Because they are single instances or pertain to a single person, case studies have limited generalizability. There are ways to check that the information provided by the study is reliable and valid.
|A description from a single observer should be regarded as tentative. When three or four people independently provide similar accounts, it seems more substantial. Cross-checking the accounts of independent observers is one means of assessing reliability in a case study. For example, a tornado victim may tell the researchers how he tried to help other people after the twister struck. Other accounts may indicate that the first individual was dazed and groggy and had to be led to safety. Do not conclude that the first respondent is an unreliable liar. Distortion becomes a topic of investigation. How consistent are people's accounts of their own actions during and after a calamity? We are led directly to a study of how people put unwelcome thoughts out of their minds.|
When doing a case study, do not challenge the respondent even when a account seems biased or incorrect. You are not the district attorney conducting a cross examination. The goal is to obtain the respondent's point of view. Verification comes afterward. You must convey the impression of being a good listener. The respondents' knowledge that you have access to neighbors, newspaper reports, and public records will keep them from straying too far from the truth. Deliberate collusion among townspeople in concocting a false story is a possibility, but does not occur often.
Differences among accounts may reflect the way each person saw the situation. The person who spend several days in an evacuation camp after a tornado may have different view of events that the person who stayed in town removing debris. Each of the views is accurate but is based on a limited range of experience. Inconsistencies in people's narratives can provide valuable leads. The researcher is a bit like the psychoanalyst who pays special attention to distortions and omissions. Asking for further elaboration is a better way of internally checking a story than challenging people as to their truthfulness or objectivity.
Cross-verification is also possible using a multimethod approach. Perhaps better than all other techniques, the case study lends itself to the use of multiple sources and techniques for gathering information. The validity of a case study is enhanced by using multiple approaches and then integrating the information through a process of triangulation or converging operations (coming at a problem from different directions using independent research methods). Interviews are usually an intrinsic part of a case study. Other useful techniques are observation, trace measures, and the analysis of public and private records.
Systematic observation can be used to confirm information suggested by a case study. For example, a case study of a child might be supplemented by systematic observation of its behavior in the classroom, or in other setting. Participant observation can be used. Although there can be problem in being objective, some researchers study organizations to which they belong, and thus operate as participant observers [see module].
Public and private records
When a case report involves a newsworthy event, other records may be available. The researchers who studied the koro epidemic found hospital records that documented earlier occurrences of the condition. Secondary data sources may be available for shedding additional light on the case under study. Archives such as census statistics, school district records, crime reports, etc. may be relevant to a case study. Often, diaries are included as part of a study of an individual. Or someone studying a natural disaster may compare diary entries before and after the event. Other personal documents described in the personal documents section may be useful.
Even with cross verification, generalization from a case study is limited. Often an event is selected because it is atypical (rare). No matter how many people were interviewed in Guangdong, China, and no matter how much time was spent in the area by the research team, it was only a single koro epidemic. It is appropriate to draw conclusions from the data (e.g., the panic was brief in the vast majority of cases) but the findings cannot be generalized to other koro outbreaks without further study.
Because so much depends on the researcher's personality and approach, a case study is difficult to repeat. The situation of two researchers conducting independent case studies of the same event is unlikely.
When case studies take place after the fact, the researcher must depend upon people's recollections of events. After a crisis, memories are likely to be selective and distorted. With dramatic events, behavioral effects may continue for years, and it will be difficult to determine when the study should end.
triangulation - technically, the process of pinpointing a transmitter by taking a bearing of the transmitter from two or more fixed points. When the bearings are drawn on a map, the point of intersection is the exact position of the transmitter. Here, it refers to "taking a bearing" by using a combination of independent research methods to study the same phenomenon.
Next section: Summary