Qualitative Case Study Research in Organizations: Interprofessional Collaboration in Legal Aid Settings

Gaining Access

After initial conversations with administrators of legal aid agencies in the city I had chosen to use as the home base for my research, one organization had agreed to participate and two others had tentatively agreed. Although the attorney administrators of some agencies I contacted saw value in this research, many were not willing to make the accommodations required for me to be present while legal business was being conducted. In the first organization that agreed to participate, I had a personal and professional connection with the administrator who was unwavering in his support for this project. The second organization that consented was also one in which I had a professional association with a staff attorney. The staff attorney was invested in seeing the results of my research and held a number of meetings with the managing partner to explain the study and his reasons for wanting to participate before the managing partner agreed to meet with me. Even with this internal support, it was only with a letter of support from the administrator of the first agency I had recruited that this administrator agreed to participate. Finally, after the other agency that had tentatively agreed to participate ultimately refused, I used personal and professional contacts to reach out to administrators of clinics in other parts of the country. With those connections and the letter of support from the administrator of the first research site, I was able to secure a third subject agency in the Southwest United States. Ultimately, I was able to gain access to three organizations that employed social workers and lawyers which, given the depth of the analysis I was planning, was a sufficient number of subjects for this multiple case study.

Equally as important as gaining access to the organizations was getting buy-in from the social worker in each organization. The primary purpose of these organizations is to provide legal services; therefore, each organization employs many attorneys, but only one social worker. Without the cooperation of the social worker, an agency would be useless to my study. In the first two organizations, I had prior professional experience with the social workers, and they had an interest in the outcome of this work. In the third organization, I did not know the social worker. However, the social worker from the first agency was acquainted with her and was able to introduce me to her. It was certainly much less complicated to gain the cooperation of the social workers, yet their participation was as essential as that of the administrators.

Data Collection

Start dates, but no end dates, were set for data collection in each of the agencies. I assumed that I would be in each organization for approximately 2 weeks, but I did not want to bind myself to a firm deadline should I need or want to spend more time in one of the settings (Glaser & Strauss, 1967). As it worked out, 2 weeks was sufficient time to gather the necessary data in each agency. With a focus on the role of the social worker, my default mode of data collection was to shadow the social worker in each organization. I took notes on what the social worker did, who she interacted with, and what meetings she attended. I scheduled time during the 2 weeks in each organization to formally interview the social worker but also took notes on details that arose in our conversations about her daily activities. I was able to set up meetings with other individual subjects, access relevant documents, and attend a variety of meetings with the assistance of the social worker in each setting. I interviewed interns, fellows, and support staff in each organization, adapting my interview protocol to include questions relevant to their particular role in the organization.

Informed Consent

Even with blanket permission to study these organizations as a whole, for ethical reasons, it was necessary to obtain informed consent from each person I interviewed. In most cases, I was introduced to members of the staff by either the administrator or the social worker so that agency employees knew that my presence in the workplace was legitimate. Every individual I approached about participating in the study agreed, though a couple of subjects seemed reticent to answer all of my questions. Some nuance in the data may have been lost because of these less than forthcoming interviews. However, in each organization, I reached data saturation, which is the point at which no new information was being gleaned from interviews. Given this, I am confident that the less detailed interviews did not have a negative impact on my findings.

Managing in the Field

I organized my days in each organization around the social worker’s schedule. I arrived in the morning when she did and left when she left, or later if I had scheduled interviews or observations. Generally, I collected data for 8 to 10 hr each day. Because I was not using an electronic recording device, I wrote many, many pages of notes each day. I then spent each evening typing notes into text files. While this made for some very long days, it followed best practices for preserving the accuracy of the data. Working with my field notes the same day I created them, filling in blanks in my written record and adding information about context, was the best way to assure that the passage of time did not cloud my memory. I also wrote research memos each evening. In these memos, I answered a standard set of questions such as, “what am I learning?” and “how does this case differ from the last?” These memos were also the place I recorded my thoughts about and reactions to what I had seen and heard that day. The memos were part of a process in which I reflected on the data I had collected each day and generated questions to ask or observations to follow up on the next day in the field (Van Maanen, 1988). In this way, I was able to use what I learned each day to deepen the overall inquiry.

Preparing Data

Creating text files and analytic memos while still in the field collecting data was the first step in preparing data for analysis. Overlapping these processes is consistent with the philosophical underpinnings of grounded theory and case study research. This iterative process of repeatedly going back and forth between the two activities allows for adjustments to data collection while it is ongoing. Preparing data in the field this way meant that there was no delay between the end of data collection and the beginning of data analysis.

