In addition, we conducted interviews with the principal, three self-contained classroom teachers (i.e., the teacher of grades 1 and 2, the teacher of grades 3 and 4, and the teacher of grades
5 and 6), three high school teachers of science and mathematics, the middle school math and
science specialist, the music and drama teacher, and a place-based educator hired by a local
We structured the data analysis in a way that would enable us to identify and verify the salience
of emergent themes pertinent to the two research questions guiding the study: a) What school
and community dynamics support and sustain place-based education? and b) What school and
community dynamics threaten or constrain place-based education? Using the qualitative analysis
software program, Atlas-TI two members of the team analyzed data from interview transcripts,
field notes, and documents. We used inductive coding and thematic analysis following guidance
from methodologists such as Boyatzis (1998) and Miles and Huberman (1994).
Initial coding supported two promising approaches to conceptual categorization of the data.
The first approach used categories for organizing all of the information about specific place-based
approaches to instruction. This categorization of the data was particularly relevant to the detailed
picture of instructional practices that qualitative case studies of schools are able to provide. It
most resembled what qualitative methodologists refer to as “content analysis” (e.g., Neuendorf,
2002). The second approach to categorization addressed the research questions more directly by
organizing data thematically to reveal the dynamics that supported and sustained place-based
education and those that threatened and constrained it. We relied on a variety of techniques of
thematic analysis such as those that Miles and Huberman (1994) discussed in their sourcebook
on the topic. Of particular value were matrices that helped us identify patterns—systematic
uses of language or descriptions of practices that were evident across participants and data
sources.
Using matrices, for example, we assembled all statements about dynamics relating to support
for PBE, opposition to PBE, implementation issues, and so on; then we created various conceptual
frameworks with possible explanatory value. Propositions characterizing each such framework
represented tentative themes. Finally, with the tentative themes in view, we returned to the
interview transcripts, field notes, and other documents to determine the extent to which actual
statements from participants or other evidence supported the salience of each. This approach
added an important product to our audit trail, providing a way to verify that the data actually
supported the themes we identified as salient (see e.g., Lincoln & Guba, 1985). The team found
sufficient evidence to support four of these themes as important to an understanding of PBE at
the school and therefore as directly responsive to the research questions.