Chapter 3: Research Methodology

Introduction

Learning about the meanings of others‘ experiences demands an eclectic and holistic paradigm of methodology ([Papadopoulou & Birch, 2009](/dissertation/references)). This chapter will present the research questions, and how the chosen methodology of phenomenology pertains to these questions. Each aspect of the phenomenological process will be discussed.

Research Questions

The purpose of this study was to determine the essence of resiliency to compassion fatigue, using paradigms from empathy, emotional contagion, resiliency, and compassion fatigue theory. In refining the research questions, I spoke individually with a panel of experts who informed the direction of the research. The meanings these five professionals ascribed to the phenomenon of resiliency to CF helped define positive traits and behaviors possibly leading to this resiliency, as well as methodology for the study. The primary research question was: What is the collective essence of healthcare professionals‘ experience of resiliency to compassion fatigue? The secondary questions were: Research Question 2: What distinctions and similarities do healthcare professionals experience between burnout, CF, and vicarious trauma? Research Question 3: What qualities and characteristics do health care professionals experience as supportive to being empathic and yet resilient to CF? A phenomenological method was used for this study. According to Moustakas (1994), one of the characteristics of a research question suitable for phenomenology is

that the answers will reveal deeper essences and meanings of human experience and behavior. Phenomenology should not seek causal relationships, models, or predictions. Instead of measurements or scores, the results are interpretative, accompanied with comprehensive and vivid descriptions, based on the participants‘ interview responses. Another aspect of phenomenology is that ―it engages the total self of the research participant, and sustains personal and passionate involvement‖ ([Moustakas, 1994, p. 105](/dissertation/references)). Semistructured interviews, using open-ended questions following an interview worksheet (see Appendix E), were used, followed by memberchecking of the interview transcripts to provide consistency and validity to the study ([Creswell, 2009; Moustakas, 1994](/dissertation/references)).

Choice of Methodology

The study used a paradigm based on the interpretative phenomenology of van Manen (1990) and descriptive phenomenology of Moustakas (1994). The phases of the study were interwoven, rather than completed in discrete steps. Keeping a field journal was an important part of maintaining cohesion and coherence. In this journal, I documented the epoché, a self-assessment regarding my own perceptions and meanings of empathy, CF, and resiliency to CF, as well as the experience of conducting the interviews and analyses. This self-analysis process continued throughout the project, beginning before recruitment, continuing through data gathering and analysis, into the writing and rewriting of the study, and dissemination of its results. The importance of

language, emphasized by van Manen (1990), influenced both the field journal documentation, and the information analysis. Following Moustakas‘ (1994) guidelines, the phases were: (a) methods of preparation; (b) methods of collecting data; (c) methods of organizing and analyzing data; and (d) methods for reporting data. The fluidity of the entire phenomenological process required these phases to interweave and iterate. As new themes emerged, new questions arose, but the overarching research question continued to provide focus. Many of these new questions are included in [chapter 5](/dissertation/chapter-5).

Role of the Researcher

In phenomenology the researcher takes a subjective role within the study ([Creswell, 2007; Moustakas, 1994; van Manen, 1990](/dissertation/references)). I bracketed my preconceptions, experiences, and theories, using epoché, following Husserl (1931) and Moustakas (1994). Doing this enhanced my ability to be fully present with each person. Although epoché is seldom achieved perfectly, the attempt to achieve it brought an energy, intention, and alertness to the process of data gathering ([Moustakas, 1994](/dissertation/references)). My views preceding, during and after the interviews were written in the field journal; the pertinent ones are summarized in [chapter 5](/dissertation/chapter-5). Although I met one of the participants in a professional setting years before the study, there were no ongoing relationships between us. No relationships developed during the study, therefore no participant was discontinued. I conducted interviews using a guided set of open-ended questions and prompts. With the participants‘ permission, I captured the interviews with digital recording, as well as handwritten notes. I reviewed the transcriptions from the digital recordings, and

verified them with the digital recordings, as well as memberchecking. Interaction with the language is an important part of the interpretative phenomenological process, according to van Manen (1990). In the interviews I used a practice Finlay (2005) termed ―reflexive embodied empathy‖ with an ―embodied intersubjective relationship‖ ([Finlay, 2005, p. 271](/dissertation/references)). This type of empathic communication involves different layers of intersubjectivity possible between researcher and participants. The first layer of this intersubjective relationship is a physical empathic reaction with each other. A second layer focuses on the cognitive aspect of empathy, where each person imagines the other‘s being, mirroring and doubling what is imagined. The deepest layer is one of full empathic engagement, combining all cognitive and affective aspects of empathic understanding ([Finlay, 2009](/dissertation/references)). The reflexive embodied empathic process was begun with a few moments of quiet centering at the beginning of each interview, which all of the participants appreciated. Once the interviews were transcribed, I horizonalized ―textural meanings and invariant constituents of the phenomenon‖ ([Moustakas, 1994, p. 97](/dissertation/references)). In horizonalization, the second step of the phenomenological data analysis process, every significant statement, or horizon, relevant to the topic is listed. Each statement, or horizon, of the experience, is given equal value ([Moustakas, 1994](/dissertation/references)). Using qualitative organizational software (NVivo 10) to assist the process, I determined themes and their relationships by clustering the horizonalized data. This led to both a structural and textural description of resiliency to CF, as experienced by the participants. Details of this process are presented later in this chapter, and in [chapter 4](/dissertation/chapter-4).

