Performing Subjectivist Studies in the Qualitative Traditions Responsive to Users - Evaluation Methods in Biomedical Informatics - page 286

To use documents appropriately, the original motivation for the creation
of the artifact must be understood. During interviews, the researcher can
refer to an artifact he may have available and ask questions about it. How
was the text recorded in the document collected? Who collected it? How
was it generated? Why was it generated? Was the artifact creation
mandated or spontaneous? Such questions must be addressed in order to
interpret the meaning of a document.
The NET meets several times to map out strategies for this study. They have
already decided to use observation and interviews, but they need to outline
interview questions and foci for observations, along with a format for field
notes. Although the consultant has agreed to interview a core set of infor-
mants herself to maximize consistency, and although she will vary the ques-
tions for each interviewee, she spends considerable time with the team
outlining six main questions to ask each person. The team agrees that field
notes will include personal, theory, observational, and methods sections when
they are in their final form. The team also agrees that the focus will be on
describing the attitudes and actions of physicians as they use computers in
general, and the EMR in particular, and also on the impact of physician use
of the EMR on other staff. In addition, the research team asks the insider for
a copy of the contract with the vendor and copies of various paper forms as
artifacts. The contract contains privileged information and the request is
denied, but paper copies of provider order forms and order sets, used only
when the system is unavailable, are provided.
Building the Argument: “Measurement” and Analysis
The end product of a qualitative/subjectivist study is a set of conclusions—
the “argument”—built on the data and rooted in relevant theory. Qualita-
tive investigators seek for their conclusions trustworthiness, confirmability,
credibility, and transferability. While objectivist studies also strive for
these characteristics, they are conceptualized, approached, and attained in
qualitative/subjectivist work in distinctive ways that are not dependent on
statistical constructs such as confidence intervals and inference. In qualita-
tive studies, trustworthiness implies total authenticity of findings. Con-
firmability denotes objectivity or freedom from bias. Credibility is akin to
internal validity or gaining a true and believable picture. Transferability is
an analog to generalizability—the degree to which the results are applica-
ble to other contexts. Dependability, like reliability, is the extent to which
the process of the study has been undertaken with consistency and care. 23
In general, approaches to qualitative data analysis employ progressive
abstracting or “progressive focusing” as described in the previous chapter.
Meaning is assigned, through categorization and coding, initially to seg-
ments of raw textual data taken directly from field notes. Initially, the
number of categories is large. Then the categories themselves are analyzed,
aggregated, and reanalyzed to develop a relatively small number of themes.