Managing Data Text Files

Data were analyzed using N6 computer software developed and sold by QSR International. While there were more updated versions of the software at the time, the practical consideration of cost weighed in favor of using the program to which I had access. This software first offered a system for organizing text files in a variety of ways. I was able to store files in configurations that corresponded to the way I planned to use the data at a given time. For example, I had files stored in separate folders that contained all of the interview data from one site, interview data for all of the social workers across sites, and all of the interview data for attorneys who practiced in the same area of law. It is possible to manage this type of filing system without computer assistance, but the software saved a great deal of time.

Coding

The first step in analyzing qualitative data is coding. Coding is the process of reading data text files created during data collection and assigning codes to significant sections. Codes are one- or two-word titles for ideas or themes found in the text. Codes are assigned to phrases, sentences, paragraphs, or more substantial sections of writing that are relevant to the research question. For this project, I used two different types of coding. First, I engaged in free coding where I worked to set aside preconceived ideas about what themes might exist in the data and read each text file with a mind open to seeing what topics emerged from the text itself. In this process, I created and assigned codes as they emerged from the text. The second round of coding involved looking for a priori constructs based on the theoretical foundations of the study. The interview protocol was designed to elicit information on these particular concepts.

For example, in all three agencies, the social workers supervised interns from local masters programs in social work. In her interview, one social worker described the process she used to assign her students to cases. In the first round of free coding, I tagged these pieces of data—phrases, sentences, and paragraphs—with the code “Assigning Students.” At that point in the data analysis, I had no way of knowing whether this code would ultimately be relevant to answering my research question, but with this concept identified and coded, I could see at a later point in the analysis if it would be significant or not. Theoretical coding was a second process in which I assigned codes to all of the text data from the list of theory-based a priori codes created when the study was designed. As with free coding, there was no assumption that any of these codes would prove significant to the final theory.

Computer-Assisted Qualitative Data Analysis

The data analysis software allowed me to create a nearly infinite number of codes and was flexible in the ways these codes could be used. For example, if an idea emerged in the free coding and was also an a priori construct, it could be coded as both a free code and as an a priori code. The software allowed me to merge those codes when appropriate or use them individually. Each code could be identified by individual subject (i.e., Lawyer A in Organization 1), by organization alone, by profession alone, and so on. With this amount of flexibility, I was able to compare and contrast a variety of codes to discern how they worked together to create overall themes that would be used in later stages of analysis to develop a new theoretical construct.

Within-Case Analysis

At the heart of within-case analysis was a detailed written case study for each site. The purpose of this step was to become as intimately familiar with each case as possible before comparing or contrasting cases with one another. The first step toward writing a comprehensive case study on an individual organization was writing answers to each of the questions on the interview protocol using data from interviews, observation, and document review collected at that agency. Various functions of the software allowed me to collect pieces of text data that had been coded as relevant to each question in one place. Compiling this amount of data was tedious, but it would have been a great deal more cumbersome without the assistance of the qualitative data analysis software. These compiled data were used as evidence to construct responses to the protocol questions. The resulting answers were organized by agency to provide a picture of the emerging themes in each organization and formed the basis for the in-depth case study of each organization. Tentative themes related to organizational influences on social worker/lawyer collaboration emerged from these narratives. These preliminary themes were noted in the narratives, as well as in analytic memos, which also became part of the case study database. Care was taken not to reach premature conclusions about the fit and value of these tentative themes, but they served as a useful starting point for the between-case analysis.

Between-Case Analysis

Between-case analysis involves comparing and contrasting cases with one another by analyzing data in three different ways. This process serves as a tool to avoid reaching premature or false conclusions that may be present when focused on a single case or when using only one method of analysis (Eisenhardt, 1989).

The first strategy in between-case analysis involves selecting themes of interest and looking for within-group similarities and between-group differences for each category. Building on the previous example, I might have chosen the code “Assigning Students.” I used the software to compile all of the data tagged with that code, and with all of these data in one place, I was able to see similarities and differences in how social work students were assigned to cases in each of the three agencies. This process of compiling, comparing, and contrasting data was repeated with all themes of interest.