I have a background as an organizational consultant for over 15 years (see Appendix J, Curriculum Vita). During this experience I developed my interviewing skills, as well as analytical and synthesizing abilities. My study of empathy and CF over a number of years, as well as learning theories of why and how CF occurs, could have brought a bias to the interviews and analyses. I tried to be alert to this possibility, and eliminate it whenever needed. The study followed a positive psychology paradigm, examining strengths rather than risks, with the aim of uncovering traits and behaviors of resiliency, rather than testing theories on risk factors of CF.

Context for the Study

Data was gathered primarily at a teaching hospital in northwestern United States. Four other local hospitals were also represented, two participants working at both the primary hospital and another, and two working solely at other hospitals. Although CF occurs in many settings outside of organized health care, this study focused only on CF resiliency in the large medical facilities mentioned. Before recruiting participants I presented the proposed study to the director of nursing research at the primary hospital. She indicated agreement in a letter of cooperation (see Appendix D for a draft of this document), after determining that approval from my university was sufficient. Although the context of a phenomenological study may change during the course of the study ([Moustakas, 1994](/dissertation/references)), the context of this study remained constant. The context from a pure, uninvestigated situation was altered due to my subjective presence within the study of the researcher as primary data collector, analyst, and synthesizer ([K. Irwin, 2006](/dissertation/references)). Due to the perception some participants may have had of me as interviewer- 87 expert, a certain power disparity ([K. Irwin, 2006; Thapar-Björkert & Henry, 2004](/dissertation/references)) may have occurred. I attempted to reduce this disparity and increase autonomy of the participants by allowing them to select their interview locations ([Irwin, 2006](/dissertation/references)), and determine the duration of interviews.

Participant Selection and Recruitment

The focus of qualitative research is the description and interpretation of human experience, which requires depth of intense, saturated depictions ([Polkinghome, 2005](/dissertation/references)). Therefore, the participants for this study were selected, not for representative requirements of statistical inference, but for their value in providing clear, in-depth description and interpretation. Although the proposed study expected to use nine to twelve participants, data saturation was achieved by the end of the seventh interview, so no additional participants were recruited. Access to participants. This study used a chain referral process for purposive recruitment. After the representative from the primary hospital signed the letter of cooperation (see Appendix D), she referred me to several key people. My intention was to include participants from departments known to have high risk for CF: palliative care, particularly for children; cancer care; long-term rehabilitation; physical therapy; pain clinics; and other areas where medical staff have longer interaction with a patient and the patient‘s family ([Abendroth & Flannery, 2006](/dissertation/references)). The original intention of the study was to meet potential candidates, as well as participants in their choices of venue. Unfortunately, at the beginning of data collection I suffered a concussion and spinal injury, requiring home rest and immobility for two

months. Several candidates chose to come to my home for their interviews, and one chose to be interviewed by phone. Recruitment procedures. I met with candidates in three different ways, in person, by email, and by phone. Most of these people had learned of the study from key people in their department, who in turn, had learned of the study from the head of nursing research. In our introductory meetings, I briefly explained the project, and what was expected of participants. I determined the candidate‘s level of interest, as well as whether or not the candidate met the criteria for being a participant. In all cases, the candidates met the study‘s qualifications to be participants. I then presented each candidate an informed consent letter (see Appendix B), and explained it. In some cases the participants received and signed this consent form, using email. In a few cases, this happened at the beginning of the interview, although a digital copy had been sent to the candidate prior to the interview. All participants signed the consent form electronically or by hand. The interviews were scheduled for a location of his or her choice, at a time convenient for each participant. Between the initial contact and the interview I answered any questions concerning the project. I asked for referrals to other people who might be interested in participating, as part of the chain sampling method. The first key person to whom I was referred by the head of nursing research was a co-coordinator of procedures in oncology nursing. After a meeting in which I described the project, she invited me to a monthly meeting of representative nurses from various

areas in oncology. I presented the project and an invitation to participate. Other than this, I did no additional recruitment. Criteria used for selection. Purposeful selection of participants was to include nine to twelve people who could be key informants. Moustakas (1994) suggested essential criteria for participants should be they: (a) have experienced the phenomenon; (b) have high interest in it; and (c) are willing to commit some time to the study process. In this study, all of the participants will have experienced CF, either having the disorder or by witnessing a coworker develop CF. The participants should have a general knowledge of the characteristics and behaviors of an empathic person, as well as what CF is. Initially the study planned on excluding anyone currently experiencing CF. One candidate asked to be included, even though she admitted to be in burnout. As this was not CF, I included her as a participant, in order to learn the differences between burnout and CF, if any existed. No minors or other vulnerable populations were approached to become participants. The study was open to any age, sexual orientation, gender, ethnicity, nationality, or socioeconomic status. Later studies may research whether any of these factors affect the occurrence and experience of CF, but none of those factors were part of the study. Although age was not be a discriminating factor, the participants were to have at least five years experience in their current specialty area. This criterion was met by all participants, as seen in Table 4 in [chapter 4](/dissertation/chapter-4). In order to provide contrast, depth, and triangulation to the study, most participants were from different health disciplines.

