Blending the Focused Ethnographic Method and Case Study Research: Implications Regarding Case Selection and Generalization of Results

Introduction

In recent times, some controversy has arisen about whether ethnographic research could fit within the definition of case study research (CSR) ( Fusch et al. 2017; Parker-Jenkins 2018). This article is based on the idea that the definition and practice of current ethnography is sufficiently broad to allow it to be blended with CSR. In fact, what we propose is a methodological blending of the focused ethnographic method (FEM) with CSR in the sense understood by Yin (2012, 2014), which are both presented in specific sections below. On the basis of a blending of this type carried out in a study of auxiliary work processes in a hospital in Barcelona, Spain, we reflect on the advantages that incorporating CSR logic and principles in the FEM offers in relation to the quality of data, the ability to make comparisons and the transferability of findings. To our knowledge, the implications of this type of blending have not yet been addressed in the literature.

The article concludes that, as previously reported by Handwerker (2001), Knoblauch (2005), G. H. Pelto et al. (2013), and Wall (2015), the FEM is very useful and efficient for studies that start from a delimited topic, a clear and focused research question, and explanatory factors that are well identified in advance because it allows a large amount of high-quality, high-density data to be collected with a relatively short-term fieldwork. Also, incorporating the CSR logic in the FEM by using theoretical criteria in the design phase to select and compare cases offers additional benefits. First, it guarantees that the cases selected are not unique and idiosyncratic, thus ensuring their representativeness regarding the phenomena under study. Second, it ensures the theoretical relevance of the collected data. Third, it allows robust conclusions to be obtained, thanks to the comparison of cases, offering good control of the influence of explanatory factors. Finally, it improves the analytic generalization (Yin 2014) and transferability (Lincoln and Guba 1985) of the findings and conclusions obtained.

The article is divided into four sections. In the first section, we explain the characteristics of the FEM and CSR and discuss the benefits of incorporating the CSR logic in the selection and comparison of cases and settings. In the second section, we explain how the fieldwork in a study of auxiliary work processes in a hospital in Barcelona, Spain, was undertaken following this logic of blending the FEM with CSR. In the third section, we discuss the benefits of the methodological strategy explained in the second section. Finally, in the fourth section, we present the conclusions.

The FEM and CSR, So Close Yet So Far

The FEM

Ethnography currently shows a wide range of developments within the social sciences (see Bernard [2011] for a review). Some of these developments, such as focused ethnographic studies (Gove and Pelto 1994; G. H. Pelto 2017; G. H. Pelto et al. 2013:37–28) and focused ethnography (Knoblauch 2005), have gained popularity in recent times in several research fields. In the anthropology of health and nutrition (e.g., Lee et al. 2019; Thuita et al. 2018), specific social, economic, and cultural factors affecting illnesses and nutrition in different cultural contexts have been identified to design more effective interventions (see G. H. Pelto [2017] for a review). In health care research (e.g., Bikker et al. 2017; Hales et al. 2018; Higginbottom et al. 2013; Muecke 1994; Wall 2015), it has been shown that people from various cultures, subcultures, and groups integrate health beliefs and practices into their lives differently, and the social and cultural factors shaping health-care use and delivery have been identified (see Cruz and Higginbottom [2013] and Vindrola-Padros and Vindrola-Padros [2018] for reviews). Finally, in workplace studies, it has been shown that technological work systems are inseparable from the workers’ local knowledge and reasoning, which produce the work actions and activities and make them understandable (see Luff et al. [2000:xii–xvi] for a review).

Here, we refer to all these approaches as the FEM. The most outstanding features of this method are the delimited and focused scope of the research question, the short-term fieldwork, and the use of intensive data collection strategies (see Table 1 for details). The FEM can be understood as a specific type of rapid ethnographic assessment method or procedure (Bernard 2011:264; P. J. Pelto 2016:267–80; Vindrola-Padros and Vindrola-Padros 2018), among which we also find mini-ethnographies (Leininger 1985), microethnographies (Wolcott 1990:64–66), and quick ethnographies (Handwerker 2001).

