Implementing Case Study Research
Once the researcher determines what to study, how and when to collect the data, and who will collect the data, and once all necessary approvals have been obtained, the plan is implemented. As Yin (2003) points out, there are no “routine formulas” (p. 57) for conducing case studies. With case study research, however, the researcher may learn something early in the study that can inform and improve the approach taken in the remainder of the study. It is important for the case study researcher to have a plan but also to remain open to discovery of the unanticipated or unexpected. For example, although the researcher is focusing on job placement in an unemployment office case study, he may also find that clients are offered more economical ways to purchase needed groceries. This is an instance of the case study researcher's being flexible in looking for certain answers but remaining open to other findings. The researcher is never sure how the case will unfold, what he will find, or what it will mean.
Data Collection
The attitude of a case study researcher during data collection remains one of an inquirer who is truly curious about the nature of the case, always searching for understanding and answers. The researcher keeps the purpose and study questions in mind and focuses on the case during data collection. The researcher understands that she is not the expert on the case, that the participants are the ones who have the information needed. Understanding elements of the case from participants' perspectives is vital as the researcher remains naive, open, non-threatening, and nonbiased. The researcher watches closely and thinks deeply, endeavoring to understand the context and the issues, searching for meaning in behaviors and other observations.
The researcher strives to ensure the validity of information collected. When information is revealed, for example through interviews, the researcher attempts to corroborate what is said through observations or document review. When one source shares information, the investigator searches for confirmation from other perspectives. When multiple sources agree, the evidence is considered more trustworthy or valid.
Like other qualitative approaches, case study research often generates a large amount of data. This high volume of data requires careful management and regular writing of field notes or keeping a journal with important information, such as the date, location, and people present at each observation or interview. The researcher seeks to maintain a “chain of evidence” (Yin, 2004, p. 85) so that any findings can be traced back to the collected data in their original, raw form.
During data collection the researcher is primarily describing, but may also make notes about potential hunches concerning the meaning behind what is observed or said. The researcher also makes notes about what she is thinking during data collection and early analysis, in the form of dated memos that eventually become part of and inform the analysis and interpretation. The two-column approach is a method of journaling commonly used in case study research (see Table 10.1). Descriptions and objective data are recorded in one column, and notes of potential meanings or interpretations of the findings are recorded in the second column.
Table 10.1 Two - Column Journaling Template
| Description | Interpretation |
| What is happening in objective terms, as others might see it | Possible meaning or meanings; other questions that arise |
| Tutor Study Example: The tutor was looking down rather than looking at the students. | Tutor Study Example: This could mean the tutor is not interested or not listening to the students. It might also be a strategy the tutor is using to discourage students from depending on her for approval and direction. Do any other tutors do this? |
Data Analysis
In case study research, as in other methodologies included in this text, data collection and analysis ideally occur simultaneously in a dynamic and interactive process. Data collection is an important part of the process, but it is useless unless the researcher can make sense of the data; this is the goal of data analysis. For example, in a case study of a community program for teenage mothers, the researcher would be collecting data from participants and those who manage the program, and would want to carefully analyze these early data for clues about what subsequent issues to pursue and which data sources to select. Learning from initial data that certain individuals are asked to leave the program should cause the case study investigator to thoroughly follow up on this issue.
The initial level of analysis often involves coding, or classifying, qualitative data from observations, interviews, and other sources. Analysis literally means pulling things apart to examine them in their smallest components. The researcher deconstructs information and then puts things back together again in a more meaningful way. By dissecting the various parts, the researcher assigns meaning to them.
The researcher constantly compares the data, incidents, interactions, or remarks for properties, such as similarities and differences, that can help identify categories. The researcher assigns codes to the various categories. Merriam (1998) provides an example of coding and categories using food items at the grocery store. She suggests comparing each food item to other food items, thereby generating categories related to their characteristics. In this way food items could be classified into categories and subcategories. For example, an orange might be classified in the fruit category and also in such subcategories as citrus and domestic. Codes help the researcher sort and organize the data, just as file folders can help with organizing a stack of papers. When the researcher sees similarities between various components, these components will be assigned the same category or code. Codes a researcher generates should relate to the study purpose and be conceptually congruent (see Exhibit 10.10). Coding is just one level of analysis; Chapter Three provides a much more thorough discussion of the process of qualitative data analysis.
