Planning Case Study Research

The elements and sequence of case study preparation are presented here so that the reader can determine the important characteristics of case study planning. It should be noted, however, that these plans are likely to unfold in many different ways depending on the case to be investigated and the study questions to be answered. The case study examples provided here and elsewhere in this chapter are excerpted from a case investigation of a dental hygiene student tutor program conducted by the first author (Moore, 2009).

Conceptualizing the Study

The first step is conceptualizing the study, which includes clarifying the purpose, defining and limiting the case, identifying the questions, and considering potential audiences for the report (see, for example, Exhibit 10.1). This step helps the researcher verify that case study research is appropriate for what the researcher wants to know, or the purpose of the study. To accomplish this step the researcher asks such questions as these: What do I want to know about the case? Why do I want to know this? How will I limit or bound the case? Who else wants to know about or cares about this? This initial step influences everything else the researcher will do.

Bounding or Limiting the Case

As described previously, the case study researcher must define, limit, or determine the boundaries for the case. The case is often bounded by time (for example, examining the initial six months of a college nursing program) or by place (for example, examining an at-risk youth program as implemented in one community).

Writing Study Questions

Study questions provide the structure to capture the essence of the case in its context. The researcher asks, What do I want to know about the case? Asking good questions is one of the most important things the case study researcher does because the questions focus the inquiry and determine the plan. The researcher considers the questions during each aspect of the study. Decisions about the types of data gathered and the strategies used during interpretation and analysis depend upon the study questions. The researcher also considers the stakeholders in the study when crafting questions. The researcher asks, Who cares or wants to know about this? What do they want to know? Who will read the case study report?

Before defining the study questions, the researcher may choose to identify what is already known and what needs to be known about the case and its context. Theoretical or conceptual frameworks pertinent to the case can guide development of the study questions and may be discovered through a review of pertinent literature. A theoretical or conceptual framework explains or suggests a relationship between concepts or ideas. Sinclair (2007) likens a theoretical framework to a map or travel plan. Before undertaking a journey to an unknown place it is helpful to learn from the previous experience of others who have been on similar trips, to hear their suggestions about what to bring and what to expect. Likewise, before initiating a study, the researcher reads accounts of similar case studies by other researchers and discovers possible links and predictions of how those researchers' concepts or ideas may influence the case he wishes to study (see Exhibit 10.2).

Stake (1995) claims that a good study question will “direct the looking and the thinking enough and not too much” (p. 15). Case study research questions often begin with how or why (Yin, 2003). For example, How are patients screened for the mental health facility? Why are these criteria used? The researcher begins to formulate questions about the situation or problem to be studied, keeping in mind that a study question is bigger than simply what one source might reveal about the case. Questions should therefore be focused on the content of a program, not on data sources. One should ask such questions as How are participants selected? rather than What do the participants think of the selection process? Note how Moore (2009) wrote the questions for her tutor study in Exhibit 10.3.

Research Plan: How to Find Answers

Once the case is defined and questions are determined, the researcher can begin to plan the details of the study by asking, How can each question be best answered? Who has the needed information? What observations are needed? and Are there documents to review? The case study researcher looks at each study question to determine what data are needed and how best to acquire them. The plan will include

• The data that are to be collected
• From whom the data will be collected
• How, where, and when the data will be collected
• Who should collect the data

What Data Will Be Collected?

The researcher determines the types of data needed to best answer the research questions. Although there are no set rules about which types of data to use in case study research, the case study purpose is often to describe, illuminate, or provide insight, which will most likely require a substantial amount of qualitative data. Case study researchers can observe persons and things, such as participants, activities, interactions, and conversations. Case studies may also examine individuals' thoughts, feelings, and experiences, which are not easily observed and may be best revealed through interviewing key informants. For example, in the tutor case study, tutor interviews helped reveal what the student tutor role meant to the tutors. These are complicated ideas that are best communicated in complete thoughts and words to obtain their full meaning.

Case studies may also examine behaviors, which are easier to observe and can be described and portrayed through more quantitative means, such as a tally of events. Quantitative data might also include results from such instruments as tests and attitude measures. It is the researcher's task to strike a balance between quantitative and qualitative data in obtaining the best answers to research questions and communicating the case to stakeholders and other audiences (see Exhibit 10.4, which elaborates on question 2 from Exhibit 10.3).