OPEN IN VIEWER

Table 1.

Main Differences between Conventional Ethnography and the Focused Ethnographic Method.

Conventional Ethnography	Focused Ethnographic Method
Open approach	Focused approach
Non-predefined variables	Few previously defined variables
Long-term field visits	Short-term field visits
Experientially intensive	Data-/analysis-intensive
Open data collection	Targeted data collection
Time-extensive	Time-intensive
Open observation	Semistructured observation
Solitary data collection and analysis	Group data collection and analysis

Source: Drawn up by the authors, building on the work of Creswell (2009), Handwerker (2001), Knoblauch (2005), G. H. Pelto et al. (2013), and G. H. Pelto (2017).

The FEM should not be understood as being opposed to other current ethnographic developments. It can sometimes be used as an alternative and sometimes as a complement to other more conventional ethnographic approaches that follow the tradition of classical ethnography more closely. In these more conventional ethnographic approaches, the data are obtained through direct contact with the group studied, with whom the researchers usually spend time in a long-term fieldwork project and usually make a precise and detailed description of it (normally with some type of participant observation accompanied by interviews and document analysis).

In Table 1, we summarize the differences in objectives and procedure between conventional ethnography and the FEM, building on the work of Creswell (2009), Handwerker (2001), Knoblauch (2005), G. H. Pelto et al. (2013), and G. H. Pelto (2017). Whereas conventional ethnography usually aims to give a holistic view of the cultural values and shared behaviors of social groups (Creswell 2009:177, 196), the FEM focuses on thematically more delimited and concrete subjects of research such as communicative actions, the use of technology, and feeding practices. In addition, following Handwerker (2001) and his proposals in relation to quick ethnographies, it could be added that to achieve high-quality results with this type of approach, it is advisable to focus the research on a few previously defined variables. The procedure is distinguished by the short-term presence of the researchers in the field compensated by a more intensive data collection, which is usually based mainly on direct observation in anthropology and sociology and on interviews in health research. The lack of a strong individual immersion in the group or community studied is replaced by a high volume of data which are analyzed intensively and thoroughly. Moreover, in the FEM, the field observation has a certain degree of structuring and targeting, and digital recording is often used as an alternative or complement to written notes. Finally, individual work is replaced by collective work in both the collection and analysis of the data.

Some of the characteristics shown in Table 1, such as the focus on a few previously defined variables and targeted data collection, show the main points of contact of the FEM with CSR. However, this connection is not always made explicit in the methodological articles on FEM. One of the reasons for not mentioning this connection may be because theory is only used in some FEM studies to build and guide methodological decisions.

Although the FEM and CSR are frequently kept separate in current ethnographic research, we contend that a stronger blending of the two approaches is possible. This blending would introduce in the FEM a research rationale in which the selection and comparison of cases or settings would be driven by existing theoretical knowledge, offering a series of benefits that are shown in the following section.

A Comparative Ethnography of Two Hospital Auxiliary Services

To illustrate the type of integration between the FEM and CSR that we advocate and the results obtained from it, below we briefly describe the main elements of an investigation carried out in a public hospital located in Barcelona, Spain. The general objective of the study was to assess how the logic of competency-based management (Lichtenberger 2003) influenced the performance of tasks in the hospital’s auxiliary work processes. ⁴ Competency-based management implies that workers’ salary is linked to the intensity of their individual commitment and their cognitive abilities to understand, anticipate, and solve the problems of their function. The work organization corresponding to this type of management logic assumes that workers apply their abilities, skills, knowledge, and attitudes according to the changing and contingent needs of the work process without a protocol set in advance (Lichtenberger 2003).