Often data collection and analysis reveal information that could not have been anticipated prior to the study and that can change the direction of the study. During the early stages of analyzing data the case study researcher makes notes about changes or additions concerning new data to collect that are not in the original study plan. The researcher notes initial insights or hunches, and then collects the data needed to confirm or disconfirm them. This evolving nature is one of the strengths of case study research design.
TUTOR STUDY CODING
In addition to coding or classifying the nature of tutor interventions as either statements, questions (with subcategories of deep or surface), acknowledgments, or clarifications and confirmations, the researcher also classified tutor interventions according to whether their emphasis was on process, content, or social issues.
Tutor facilitation style when requesting or encouraging students to act was also classified as directive, suggestive, or empowering.
- A directive facilitation style was one whereby the tutor decided which direction students were to take or directly told the group or student what to do (for example, “Look up this,” or “Don't focus on that”) or controlled the process without giving options or choices (for example, “Why don't we just do this,” or “We'll start with you”).
- A suggestive facilitation style was one whereby the tutor suggested a single direction to take but used softer language (for example, “Maybe you could do this”) or left the decision somewhat open (for example, “We could do it this way,” or “Does somebody want to do that?”).
- An empowering facilitation style was one whereby the tutor encouraged students to make their own decisions (for example, “How do you want to do it?” or “Do you know what to do next?”) or offered more than one option from which to choose (for example, “You can do this first or do that first”).
During data analysis the case study researcher also makes a concerted effort to remain open to findings that are contrary to preconceived notions identified prior to the study, and he attempts to disconfirm his own interpretations (see Exhibit 10.11). If no contrary evidence can be found, the researcher's interpretations are more strongly supported.
Data analysis leads the researcher to the findings that will be reported, most often including descriptions of the context and identification of meaningful themes or patterns, and sometimes including the application of models and theories to the data and the case, as discussed in the next section.
TUTOR STUDY CONTRARY FINDINGS
Tutor facilitation style appeared to support the researcher's prestudy belief that tutor behavior would improve over time; however, week 9 showed tutors becoming more, rather than less, directive. Wondering if time pressures placed on tutors during week 9 had obscured a real trend, the researcher analyzed transcripts of observations from week 8. Her initial assumption, that tutor facilitation style might become less directive over time, was disconfirmed when the number of directive interventions per hour for week 8 also increased.
Validity
Tactics to improve the validity and trustworthiness of case study findings may include triangulation (the collection of data using a variety of methods [for example, interviewing, observing, and reviewing documents] and multiple sources [for example, tutors, tutees, and administrators]) and using piloted and field-tested (or standardized) protocols for interviews and observations. Conducting appropriate data analyses; examining researcher preparation and bias (for example, determining the extent to which the researcher's preconceived beliefs may have influenced the study findings); member checking (for example, reviewing draft findings by key informants to see if they affirm the validity of the report and recognize their contribution); and, if necessary, undertaking an external review and interpretation are also important measures to improve the validity and trustworthiness of case study findings. In the tutor study, member checking was used to ensure validity (see Exhibit 10.12).
REFLECTION QUESTIONS
- How might you explain the processes of data collection and analysis in case study research to a friend who has not read this chapter?
- What are the most important characteristics that a case study must have for you to trust its findings? Why?
TUTOR STUDY MEMBER CHECKING
After initial data analysis, but prior to final analysis and report writing, tutors responded to preliminary patterns that had emerged at a focused group interview. Following a presentation, tutors were invited to react to and interact with the findings. The tutors contrasted the researcher's findings with their perceptions of the tutor role and experience. They also identified discrepancies between the preliminary analysis and their perceptions, which enabled the researcher to return to the data, applying new insights to a final analysis.