The case study researcher creates a plan that incorporates a variety of data gathering methods to answer the questions. Triangulation, or finding agreement among evidence collected from multiple sources and using various methods, increases the validity and trustworthiness of findings. When the researcher obtains similar findings through two different methods, such as interviewing and observation, that information is considered more trustworthy or credible (for example, if interviews indicate that patients spend a considerable amount of time in the waiting room, and visits to the facility reveal standing room only). However, if the researcher observes something different from what an interview reveals, more investigation is needed to understand the discrepancy (for example, if interviews indicate that patients spend a considerable amount of time in the waiting room, but visits to the facility reveal an empty waiting room). Likewise, if two different sources agree about something (for example, if patients and doctors agree that the health care system is meeting the needs of the community), that information is more credible than when two sources disagree (for example, if doctors feel the health care system is meeting the needs of the community but patients do not).

From Whom Will Data Be Collected?

An informant is a data source, or someone who knows about the case and can help the researcher learn about the case. Stake (1995) defines an informant as “someone who knows a lot about [the case] and is willing to chat” (p. 67). The case study researcher creates a plan that incorporates collection of data from multiple sources and various perspectives to answer the questions. For example, a case study of a police academy training program would most likely include seeking data from the trainees and trainers, and could use interviews, questionnaires, and direct observation. In the dental hygiene tutor case study, this use of multiple methods and sources included interviewing and observing the tutors and those tutored.

Case study researchers often identify participants using purposeful sampling as opposed to random sampling. With random sampling, each item or person has an equal chance of being selected for study. For example, participants may be selected by drawing names from a hat. Purposeful sampling lends more strength in case study research because data sources, participants, or cases are selected by how much can be learned from them. This approach is described as seeking “information-rich” sources (Patton, 1987, p. 58) rather than producing representative samples. Additional considerations about sampling are outlined elsewhere in this text (see, for example, Chapter Four).

The case study researcher considers how many or how much to sample in order to answer the questions. Sometimes an estimate is made prior to the study, and then actual numbers are determined based on the data that are obtained. For example, the researcher may continue to search for data until saturation is reached, that is, the evidence becomes redundant, with no new information coming in (Lincoln & Guba, 1985).

How, Where, and When Will Data Be Collected?

The case study researcher concentrates on the case, collecting data from various sources using various methods for days, weeks, or perhaps even longer. The researcher must decide whether a longitudinal or cross-sectional approach will best reveal the complexities of the case. As previously discussed, longitudinal case studies are an exception in regard to time frame in that they may last for months or even longer depending on the purpose of the study. For example, if studying a health care system, the researcher must decide if he should spend an extended amount of time following the same patients over time from their acceptance to release (longitudinal), or if he should study different people at various stages, including some patients who are being accepted, some who are being treated, and others who are being released from care (cross-sectional).

Once the researcher has determined the kinds of data required and the data collection strategy, a timeline can be constructed. Depending on who and what are being studied, it may be necessary to make contacts and ensure access to the study site, a process similar to that required for ethnographic research (see Chapter Seven). The case study researcher also makes plans to protect participants, secures their consent to participate in study activities, and obtains necessary approvals (for example, from an institutional review board as discussed in Chapter Two).

Who Should Collect the Data?

The case study researcher needs a variety of skills in addition to information about the issues related to the case. A review of the literature helps familiarize the researcher with the case, the context, related issues, and theoretical or conceptual frameworks that can inform the study. Prior to conducting the study, a researcher who is new to case study research may need to acquire knowledge and gain skill in the methods commonly used in qualitative research, including interviewing, listening, observing, describing, and interpreting. An example from the tutor study may be useful (see Exhibit 10.5).

According to Seidman (1998), “There is an inherent paradox at the heart of the issue of what topics researchers choose to study” (p. 26), because researchers' interest in a topic is often related to how closely they are involved with it. Qualitative designs often call for the “persons most responsible for interpretations to be in the field, making observations, exercising subjective judgment, analyzing and synthesizing, all the while realizing their own consciousness” (Stake, 1995, p. 41). It is important for the researcher to carefully examine and reveal her researcher position. If the researcher has a close relationship or a past history with the case being studied, this information should be made transparent. Researcher biases or predispositions can be made explicit in a bracketed interview prior to the study. In a bracketed interview, the researcher reveals what she believes the study may reveal, what she thinks the answers to the study questions might be, and other potential biases or beliefs that can influence the researcher's interpretation of study findings. The researcher and case study audiences must examine more carefully any results that match the researcher's preconceived expectations.

In addition to revealing and documenting predispositions and biases, depending on her position the researcher also considers who is best to collect the data. It may be necessary to involve multiple investigators. When different observers find similarities, there is strength; where they differ, there is a need for further investigation. Exhibit 10.6 describes how Moore (2009) minimized bias in her case study.