The theoretical knowledge on competency-based management led us to consider two important intervening factors for understanding the question studied: the complexity of the work processes—reflected in the number of agents, interactions, and levels involved in the tasks and functions and in the simultaneity and fractionation of the work processes (Hutchins 1995)—and the type of skills applied in these tasks and functions.

The fieldwork was inspired by the objectives and procedures of the FEM and specifically by the previously mentioned workplace studies (Luff et al. 2000). Following the methodological principles of the FEM, the selection criteria for choosing the setting(s) to be observed considered the key theoretical elements that the study was focusing on. The first, and most obvious, was that we needed to observe work processes in which the competency-based work organization was applied; this was clearly the case of the auxiliary work processes in the chosen hospital (all performed by women) but not of the work performed by doctors; precisely because of this, the former were chosen. Second, we wanted to observe the performance of workers in two different but comparable work processes and assess the influence of the complexity of the work process and the type of skills used by workers, both of which were considered intervening factors in our theoretical framework. This objective led us to choose the work performed by both the administrative clerks of Outpatients and the nurses in the Sterilization Unit, which were both auxiliary work processes located within the hospital building.

The work carried out by the administrative clerks consisted of attending to the users (receiving them, taking note of their requests, and giving them appointments for medical tests and with the doctors), managing the patients’ medical records, and organizing and recording the day’s appointments. The administrative clerks carried out their tasks in front of the public, behind a desk at which patients were making queries or asking for appointments, so their work actually alternated between attention to the public and purely administrative work when there were no patients in the queue. This meant that the workflow was set by the public waiting to be attended. In the Sterilization Unit, instruments and surgical materials were sterilized and redistributed to the units that had sent them, so the workflow was set by the work in the operating rooms. The number of agents and interactions involved in the administrative work in Outpatients made the work process more complex (according to the definition we were using) than that in the Sterilization Unit, and the interaction with patients demanded interpersonal skills that were not required in the Sterilization Unit (see the Online Appendix for more details on the work processes and the characteristics of the workers).

The two work processes were chosen following the idea of multiple case design as proposed by Yin (2014) in which the differences between the contrasting cases are taken to test the influence of the key causal and intervening factors considered in the initial theoretical framework. The fact that the two work processes were in the same hospital meant that the higher management levels were fixed, thus avoiding the introduction of additional variability that could affect the work processes analyzed. The choice of a different health institute or a second hospital within the same institute would have included new variables that would have had to be related to the microlevels that were observed and analyzed. Table 2 shows the levels considered in the fieldwork, including those that were left “fixed” and those that were compared. Also included are the methods used to empirically address each level considered.

OPEN IN VIEWER

Table 2.

Levels Chosen for Carrying Out the Fieldwork and Methods Used to Obtain Data.

	Contextual Levels of the Work Processes (Fixed)			Auxiliary Work Processes within the Hospital that were Compared
Levels and settings	Catalan public health sector	Municipal health institute	Hospital	Sterilization of medical instruments	Individual tasks in the work process
				Administrative work in Outpatients	Individual tasks in the work process
Data collection methods	Documents	Documents	Documents and semistructured interviews	Documents, semistructured interviews, and open observation	Semistructured observation and recording with video camera and narrative interviews

Source: Drawn up by the authors.

As shown in Table 2, we based the data collection on various methods, giving a central role to the observation of the two work processes. Initially, we addressed the levels considered as fixed using internal and public documents and semistructured interviews with people with different levels of responsibility in the hospital. The main objective was to obtain an overview of the whole hospital and contextualize the tasks and work processes in the units studied within the competency-based organizational model. Subsequently, we formed two subteams of three analysts, each subteam taking data from one of the two units but working in coordination. In both units, the initial observation sessions were open and exploratory so that the observed workers could get used to the presence of observers and the observers could become familiar with the spaces and processes that they would later observe more thoroughly and systematically.

