The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
Exploring Innovation and Quality Improvement in Health Care Micro-Systems: A Cross-Case Analysis
Personal insights by the researchers are the essential information derived from the interview data and they are critical to understanding the complexities of micro-systems and the organizations in which they are embedded. However, the research must approach the phenomenon under study with what Patton calls “empathic neutrality.” 18 To be neutral to the findings means not approaching the phenomenon with a set of preconceived ideas. That means one approaches micro-systems with a desire to learn about them as interrelationships emerge.
In qualitative research, it is important to separate the description of the data from the interpretation of the data. Geertz 19 and Denzin 20 discuss “thick” and “thin” description. “Thick description” depends on presenting descriptive data or recording verbatim comments so that researchers can make their own interpretations later. “Thin description,” on the other hand, summarizes the facts without including any of the context. Thick description sets up analysis and makes possible interpretation. 21Appendix A shows examples of each type of description. For this research, thick description was used and later coded. Each micro-system was recorded and presented in sufficient detail so that the micro-system, or “case,” could be understood in its local context.
This study used two methods: first, descriptive summaries of the interviews derived from thick description; and, second, cross-case analysis. Cross-case analysis offers a way to reconcile the need for “thick description” of uniquely individual cases yet captures the themes and patterns that emerge across cases. 22 Two approaches to cross-case analysis are available: case-oriented analysis and variable-oriented analysis. 23 A case-oriented approach starts by considering each case as its own entity. Only after understanding the relationships, configurations, associations, and the like within the case does the researcher move to a comparative case analysis. The goal is to discover the underlying themes, similarities, and associations that hold across cases.
A variable-oriented approach to cross-case analysis starts with a framework of several variables or themes that cut across cases. For example, variables that may be relevant to a study of health care micro-systems may be the use of information, the role of information technology, or coordination of patient care. Although the study starts with key variables in mind, the variables may evolve and be clarified as the study progresses and as cases are included in the analysis. The variable-oriented approach is more conceptual and theory-centered from the beginning, and less emphasis is placed on the specific details of a particular case.
Neither approach to cross-case analysis—case-oriented or variable-oriented—is necessarily better. As Miles and Huberman point out, the process is one of alternating, combining, or integrating methods as a study progresses. 24 They suggest a mixed strategy that combines the two approaches and uses a “stacking” technique. Such a process was used in this study. To use this technique, the researcher writes up a series of cases using a more or less standard set of variables. Matrices are used to display the data for each case. Without losing any of the individual case-level data, cases are then “stacked” in a “meta-matrix.” Analysis continues by systematically comparing the stacked cases and condensing the meta-matrix.