During the first screening interview, I discussed briefly the areas of CF and empathy, in order to assess the value of the candidates‘ interpretation and knowledge. The meanings for CF, as well as for empathy, differed from individual to individual, adding range and breadth to the data gathered. It was not necessary or desirable for the study that a consensus of definition be reached. The most important selection criteria was the amount of information each person could provide describing the phenomenon of empathic people‘s resiliency to CF, and their interpretation of it. The number of participants depended on data saturation, as indicated above.

Establishing Working Relationship with Participants

In qualitative research it is very important to establish good working relationships with participants ([Creswell, 2007; Moustakas, 1994; van Manen, 1990](/dissertation/references)). I was the primary data collector and interpreter. Without a good relationship with the participants, the depth of their experiences and what it means to them cannot be attained. In the hermeneutic interview within van Manen‘s methodology, the interview process tends to turn participants into collaborators on the project ([van Manen, 1990, p. 63](/dissertation/references)). Without this type of relationship it may be difficult to hear a ―lived-experience description‖ and interpretation from the participant ([van Manen, 1990, p. 64](/dissertation/references)). To achieve this type of description and interpretation, I scheduled the interviews at times determined by the participants, and at locations they choose.As the depth of information an individual may provide can be negatively impacted by the presence of the other person, or the account colored by the other person‘s reactions, the interviews were for one person at a time.

I planned to use whatever method of communication worked best for each participant during the interview, whether simply talking about feelings, experiences, observations, or supplementing this with graphic displays, charts, data sheets, or any other information source. None of the participants needed anything more than being listened to. I encouraged the participants to avoid causal explanations or generalizations concerning resilience to CF. The participants were encouraged, instead, to describe particular incidents, and provide specific examples. The questions in the interviews followed the train of thought each participant presented, allowing the participant to reflect on what was said, and make amendments and additions. Memberchecking was used to refine their descriptions and interpretations. As the interviews progressed, any misconceptions which appeared were resolved. I accepted and supported open disclosures, as recommended by Moustakas (1994, p. 110). At the completion of the data gathering and memberchecking, I sent each participant a letter of gratitude for his or her contribution to knowledge concerning the nature, meaning, and essence of the phenomenon, empathic people‘s resiliency to CF. Compensation. No compensation in the form of monetary payment or gift was given to the participants. Instead, I offered each participant a copy of the summary report, and if they were interested, a short presentation and discussion session of the results to their departments.

Ethical Protection of Participants

The ethical issues of qualitative research overlap with those of quantitative research, but are more based in the relationship between participants and researcher.

Qualitative researchers need to be more subjective than in quantitative studies, which can create boundary issues. I followed Walden University‘s Institutional Review Board (IRB) guidelines in ethical protection. As part of this,I strove to be honest and open in relationship with the participants, to minimize any blurring between roles as professional researcher, professional acquaintance, and friend. The study presented minimal risk to participants in terms of unintended disclosure of information, physical or psychological harm. No vulnerable populations were involved. During the research study the identities and names of participants were masked. The process of phenomenological research may extend over time, creating a climate of exchange of intimate details between researcher and community. This exposure may lead to harmful repercussions ([LeCompte & Schensul, 1999a](/dissertation/references)). All possible ways to minimize external disclosure of these intimacies was taken. A coded identification was given each participant. The cross-list was available only to me for use in crosschecking information with the individual participants, and was destroyed at the end of the project. I was the only person gathering, analyzing, and summarizing the data. The professional transcriber received digital recordings with all identifying data removed. In addition, she signed a confidentiality and nondisclosure form. I verified the transcripts against the digital recordings of the interviews. The participants each verified their individual transcripts. Although the final report is available to the participants and their respective departments, any identification of who said what has been removed. It is possible during the interviews, despite precautions to minimize harm, the recollection of upsetting experiences may be uncomfortable to some participants. This