These initial observations were performed on different days during a week and different moments of the day (Verd et al. 2007) to identify possible variations in the workflow and the existence of small crises or special days or moments that broke with the everyday routine. Once the standard pace and workflow were identified, the analysts performed a more focused and semistructured observation of the tasks performed in the workplace. They first obtained data with field notes and then, when they had gained the trust of the people observed, with a video camera. According to the needs, the cameras were kept in a fixed position or hand-held by the observers. In addition, narrative–professional interviews were conducted with the observed workers.

Finally, all the information was analyzed following a qualitative content analysis procedure performed with the ATLAS.ti version 5 software that proved to be especially suitable for storing and handling all the types of material obtained (documents, interviews, field notes, and audiovisual recordings) together and for team analysis. The data of contextual levels were analyzed by the six members of the research team, while the data from each unit were jointly interpreted by the members of each subteam. The sharing of the same theoretical perspective by all team members based on the readings of common texts ensured the general coherency of the analysis. The slight differences in the interpretation of some details of data made by different researchers—probably associated with the gender and experience of the researchers (May and Pattillo-McCoy 2000)—allowed a richer and more complete interpretation of the data as a whole, as all interpretive discrepancies were discussed among the researchers to reach a common understanding.

Discussion and Evaluation of the Methodological Strategy

The methodological strategy described above offered several beneficial results for the study. First, closely following the research logic of the FEM provided the efficiency that is attributed to this type of ethnography when focused and thematically delimited subjects are studied (Knoblauch 2005; G. H. Pelto et al. 2013; Wall 2015), a clear research question has been posed, and focus variables are identified (Handwerker 2001). In our case, the theoretical knowledge we had on competency-based management led us to focus on the organization of work processes and the performance of workers. With a set of short visits to the units studied, the team of six researchers obtained a large volume of high-density data on the way tasks were performed and how the complexity and type of skills applied influenced the performance of workers. This was possible, thanks to intensive data collection through the combined use of note-taking and audiovisual recordings. Furthermore, the data analysis allowed the findings to be mutually validated by the researchers, while contributing to a more detailed portrait (May and Pattillo-McCoy 2000) and a more holistic understanding of the research topic (Bikker et al. 2017), which was facilitated by the use of the ATLAS.ti program. This use of the FEM was possible because sufficient theoretical information was available to researchers to frame the research question and to identify the relevant key factors for addressing the phenomena under study. In other ethnographic studies, this theoretical information is not always available, in which case a more open approach closer to conventional ethnography is necessary.

By incorporating in the FEM the CSR logic of selecting the cases using theoretical criteria in the design phase prior to the data collection, we obtained three added benefits. First, it guaranteed that the two settings had the appropriate characteristics for the data to correspond to the study problem. In fact, the processes of both the administrative work of Outpatients and the Sterilization Unit provided relevant data on the workers’ performance, the effects of work process complexity on this performance, and the type of skills used by the workers. It seemed that as both workflows were continuous and not fully predictable, the management found that it was more efficient to leave the specific work organization to the workers’ initiative without setting a work protocol in advance. Moreover, when work processes were complex (as in Outpatients), the workers were left alone to decide how to solve unexpected problems or conflictual situations and the kind of skills used to overcome these problems or conflicts. These skills were identified during the observation sessions but were invisible to the workers themselves because they were not acquired by means of formal training but informally through gender socialization (Jenson 1989). In short, whereas the FEM made it possible to obtain thematically well-focused data, CSR’s theory-driven selection of cases made it possible to directly obtain appropriate information on the focus variables of the study such as the pace of work, the degree of complexity of the work process, and the type of skills used by the workers.

Second, the combined approach allowed us to compare the two cases, controlling the influence of the variables that might introduce variability in the results (Yin 2012:33–38) and taking full advantage of the replication of cases. The fact that both units had the same institutional and management framework made it easier to see the specific effects of the technical and social organization of each work process over and above the common effects of a competency-based work organization (Lichtenberger 2003) on both units. For instance, in Outpatients, the pace of work was marked by a tight flow (Durand 2001) in the queues of users and medical requests, also involving frequent interruptions, simultaneous tasks, and cognitive overload as a consequence (Cicourel 2002), while in the Sterilization Unit, the pace was more regular and sequential with very few interruptions from outside the unit.

