five broad categories: (1) personal biography, (2)
understandings of teachers, (3) understandings of race and racism, (4) understandings of
the job, and (5) understandings of leadership. I explored within-category and crosscategory
connections to surface possible themes