happened with one. I suggested psychological counseling to the participant, who affirmed this was already being used. I did not attempt to provide counseling. As part of the informed consent letter, each participant was given access information for both me and my dissertation committee chair in case of any difficulties or questions. Informed consent. In the informed consent letter (see Appendix B), details of the project such as the general purpose and benefit, duration of the study, as well as the extent and type of participation, was presented. The reasons why people were being approached to participate was given, and the possible minimal risks. The informed consent letter also disclosed how information was kept secure and confidential. The informed consent letter told potential participants that data would be securely stored in password protected files, on a peripheral hard drive kept in a lock box. Also, none of the individual‘s confidential information would be used for any purpose outside of the proposed project. At any time during the study, participants could withdraw with no consequence to them, but none did. I explained the informed consent letter when I gave it to each participant. Protective measures. Even when there appears to be minimum risk to the participants, I needed to be alert to the potential dangers, and protect the participants from harm, intrusion, or loss of privacy whenever possible ([Fisher, 2012](/dissertation/references)). Autonomy of participants was preserved, enhancing the validity of the study by increasing the potential for depth of description. Interviews were scheduled in private areas selected by the participants at times and durations convenient to them. Each participant was assigned a pseudonym. This link to identifying information was kept in a password-protected digital

file, available only to me, until the project was completed. This was done in case a participant wanted to remove information after deciding not to continue participating. It was also critical that no aspect of the study disempowered the participants. One of the paradigms on which the study was based on strengths development in positive psychology, so there was no risk of disempowerment. APA Standard 3.06 (Conflict of Interest) provides a basic guideline against embarking on a research study when any other roles, interests, or relationships could lead to clouding researchers‘ judgment, impeding their effectiveness, or exposing the participants or site to any exploitation or harm ([Fisher, 2012](/dissertation/references)). For the study, the issue of boundary issues was brought up in the initial screening interview. If multiple roles appeared to be a problem between the candidate and me, I did not ask the person to participate. Institutional review. Before making application to the Walden University IRB, I requested and received a letter of cooperation from the primary hospital‘s nursing research office; a draft of this letter of cooperation appears in Appendix D. This hospital found my Walden University IRB approval, but limited my study to nine or less nurses, rather than a variety of healthcare positions. The two participants who represented two other hospitals volunteered when they learned of the study, and did not represent their hospitals, so no further institutional review and approval was sought.

Data Collection

The typical data collection method for phenomenology uses semistructured interviews, involving participants as coresearchers. Although I brought an interview worksheet with open-ended questions and prompts (see Appendix E), the interviews

followed the participant‘s story without direct reference to those specific questions. The interview guideline questions were used to return focus to the interview.

Data Collection Procedures

The screening process for candidates based on participation criteria was conducted either by email or by phone, and was not recorded. During this process, I provided a brief overview of the study, and inquired about the candidate‘s interest in the study, in which department they worked, and for how long. In all cases, candidates became participants. In the data collection interview, after a brief social conversation and restatement of the purpose of the interview and the study, we spent a moment of centered reflection. The interviews, which proceeded for an average of 47.5 minutes, were digitally recorded, with permission from the participants. I also took handwritten notes regarding the context and process of each interview, trying to be alert to bias and avoid it. Additional information was gained during the memberchecking process. The phenomenological researcher, as primary data gatherer, needs to have self- efficacy in being with others, their emotions, and their issues ([Ivey, Ivey, & Zalaquett, 2010](/dissertation/references)). To be skilled in the data collection process of interviews, the researcher needs to follow the trend of the participant‘s thought, with little interruption, using additional probes in order to enhance the richness of the data ([Ivey et al., 2010](/dissertation/references)). I followed these guidelines in conducting the interviews.

Recording and Transcription

As noted above, the interviews were audio-recorded, with the participants‘ permission, as well as hand-noted. The audio-recordings and transcriptions were stored on a password-protected peripheral hard drive, and kept in a locked box, separate from the computer. Before transcription, I removed any identifying information. The audio- recordings were digitally keyboarded by a professional transcriptionist, who signed an agreement of confidentiality and nondisclosure. I had begun transcribing the first interview, done over the phone, but found the digital recording was not clear, and the transcription was taking too long. I asked a trustworthy acquaintance of mine, and she took on the project. Before giving me the transcriptions, she double-checked them against the recordings. I also checked the transcriptions against the recordings, and removed additional phrases and words which could be used as identification. These edited transcriptions were then memberchecked. Three of the participants told me no changes were needed; the other four, particularly the participant who had been interviewed by phone, provided corrections and additions. These modified transcripts were entered into NVivo10 as cases. The raw data and edited transcripts will be kept for five years after the completion of the study. In support of the transcribed interviews, my field journal provided additional insight into the content and process. The reflexive nature of this journal is part of the transcendental phenomenological process ([Moustakas, 1994](/dissertation/references)). Analysis of the specific language used also appeared in the field journal, as part of the interpretative process ([van Manen, 1990](/dissertation/references)).