Research Protocols and Instruments

Design and choice of protocols and instruments will have a great impact on the overall quality of the study. For example, the researcher wants to select or design an instrument that is valid, or measures what it intends to measure. Standardized instruments, such as questionnaires or observation protocols, may be identified from a review of the literature or special sources that report research tools (see Chapter Four). If an appropriate instrument cannot be identified (which often happens), the researcher may need to create an original instrument, or modify an existing one, to meet the study needs.

Surveys and Interviews

Interview and survey instruments are often used in case study work. Good interview and survey questions should be neutral rather than leading or implying answers. For example, “How have students reacted?” is much better than “What do students dislike?” Also, each question should ask only one thing rather than multiple things. For example, “Tell me about your training” is better than “Tell me how you were selected and trained.” Open-ended questions will result in more detailed and useful data than questions that can be answered with a yes or no. For example, “What things are you trying to achieve with your students?” will lead to entirely different types of data than “Do you feel your main goal is to teach content?” (a good example of a leading question as well). Some styles of questions generate even richer data. These include (1) hypothetical questions, such as “Suppose that I am a student who is resisting the role, what would you say to me?”; (2) devil's advocate questions, such as, “Some would say that students shouldn't be tutors; how would you respond to that?”; and (3) questions that ask for interpretation or speculation, such as, “How would you say that tutoring was different than you expected?” or “If you were the director, what changes would you make in the program?”

In addition to using valid instruments for data collection, the researcher also specifies how the instruments will be used. Interviews can be highly structured, with specified wording and order, or conducted more like informal conversations, with room for flexibility or exploration. In either case the researcher's goal is to use the best approach to achieve the purpose of the case study. The order of the questions asked as well as the specific wording can also have an impact on the results.

Once data collection instruments are selected or designed, the researcher tests them. This helps ensure success in obtaining the desired information and allows problems to be worked out prior to the study. A pilot test may be as simple as sharing the instrument with a colleague to ensure that questions are worded appropriately and are understood and interpreted as intended. A field test involves using the instrument on participants similar to the actual participants who will be studied. Exhibit 10.7 explains this process in the context of the tutor study.

Moore (2009) developed an interview protocol, or guide, for her tutor case study, which included instructions as well as the questions. The interview guide was designed to allow for exploration of participants' experiences. Exhibit 10.8 is an excerpt from the revised tutor interview protocol, after pilot and field testing.

Observations

An observation plan helps the researcher consider what exactly to observe, when, and for how long. In case study research the researcher observes and describes details, including specific observations about the context—the surrounding environment and physical, political, and social setting.

Depending on the purpose of the study, the observation can require various amounts of structure. Perhaps the study will require that the observer complete a specific protocol or instrument that has been validated for use. An educational example of an observational tool is Flanders Interaction Analysis (Flanders, 1973). This protocol asks the observer to count and code teacher-student interactions using different numerical codes depending on whether a teacher uses direct influence (such as lecturing or giving directions) or indirect influence (such as asking questions, praising, or encouraging), and depending on the nature of student talk (such as responding, initiating, or silence; see also www.nova.edu/hpdtesting/ctl/fia.html for more information).

Some case studies call for less structured observations. For example, an experienced observer may begin with a blank page, draw maps and diagrams, and record whatever she feels is noteworthy. This offers details of the unique surroundings that could not be known until the observation begins.

The researcher must differentiate between what is observed, or what is happening in objective terms as others might see it, and what is interpretation. It is important for the researcher to note impressions and questions that arise in addition to describing more concrete observations, and to understand the difference. For example, “He began talking louder than everyone else and pointed his finger as he said, ‘You will be hearing from me’” is descriptive; “He acted like he was angry and threatened someone” is interpretive.

Data Recording

Prior to the study, the case study researcher also chooses a method for recording information from interviews and observations. For example, will the interview be audiotaped, or will the interviewer take notes during or after the interview? Perhaps she will do both. Will observations be videotaped, or will the observer take notes or make tallies in real time? These decisions are made based on the nature of the observation. For example, if there are many things to observe at once, videotape, which can be reviewed multiple times, may best facilitate this but also requires additional time (see Exhibit 10.9). Studies suggest that people become accustomed to the camera quickly when there is no operator behind it and when people are intensely involved in what they are doing (Jordan & Henderson, 1995). If the observation is more targeted, it may make more sense to observe in real time (live observation). As with many studies that use direct observation in which the researcher is present, the researcher may actually serve as a participant observer and interact with participants, but it is more common for the observer to be as unobtrusive as possible to reduce her influence on the natural situation being observed. (See Chapter Four on methods and Chapter Seven on ethnography for more about participant observation.)