Retrospective Case Study

Conceptual Overview and Discussion

All retrospective case studies have three factors in common: (1) data are collected after the significant events have already occurred, (2) researchers have access to both first-person accounts and archival data, and (3) the final outcomes—which were presumably influenced by the variables and processes under study—are already known when data collection takes place. Although investigators cannot observe events that occurred prior to the outcome (which would have helped them construct a time line of events or changes in variables), they do have access to individuals who did directly observe and/or participate in the events. Retrospective researchers can, of course, include observations about the final state of a process outcome. Todd Chiles and his colleagues did this when they observed the current end state of the tourism industry as it centered around Branson, Missouri, at the end of their time line analysis in 1995, within a retrospective case design that also covered 100 years of regional development.

Retrospective case studies are one of three types of longitudinal case study designs. They are different from concurrent designs, in which data are collected in real time over successive periods as the events and activities in an organization are occurring. The 18-month case study Tyrone Pitsis and his colleagues conducted on large-scale project [Page 825]management practices for the 2000 Olympics in Sydney is one example of a concurrent design. Retrospectives are also different from historical designs, in which the events in question occurred far enough in the past that no relevant informants are available and investigators must rely solely on secondary documents and cultural/physical artifacts. Andrew Hargadon and Yellowlees Douglas's case study of Thomas Edison's efforts to commercialize the electric lighting system in the United States is one example that illustrates the historical design.

These three types of longitudinal case studies differ along two dimensions: (1) whether the events being studied have already occurred and (2) whether researchers have access to informants who were involved in the events or phenomena being studied. Both retrospective and concurrent case study designs typically include informant interviews in the data set, whereas historical designs do not. In both retrospective and historical designs the events have already occurred, and process outcomes are known, but in concurrent designs the final outcomes are not yet known. Retrospective designs are often more efficient than concurrent designs, because the latter require researchers to wait until the passage of time creates the intended changes in processes or variables. Retrospective designs take advantage of the fact that data are collected from multiple prior periods all at once.

Application

Under a variety of research philosophies—positivist, interpretivist, and critical, for example—retrospective case study designs have several practical advantages that make this methodology well suited in at least three basic situations for studying changes that took place over time.

...