Verification Process

As soon as the first participant was interviewed, I began transcription, data organization, and verification. This iterative process continued until after saturation of data had been achieved and verified. Each edited transcribed interview was delivered to the pertinent participant for verification, amendment, and addition. Once the reviewed data had been verified individually, NVivo10 was used to organize the analysis process, as discussed in the next section. An executive report, a summary of the findings, was prepared, and made available to each participant for further verification. Data Analysis

Data Analysis Techniques

The analytic process was similar to ethnographic research in that the coding and study of interrelationships began when the first interview has been conducted, and was refined as the data was collected ([LeCompte & Schensul, 1999b](/dissertation/references)). Following van Kaam (1959, 1966), Moustakas (1994), and van Manen (1990), organization of the data began at the time of interview transcription. As I listened to the audio-recordings and coded the transcripts into NVivo10, I notated in my field journal trends and observations of the interview, as well as my own assumptions, biases, experiences, and theories regarding the phenomenon of resiliency to CF. This process, termed epoché, allowed me to bracket the phenomenon as experienced by the participants, separate from my own views. This approach was presented by Moustakas (1994) as a modification of models developed by Stevick (in 1971), Colaizzi (in 1973), and Keen (in 1975). This self-analytical process followed the same steps as the analysis of the transcripts from the participants. As part of

this process, I had a friend interview me, using questions from the interview guidelines. A summary of this epoché can be found in Appendix H. In the first stage of data analysis, horizonalization, I listed every relevant expression relevant to the phenomenon ([Moustakas, 1994; van Kaam, 1959, 1966](/dissertation/references)). These horizons were entered into NVivo10 as free nodes, such as empathy, burnout, support for compassion. The concept of horizonalizing is characteristic of phenomenological analysis. An expression qualifies as a horizon if it contains ―a moment of the experience that is a necessary and sufficient constituent for understanding it‖ ([Moustakas, 1994, p. 121](/dissertation/references)). Rather than using the reduction and elimination process recommended by Moustakas, I coded every significant statement, or relevant horizon, giving each equal value. A cumulative value of overlapping or repetitive statements developed, as suggested by Giorgi (2006), under each node. From this preliminary analytic process, meaning units were selected ([Moustakas, 1994](/dissertation/references)). This iterative process continued as more information was gathered from each interview ([Creswell, 2007; Schensul, Schensul, & LeCompte, 1999; van Manen, 1990](/dissertation/references)). I clustered these meaning units into common concepts or themes: these invariant horizons, or nonrepetitive, nonoverlapping statements became parent nodes. Using memos and annotations within NVivo, these parent nodes were accompanied with a synthesis of descriptive terms and phrases. As the analysis proceeded, child nodes were identified and brought into the hierarchy, headed by the parent nodes. As part of this development of free nodes into hierarchical or parent nodes, interrelationships between the nodes were

identified, using NVivo10. A hierarchical structure also occurred with relation to the interview questions and subquestions, as displayed in Table 5 in [chapter 4](/dissertation/chapter-4). Using ―imaginative variation‖ I took various points of view to examine the textural themes and qualities in order to develop a structural analysis ([Moustakas, 1994, p. 122](/dissertation/references)). By varying frames of reference, reversing positions, changing roles or functions, I investigated ―underlying and precipitating [components of the experience of each participant]. . . . The thrust is away from facts and measurable entities and toward meanings and essences; in this instant, intuition is not empirical but purely imaginative in character‖ ([Moustakas, 1994, pp. 97-98](/dissertation/references)). Consideration of invariant structures such as time, space, physical presence, relation of the phenomenon to the person, as well as interrelations with others in analyzing the participants‘ responses developed into a structural description of how resiliency to CF was experienced. To aid in this phase of analysis I created various diagrams and charts, included in [chapter 4](/dissertation/chapter-4), of the process by which emotions, thoughts, and behaviors react to the phenomenon of resiliency to CF. A textural description and structural description of the phenomenon was created for each participant, as well as for myself. I combined these into two composite descriptions representing the group, one textural, the other structural. The final executive report (Appendix XX) was a combination of these two descriptions into a holistic view of empathic people‘s resiliency to CF, accompanied with its interpretation. Data interpretation. The earliest interpretations in the study, beginning with the analysis of the first interview, were accompanied by intentional consciousness of my own views, as part of the epoché process. The interpretation was modified and refined as data

gathering and analysis continued, in the tradition of qualitative research. One of the primary advantages of qualitative research is the inclusion of participants‘ multiple perspectives in data interpretation ([Bennett, 1998a; LeCompte & Schensul, 1999b](/dissertation/references)). Once the data were gathered into composite reports, I used various theories of CF, empathy, and ecological theories of social network support in order to extract valid meanings. This interpretation style, based on the phenomenological methods of Moustakas (1994) and van Manen (1997), provided: (a) a description of my personal experiences with CF and people who do not develop it; (b) a list of horizons, significant statements from the participants; (c) a grouping of elements into meaning units or themes which emerge; (d) a textural description of the what the participants experienced or observed; (e) a structural description of how their experiences occurred; and ends with (f) a composite description and interpretation of the phenomenon. This composite description and interpretation included a set of propositions and a list of features or elements which were common to all points of interpretations, couched within an essay. As the sample was limited to seven nurses, a frequency table was not pertinent. Because graphics can convey a great deal of meaning ([LeCompte & Schensul, 1999a](/dissertation/references)), particularly in showing relationships between elements, I created prototype models which indicate these relationships. ***

