336 ANALYZING DATA
agency. Furthermore, I felt I had to share our analysis with BLM personnel who had provided me
with access to the agency.^2 I wondered whether the study results would have a negative impact on
my relationships with them. I provided copies of our findings to several key people at BLM. Their
responses were mixed. Although one person thought there was some validity to our findings,
others were polite but lukewarm. One person expressed confusion. This was not what he thought
I was working on in my research. He felt that the nature of the study did not match the content of
interviews I was conducting or the nature of my other activities in the agency. More important,
the findings did not reflect his experience with the EISs. With some awkwardness, we moved past
the breach as I explained that the statistical study was distinct from my dissertation research.
Three years passed before I confronted this discrepancy and undertook to analyze the EISs using
interpretive methods.
In assessing the relevance of the EISs to implementing wilderness policy in the BLM, I had to
address the fact that the statistical analyses did not convey much about the processes through
which the documents were generated. They also did not offer insight into how the EISs were
connected to efforts that had been occurring inside the agency since 1976 to make sense of wil-
derness policy. As I faced down these missing pieces, three things occurred to me: (1) in coding
information from the documents for the statistical analyses, the team of students had had more
difficulty with some documents than others because they were organized differently; (2) from my
experiences teaching technical writing, I knew that a different organization meant a different
story line; and (3) if any patterns existed in changes in the organization of EIS documents from
draft to final report, these might provide evidence of a change in the story about wilderness
presented in the documents and about shifts in public lands policy. The interpretive content analysis
that followed took shape quickly.EISS, SENSE MAKING, AND INTERPRETIVE ANALYSISTo understand the range of options available for conducting content analysis, it is useful to recog-
nize that when we write, edit, and read written material, we are working with multiple layers of
documentary organization; that is, words are organized into sentences, paragraphs, sections, and
full documents. We can assess word choices and sentence-level grammar. We might ask whether
the sentences in a paragraph provide a flow from one idea to the next, linked by a main point. We
can consider whether the text is organized into a well-structured outline and what messages that
overall outline sends. Relative to the substantive content of a document, we can evaluate the
information included from the perspective of logic and whether evidence is provided to support
premises and conclusions. We might examine a paper to see whether it conveys contradictory
ideas that are not resolved. We can identify technical language and rhetorical devices, such as
analogies and metaphors. We also can extend our evaluation to consider the context by inquiring
about who produced a document, how, and why, as well as about the intended audiences and
whether and how the text is adapted to communicate with them. It is possible to address these and
other questions in the writing, reading, and analysis of a text.
In my analysis, I focused on the outline or overall structure of the EISs to identify their story
line. I assessed the types of technical information included in them as these related to the story
line, and assessed the structure and normative positions of arguments that linked the technical
information to wilderness recommendations. This evidence became a part of my effort to under-
stand how agency personnel interpret and make sense of wilderness policy. I drew data from 48
pairs of draft and final documents, 1 unpaired draft, and 4 unpaired final EISs for a total of 101
documents. They covered 670 study areas (78 percent of the total) and were chosen based on their