Multiple Sources of Evidence

Conceptual Overview and Discussion

Robert Yin has described strengths and drawbacks of sources of evidence that is often used in case studies (e.g., documents, archival records, interviews, direct observations, participant observations, and physical artifacts), and the need for corroborating and supplementing findings using a diversity of approaches. A brief overview of these issues is provided next.

Documents—such as memos, internal reports, and proposals—though not always accurate in their representation of actual events, are useful for confirming and augmenting information found by other means. Weaknesses such as reporting biases and researcher selectivity are qualified by strengths arising from the potential breadth of data availability and the unobtrusive nature of data collection. Archival information, both quantitative and qualitative, while indicative of trends and contextual perspectives, should also be subjected to the same considerations and concerns associated with document use. Interviews tend to be either open-ended (where interviewees provide their opinions, factual information, directions for further investigations, or additional sources of evidence), focused (where the interview is loosely structured around a set of questions), or formal surveys (where the interview follows a well-structured survey format that can be used in conjunction with additional assessments of the situation). While interviews generate direct insights into issues, biases can be introduced by interviewers as well as by interviewees who respond in ways they perceive are expected of them. A direct observation of processes as they unfold grounds data within a context. This, however, can be burdensome in terms of budgetary and time investments; moreover, a mere realization of being observed is capable of distorting people's behaviors. Participant observation is a variation on this approach where researchers actually assume roles within events and are not passive observers. Finally, physical artifacts are to tangible items such as technological devices and instruments, and allow for analyses of their potential applications.

tradeoffs between strengths and weaknesses, as well as researcher training and expertise in collecting information by these and other means are typically considered while choosing sources of evidence to access. A key benefit of using multiple sources is that of data triangulation where a broad range of inquiries converges on the same set of conclusions. Construct validity (i.e., the extent to which a construct corresponds with variables used to measure it) is also enhanced as additional measures are generated through each investigation. Karen Eisenhardt further discusses the importance of multiple sources of evidence in building theories from case study data.

Application

Two examples of the use of multiple data sources are provided—specifically: (1) the use of interviews, direct observations, participant observations, and documentation by Tim Foster in examining how different levels of an industrial Web site created value for a mining company's supply chain members; and (2) the use of interviews with different types of frontline employees and internal documentation by Loïc Plé in uncovering customer opportunism at a retail bank.

...