Use of NVivo 10 Software

Use of qualitative data analysis (QDA) software aided me in a number of ways, such as providing a single location for storage of data, and use of consistent coding

schemes. I have considered the possible disadvantages and limitations, beyond the time needed to become proficient in using the program. Being aware that some researchers use the QDA software to take shortcuts, or may become distant from their data ([Bergin, 2011](/dissertation/references)), I will remain true to the phenomenological approach I would have used without the QDA. I can avoid analytical pitfalls by running queries of the database as I progress from early coding to theme clustering, rather waiting for a mechanical data analysis with artificial results ([Bergin, 2011](/dissertation/references)). In qualitative analysis the researcher, not the QDA software is the analyst. My choice of NVivo 10, a QDA software developed and supported by QSR International ([QSR International Pty Ltd., 2010](/dissertation/references)), is based on several factors: (a) NVivo 10 will read digital audio input from my interviews, and provide summary transcripts of each; (b) NVivo 10 will aid me in managing and organizing the data as I gather it, and will manage ideas that come to me in the data gathering process; (c) NVivo 10 will provide me many methods of reporting data, including graphs, models, and tables, based on varied querying modes I will use ([Bergin, 2011; Leech & Onwuegbuzie, 2011](/dissertation/references)). The summary transcripts of the interviews will be part of the analytic triangulation, compared with verbatim transcripts I will be able to create from the audio playback through NVivo 10. Additional strengths of NVivo 10 include allowing me to import chunks of my literature review and create links between previous studies and theory with what will emerge in my study. This QDA software will provide me with flexibility in choosing which types of analysis are appropriate for data and research questions of my study

([Leech & Onwuegbuzie, 2011](/dissertation/references)). The highest advantage NVivo 10 will offer goes beyond shortening the amount of time coding and analyzing, and providing highly consistent coding and analysis, avoiding the common pitfalls of noncomputer-assisted qualitative analysis ([Bergin, 2011; Robson, 2002](/dissertation/references)). The qualitative analysis pitfalls include data overload, misjudgment in coding and grouping, and uneven reliability ([Bergin, 2011](/dissertation/references)). Misjudgment may occur due to the stronger influence early impressions may have, leading to resistance of later revision. It may also occur due to instances which confirm previously held beliefs, one of the reasons epoché is a critical part of phenomenology. Although NVivo 10 will be a useful tool for organizing information, I will need to be alert to the possibility of placing too much confidence in the QDA software, particularly in my coding. Coding procedures. NVivo 10 uses the creation of various types of nodes in coding information. I will categorize these nodes as I read through the transcript summaries and verbatim transcripts. Free nodes, as yet unassociated with others, are the initial phase. Each free node may be a collection of references to a theme, location, person, or other item of interest ([Bergin, 2011; Leech & Onwuegbuzie, 2011; QSR, 2010](/dissertation/references)). For example, some keywords which I may use for nodes are CF, resiliency, empathy, and emotional contagion, or emergency room, palliative care, and cancer clinic. In the next phase, tree nodes which are groupings of nodes (no longer free and unassociated) will be organized into hierarchical structures, showing linkages from a parent node to more specific categories, or child nodes. I will make judgments of pertinence and linkage before moving a free node into a tree node folder ([Bergin, 2011](/dissertation/references)).

Case nodes have attributes, such as participant type, or location, or other distinguishing factors. Relationship nodes show connections between two or more case nodes or items in the study, such as between resiliency to CF and a healthcare provision area. From these various types of nodes, matrices will be created, using a matrix-coding query. Each cell of the matrix will be a node used to further explore and code the data. Analysis procedures. Analysis of the data may begin after the first interview, due to using NVivo 10 for organization of data. Various modes of query exist, which will allow me to probe the data as I gather it, identifying patterns which I might follow in gathering more data. Each query can be saved in the QDA, run again with the new data, with a tracked progression of results ([QSR, 2010](/dissertation/references)). In a 2005 paper presented at the annual meeting of the American Educational Research Association, Leech and Onwuegbuzie (2007) identified 21 qualitative analytical methods. Constant comparison analysis, which examines and compares any type of input of data, is one of the most frequently used qualitative analytical tools ([Leech & Onwuegbuzie, 2011](/dissertation/references)). Constant comparison analysis was developed as a coding system for grounded theory research by Glaser and Strauss in 1967, but can be used for any qualitative research method as a systematic search for meaning. Deductive coding, where codes are identified prior to analysis, can be assisted by NVivo 10, which will also search for these codes already in the data. The QAS will also search inductively for codes which emerge, and abductively, seeking iterative occurrences of codes ([Leech & Onwuegbuzie, 2007](/dissertation/references)). This process is what most qualitative analysts refer to as coding, chunking, grouping by similarity, and identifying themes.

