restricted access Stage 2 : Ensuring Accuracy of Results
In lieu of an abstract, here is a brief excerpt of the content:

STAGE2 ENSURING ACCURACY OF RESULTS The accuracy of the results must be a paramount concern for the researcher from the beginning to the end of the study. Therefore, the steps that will be taken in the course of the project to ensure accuracy should be considered at the very outset, as soon as the appropriateness of the case method has been established. The purpose of this vital stage, which should inform all the others, is to demonstrate not only that the results were obtained through a rigorous process, but also that they correspond to reality. The value of a scientific study depends in large part on the investigator’s ability to demonstrate the accuracy of the results. This is particularly true for qualitative research methods such as case studies : since these methods are more flexible, they can attract sloppy researchers who hope to avoid the direct evaluation to which the results produced by quantitative and experimental methods are automatically subjected (Hlady Rispal 2002a, b; Kvale 1987). The concept of accuracy embraces two components: reliability and validity. Reliability relates to the consistency of the observations, meaning the replicability of the results: if the same phenomenon were investigated by other researchers using the same methodology, they should arrive at roughly the same conclusions (Kvale 1987). Validity relates to the connection between the results and reality. A study is valid if the constructs developed by the researcher are good representations or measures of the categories of human experience under observation. Reliability is a necessary condition for validity but does not guarantee it (Bachelor 1992; Light, Singer and Willett 1990). When we conduct a case study for research purposes, we are examining the interactions among a number of variables in a natural setting, often without any preconceptions. The researcher 22 The Case Study as Research Method plays a decisive role at every stage of the process: data collection, analysis and interpretation. Therefore, the researcher’s actions and personal characteristics can have a significant direct impact on the accuracy of the results. There are three classic types of bias that investigators are liable to introduce: the holistic illusion, which consists in ascribing greater connectedness and consistency to events than they actually possess (for example, by ignoring facts that do not fit); the elite bias, which consists in attaching greater weight to information from informants who express themselves clearly than to reports from less articulate informants; and over-assimilation, whereby the researcher accepts the facts and perceptions conveyed by local informants whole cloth, surrendering his or her own vision and critical faculties. Personal characteristics that can introduce bias into the analysis and interpretation of the evidence include individual differences , gender, age, theoretical orientation, the investigator’s history with the object of study, and level of experience (Hill, O’Grady and Price 1988; Landry and Farr 1980). Psychological variables such as self-confidence, level of anxiety and cognitive complexity can also come into play (Landry and Farr 1980). The researcher’s history and expectations can therefore introduce idiosyncratic elements (Bachelor 1992; Beutler and Hamblin 1986). To ensure the accuracy of the results, the researcher’s first task is therefore to detect potential sources of bias or contamination and address them. To do so, the researcher may have to apply measures that can appear contradictory at first glance. It is illusory to think that we can obtain results that are entirely reliable and valid. For one thing, in qualitative research it is difficult if not impossible to demonstrate reliability and validity in a precise, numerical manner, as one can in quantitative research. For another, techniques for enhancing reliability and validity operate, more often than not, by reducing the impact of the researcher’s subjectivity as far as possible, or by providing information to enable others to check the process through which the results were obtained (Guba 1981). This chapter, which is largely based on Lecompte and Goetz (1982), covers the factors that must be borne in mind and the steps to be taken in order to increase reliability and validity. Reliability can be divided into two components: internal reliability and external reliability. Internal reliability means that other investigators would arrive at essentially the same findings if they were to analyze and interpret the data produced by the study. In other words, the conclusions drawn from the evidence by multiple Ensuring Accuracy of Results 23 independent observers and coders would be sufficiently consistent to describe the phenomenon in a similar way and come to the same conclusions about each...