Third, the theoretically guided selection of cases guaranteed that the work processes selected were not unique and idiosyncratic and that the observed work organization was shared by other work processes in which a similar competency-based rationale was applied. The representativeness of the setting was thus ensured in advance.

The final outcome of the abovementioned characteristics was the possibility of achieving a good analytic generalization of the results obtained. For instance, one of our findings was that when competency-based work organization is coupled with a continuous workflow that workers cannot control, this produces a sort of “imposed self-organization” in which the decision-making power of workers is minimal. ⁵ In the context of this work organization, the tighter the flow and the more complex the work process, the higher the mental effort and cognitive overload of workers. This finding can be generalized to many other work processes in which a seemingly high margin of decision-making is granted to the workers, while at the same time, they are subjected to a pace of work that is beyond their control. This analytic generalization fulfills the objective of obtaining the transferability (Lincoln and Guba 1985) of the results and conclusions to other contexts and settings.

Conclusion

This article has presented and reflected on the advantages of blending the methodological characteristics of the FEM and CSR in a study on the influence of competency-based management in the organization of work processes. Specifically, it has shown the benefits regarding the quality of data, the ability to make comparisons, and the transferability of findings when CSR’s logic of using theoretical criteria in the design phase for selecting and comparing cases are incorporated in an FEM approach. We conclude that this combination offers a series of advantages for studies that start from a clear and focused research question, in which researchers have sufficient theoretical information available and key factors are well identified in advance. The benefits are summarized below.

First, the theory-driven selection guarantees that the cases chosen are not unique and idiosyncratic, thus ensuring their representativeness regarding the phenomena under study. Second, it ensures the theoretical relevance of the collected data. Third, it provides robust conclusions, thanks to a comparison of cases (which Yin [2014] calls replication) that allows the influence of explanatory factors to be controlled. Finally, it improves the analytic generalization of the results and conclusions, thus fulfilling the objective of external generalization (Maxwell and Chmiel 2014) or transferability (Lincoln and Guba 1985) of the research findings.

These better research outcomes are achieved because the characteristics of the CSR design reinforce many of the strengths of the FEM, such as its thematic concreteness and its targeted data collection. The theoretical framework of the study thus strengthens the theoretical relevance of data collection and makes it easy to focus on the key variables previously highlighted in the literature in addition to offering guidance for multiple case comparison.

However, the blending of the FEM with CSR does not transform the central characteristics of fieldwork in the FEM. The FEM’s short-term fieldwork reduces the chances of collecting information that has not been previously identified as relevant or that arises as a consequence of “exceptional” or unexpected situations. The chances of collecting this kind of information would be higher with longer periods of observation, so this limitation associated with FEM (even if the fieldwork is carried out at different times of the day or week as in our study) remains when it is combined with CSR. Another characteristic of the FEM that remains when the CSR logic is incorporated in it is the collective nature of the data collection and interpretation, which could imply differences in the understanding of data made by each researcher. However, in this case, the sharing of the same theoretical perspective based on the reading of common texts and the joint interpretation of data by all team members may help ensure the general coherency of the analysis.

Furthermore, the theoretically guided selection of cases and the focus on some key variables that are characteristics of CSR may cause the researchers to miss information that could subsequently be relevant to the particular phenomenon studied. Thus, the incorporation of the CSR logic into the FEM brings to the combination a limitation that could originally be mostly placed in CSR. However, although this may limit the identification of new relevant variables not considered in the literature, it produces in exchange a better and more thorough understanding of the relation between the variables initially considered.

As a final comment, the proposal presented in the article is not free of drawbacks, but according to our experience, the blending of the FEM with CSR brings together complementary strengths without increasing their limitations.

Supplemental Material