A second analytical tool which NVivo 10 can use is keywords-in-context ([Leech & Onwuegbuzie, 2011](/dissertation/references)). This method compares words which appear and after key words I will have designated. It is useful in identifying implied connections, and will be part of my hermeneutical analysis. Supplementing this, I will also use word count analysis to help identify patterns, to set usage values where pertinent, and to assist in maintaining analytic integrity ([Leech & Onwuegbuzie, 2007](/dissertation/references)). Word count analysis may be used to establish usage patterns of value-setting words, such as ―many,‖ or ―often.‖ As word count analysis may decontextualize the word, theme, or phrase, it can be misleading. It is based on the assumption that frequent use of a word by an individual indicates importance to the person‘s view. Although I will use this NVivo organizational tool, I will view the results through the lens of my field journal notes regarding each interview, and question the results for significance. Classical content analysis, instead of reporting emergent themes, can be used to count the number of times each node occurs. I will establish the coded nodes deductively, at the beginning of the analysis. This method is useful if there are many codes ([Leech & Onwuegbuzie, 2007](/dissertation/references)). As part of the structural analysis I will use domain analysis, taxonomic analysis, and componential analysis. Domain analysis uncovers semantic relationships between nodes, such as antonymy (X is defined as negative of Y), comparison (X is defined in contrast or similarity to Y), and attributive (X is defined with regard to attributes of Y) ([Leech & Onwuegbuzie, 2007, p. 568](/dissertation/references)). The components of each domain in NVivo 10 are tree nodes. The reasons I will use domain analysis are to uncover relationships among

concepts, giving myself another way to triangulate my analysis, and to develop additional probes for later interviews. Taxonomic analysis will help me understand the usage of specific words. Most people have different definitions of CF, empathy, and resiliency. Taxonomic analysis will show me the relationships of tree nodes within a domain, reveal branches or subsets, and how these branches relate to the whole tree ([Leech & Onwuegbuzie, 2007](/dissertation/references)). Componential analysis is a more detailed examination of the relationships between words. Using matrices, with one dimension a ―contrast set,‖ and the other ―dimensions of contrast‖ I will be able to determine gaps in the information ([Leech & Onwuegbuzie, 2007, p. 571](/dissertation/references)). NVivo 10 will be a very useful tool for me, but in the analytical process I recognize I will be the analyst, not the software program. As a vital part of the triangulation process, I will return to the participants for verification of the textural and structural prototypes I have developed.

Verification of Authenticity and Trustworthiness

Credibility, or internal validity, can be damaged in a study if sampling bias or omission of point of view occurs. In quantitative research, reliability, validity, and replicability are the parameters of authenticity and trustworthiness, but using those terms in qualitative research makes little sense due to the research paradigms used ([Williams & Morrow, 2009](/dissertation/references)). To establish authenticity and trustworthiness in the proposed qualitative study, I shall be attentive to the integrity of data, maintain a balance between reflexivity and subjectivity, and communicate the findings clearly, as recommended by Williams and Morrow (2009). To avoid sampling bias, I plan to take time with the study, and

gather data until a saturation point occurs, with inclusion of diverse perspectives. Providing details of the procedures I use will allow others to investigate the phenomenon with other populations. To further establish validity, I will use direct quotes, showing how data support my interpretation. Part of the validity of the proposed study will be based on the trust established ([Creswell, 2007](/dissertation/references)). Triangulation in qualitative research reinforces trustworthiness. Gathering information from varied bases, rather than from a homogeneous group, will be one leg of the proposed study; memberchecking will be a second leg; the third leg of triangulation will occur in comparing structural and textural descriptions of the phenomenon ([Davis & Paulson, 2011](/dissertation/references)). The reflexive field journal, described earlier, will add to the trustworthiness of the proposed study (Moustakas, 1994; van Manen, 1990 The journal will contain personal notes, describing my own answers to the questions asked in the study, my own biases regarding CF, empathy, and the phenomenon of empathic people not developing CF, as well as notes about the setting of the interviews, the facilities themselves, reactions to the participants, personal interpretations, and accountings of prior experiences with CF, and the role empathy plays in it. Transferability, or external validity, of a study yields sufficient description for others to decide whether or not the study‘s findings are generalizable to other situations ([Creswell, 2007](/dissertation/references)). For the proposed study, the structural description of the phenomenon and detail of procedure should strengthen the possibility of replicability with other populations ([Davis, 1998; Moustakas, 1994](/dissertation/references)). The report of findings for the study will

describe the contexts in which the interviews occurred, as well as the reported experiences and observations. This will include the environmental and management factors, as well as the nature and relationship of the participants to the people they are describing. External reliability cannot occur when concepts or data gathering methods are inappropriate for the group being studied ([Schensul et al., 1999](/dissertation/references)). The use of semi- structured interviews with open-ended questions is considered appropriate for phenomenological study ([Creswell, 2007](/dissertation/references)), and is a method with which all participants will be familiar, due to the nature of their professions. If there is any indication that the experiences reported by the participants are idiosyncratic to that group, this will be carefully documented in the study‘s report ([Schensul et al., 1999](/dissertation/references)). I will strive to maintain the clarity in definition of concepts, questions, factors, and elements at all levels, needed for both internal and external validity. Gaining a full perspective of a phenomenon requires allocation of sufficient time; conditions may change, and attrition may occur. I will be wary of making premature conclusions, and continue to modify interpretations as new, conflicting data emerges ([Moustaka, 1994; Schensul et al., 1999](/dissertation/references)). The overall validity of the study will also be enhanced by checking the transcriptions for accuracy against the audio-recordings, and ascertaining I did not influence the responses and descriptions in the interviews. The confirmability of the study by people other than participants and myself is less easy to control without damaging confidentiality: the only audit trail available would be to compare the audio-recordings and transcriptions.