The second tactic in between-case analysis is to select pairs of cases and to list similarities and differences between each pair. By looking at all possible pairings of my three research sites, I was forced to look for differences where I may have been convinced that there were none and to look for similarities where it seemed there were only differences. An example of this was that one of my research sites was a federally funded provider of legal assistance and another was a university-based legal aid clinic. At first glance, these organizations seemed vastly different in the ways that they were organized, funded, and staffed. However, when tasked with the specific goal of finding similarities, I was able to see that although the organizational charts in these agencies looked very different, in practice their work was organized in a similar way.

The third procedure in the process of between-case analysis involves dividing data from all three organizations by form. In this approach, data forms (interviews, observations, and documents) were examined individually with an eye toward exploiting each for all possible insight. A pattern found in one data form could be strengthened by corroborating evidence from another form. An apparent pattern in two forms could be strengthened or rejected based on conflicting evidence from the third form. In one case, when asked in the interview about their agency’s mission, not one lawyer said anything about interdisciplinary practice as a part of the core mission. However, data from document review showed that the written mission of the organization was to provide legal services in an interdisciplinary team. This analysis strategy was vital in bringing this discrepancy to light.

As the above examples demonstrate, the methodical application of between-case analysis strategies augments information gleaned from within-case analysis. The analytic memos that result from this process form the basis for the next steps of data analysis.

Shaping Hypotheses

The ultimate goal of this process was to construct theory about the influences of individual, organizational, and institutional factors on collaboration between social workers and lawyers. Analytic memos created during within-case and between-case analysis were the starting place for shaping hypotheses and developing theory. Theory was emerging from the data and was fleshed out in these memos. It was then compared with the evidence from each case to assess fit. Key to this step in the process was continually comparing theory and data while iterating toward a theory that tightly fit the data. As is seen in the previous descriptions of the data analysis process, these steps did not proceed in an orderly fashion one following the next. Instead, each step of the data analysis process was visited and re-visited as themes were made more explicit by other stages of the process.

Results

The theme that surfaced time and again in these data was that there was no clear definition of the social worker’s role in any of these agencies. In different ways, social workers themselves, administrators, and attorneys said that they did not have a clear comprehensive idea of how the social worker fit into the legal work of their organizations. It became apparent that many of the areas that lacked clarity were those that had been identified in the scholarly literature as areas of conflict between the professions of social work and law. These data indicate that the fundamental characteristic of the collaboration between social workers and lawyers is not conflict, it is ambiguity. The three social workers in this study experienced the ambiguity regarding their role in these legal settings in one of two ways, either as an inherent problem or as an opportunity. One social worker expressed discomfort with the ambiguity. She wanted her agency to provide more explicit and formal structures for mundane tasks such as scheduling meetings, as well as more abstract activities like negotiating ethical conflicts and deciding when it was appropriate for her to work on a case. The other two social workers experienced the uncertainty as freedom to perform their work as independent and trusted professionals. Each offered examples of the ways that they used the latitude provided by ambiguity about their role to creatively work with clients in ways that would not have been possible in other settings. This finding that the working relationship between social workers and lawyers is characterized by ambiguity rather than conflict has the potential to change both the academic conversation about and the practice of collaboration between social workers and lawyers.

Be Prepared

I had written and re-written my research proposal. I had planned and re-planned my time in the field. In spite of that preparation, things happened in the field that I had not anticipated or that were simply not covered in the books. When this happened, I relied on the fact that I was not only prepared with a research plan, I was prepared with theoretical and practical knowledge which informed every step of my research process. I had read many books and articles about case study research, qualitative research, and organizational theory. I read case study after case study, and I had consulted with researchers who had carried out their own qualitative research. I may not have anticipated the precise dilemmas I would face in the field, but I had the knowledge I needed to make informed decisions when they arose. Was information shared over lunch with members of the staff confidential? Was it legitimate to interview a former employee of the organization because it was not in my original plan? How did I deal with the subject who had second thoughts about participating? It was only because of the preparation I had done before entering the field that I was able to make accurate and ethical decisions about these and other issues that arose. There is no substitute for being prepared.

Be Authentic

I was particularly cognizant of the fact that I was infringing on professional relationships that rely on confidentiality. I was asking detailed questions on private work matters. I was following people around their workplace recording their every move. I was doing all of this with people I had, for the most part, just met and with the end goal of putting all of this information into writing. Negotiating this process was often not about theoretical understanding of the research process, it was simply a matter of relating to research subjects as human beings. People needed to see that I could be trusted with their private information and that I cared about them and their work. Without a basic level of trust, there would have been no research. There is no substitute for being authentic.