Supported by the literature ([S. Smith, 2009](/dissertation/references)), the phenomenon of resiliency to CF is a valid phenomenon to investigate. The reflexive nature of transcendental phenomenology requires that I be alert to all potential predefined or biased attitudes regarding the phenomenon, its components, the people being interviewed, and the facilities in which the study will occur, thus decreasing the likelihood of unreported bias.

Dissemination of Findings

Each participant will receive a copy of the executive summary report, as will each interested participating department. For any department desiring it, I will make a presentation of the findings from the study, and/or lead a workshop in resiliency development. At the conclusion of the project I intend to condense the study into an article, and offer it to a peer-reviewed journal in the healthcare professions.

Summary of Processes and Security Precautions

The steps to be taken in the proposed study, after approval by the appropriate IRBs are: 1. Recruitment: Before attending a monthly meeting of the Oncology Local Practice Council, to which I have been invited, I will talk with the coordinator for this meeting. At the monthly meeting I will make a brief presentation of the project, and request referrals to potential participants. Using personal inquiry and the snowball method, I will identify and recruit nine participants. This will be done initially by speaking with nursing administrators or coordinators in various health areas: palliative care, cancer care, long-term rehabilitation, physical therapy, pain clinics, and long-term

pediatric care. In this initial phase, I will request referrals to people who might serve as key informants for the project. The potential key informants will be approached initially using email, phone, and in-person requests, over a period of two to four weeks. A letter of introduction (see Appendix C) will be used. Each participant will be asked to sign a consent form (see Appendix B) 2. Data gathering: During the recruitment period, I will schedule interviews with the participants, selecting sites which are recommended by the participants, usually at their places of work. Each participant will be assigned a pseudonym. In order to be able to remove information should a participant decide not to continue, the link between participant and pseudonym will be kept securely until the project is completed. Each interview will be audio-recorded. The interviews will last 45 minutes, or longer if the participant wishes. The audio-recordings will be stored in a locked box, available only to me. 3. Transcription: I will make a verbatim transcription of the audio-recordings of interviews. I will keep the original verbatim transcript stored on a peripheral hard drive in a locked box, with the audio-recordings. A second transcription, removing any identifying information, will be stored on a separate password protected flash-drive. 4. Analysis: I will begin this phase of the study after the first interview, rather than after all initial interviews. First, I will describe in my field journal each

participant, as well as the context of each interview. Using NVivo 10, I will enter the reduced transcript, with identifying information replaced by code, into a database, which will then be processed to cluster elements of data into themes and overarching motifs, leading to a textural description for each participant. Using imaginative variation, I will create a structural description for each participant. 5. Composite description: Combining the individual descriptions, I will develop a group composite of both textural and structural findings, followed by a holistic narrative which describes the phenomenon holistically. 6. Verification: Using in-person interviews, I will bring the composite textural and structural findings back to the participants for additional feedback and correction. 7. Revision: I will revise the textural and structural narratives, based on new input from the participants. This will lead to revisions of the composite descriptions. 8. Dissemination: If requested by participants, I will make presentations of the findings to the interested departments. Each participant will receive a copy of the final report. Many articles have been written describing studies of CF ([S. Smith, 2009](/dissertation/references)). The majority of the studies have been done in medical settings, frequently by nurses, with the intended audiences being policy makers, administrators, and other nurses ([Aycock & Boyle, 2009](/dissertation/references)). For the proposed study, these will also be the primary audiences. As this

study may be preliminary to other studies, including a grounded theory study to develop a revised theory and model of CF in the health profession, I do not plan to disseminate the information widely until I have performed additional research. The next two chapters will be completed after data have been gathered and analyzed. [Chapter 4](/dissertation/chapter-4) will discuss the process by which data were generated, gathered, and recorded. I will describe how I kept track of data, as well as the use of my field journal. I will present the findings from the analysis, tying relevant discoveries to the individual research questions. In addition, if there are any discrepant cases or themes, I will disclose them, including the presentation of all salient data. In my discussion of evidence of quality, I will show how this study followed procedures which ensured data accuracy, so my study will have merit, trustworthiness, and credibility. [Chapter 5](/dissertation/chapter-5) provides a brief overview of why and how the proposed study was done, followed by an interpretation of the findings. I indicate the limitations of the study, where additional study might be done, as well as recommendations for action. I will present the implications for social change which the findings underscore. As part of this chapter I will also write about my experiences in the research process, my own biases, and any possible effects I may have had on the interviewing process and gathering of data.