Abstract
Accepted methods of analysis might be viewed as inadequate for analytical work on field notes and interview transcriptions when the aim is to analyze these as a complete body of material. The purpose of this article is to present a new method for systematic data integration in which the key component is the interweaving of observation data and interview data derived from sequences of interactive situations. The method has been developed in a study where groups of user representatives and professionals collaborated for some time on planning, implementing, and evaluating patient education programmes. The article briefly presents the study on which the method is based, after which the Systematic Data Integration method is introduced with examples from the study. The method entails analytical work on field notes and interview transcriptions, during which raw data and preliminary analysis results are integrated. It will be argued that interweaving the data clarifies the complementarity of the types of data and increases transparency for further theoretical analyses and interpretation. In addition, the interweaving contributes to ensuring quality in choosing quotations used to illustrate and emphasize the study's findings.
Accepted and frequently used methods of analysis may be viewed as inadequate for analytical work on field notes and interview transcriptions when the aim is to analyze these as a complete body of material. The purpose of this article is to present a new method for systematic data integration in which the key component is the interweaving of observation data and interview data from sequences of collaborative situations (Figure 1). The method entails analytical work on field notes and interview transcriptions, during which raw data and preliminary analysis results are integrated. The first part of the integration was inspired by Vatne and Fagermoen (2008), who developed a method for analysis of field data from single situations (Event-Oriented Data Integration). Systematic Data Integration comprises an integrated analysis of sequences of situations (hereafter called pathways).
An Overview of the Systematic Data Integration Method
Systematic Data Integration was developed within a study of groups of user representatives and professionals who collaborated for some time on planning, implementing, and evaluating patient education programmes (Strøm, 2010). We consider the method applicable to studies of processes—in both dyads and groups.
Fieldwork is recognized as an appropriate approach in studies of interactions that are not structured by the researcher. Observation is often combined with the qualitative interview to acquire more perspectives or in-depth information on a phenomenon. Field notes and interview transcriptions are usually analyzed separately, but literature on methodology also encourages researchers to view field notes and interview transcriptions as equivalent texts, and as a complete body of material for analysis (Fangen, 2004; Polit & Beck, 2004). Whether observation data or interview data best communicates what “really” happened is part of a long-standing debate in social research. From the beginning, participant observation was viewed as the “gold standard” due to the method's completeness. Observation was considered special because of the richness of direct interaction between study participants (Atkinson & Coffey, 2003). Doubt was cast on the interview as an appropriate method with the claim that it produced explanatory, dismissive, and staged data—even untrue data. Social researchers have gradually turned the discussion toward the differences in the various methods and their relative strengths and potential complementarity (Atkinson & Coffey, 2003; Polit & Beck, 2004). Other literature on methodology claims that observation data are often assigned less importance in the analysis and presentation of findings than interview transcriptions (Fangen, 2004; Prieur, 1993).
The following briefly presents the study on which the method is based. The Systematic Data Integration method is then introduced, with examples from the study mentioned. It will be argued that interweaving the data clarifies the differences between and complementarity of the types of data, and increases transparency for further theoretical analysis and interpretation. In addition, this integration helps to ensure the quality of the choice of quotations used to illustrate and emphasize the study's findings.
Background of the Method
A Study of Collaboration on Patient Education in Hospitals
Learning and Mastery Centres (LMCs) at hospitals in Norway are competence centres that play a key role in the guidance and training of those who suffer from long-term illness. These patients require more than medical help to enable them to live with the changes they must cope with in daily life. In this work, the experience of the users is regarded as equally important as the contributions of the professionals. The centres aim to achieve equal relationships between users and professionals. The user representatives are thus given a completely new role as contributors to the service, which entails the systematization of collaboration between users and professionals. A collaboration group is set up for each learning programme that is developed at the LMCs. The group normally consists of at least two user representatives, relevant professionals (e.g., a doctor, nurse, and physiotherapist), and the director of the centre. The planning, implementation, and evaluation of each learning programme are the responsibility of this same group.
The purpose of the study was to unravel what characterizes this collaboration. The result of the study was an inductively developed collaboration model that provided an overview of various forms of collaboration and associated power aspects. The model consisted of three axes: (a) mutual acknowledgement and sharing, (b) decisions, and (c) knowledge, and was compiled via newly developed concepts of types of user representatives, professionals, and collaborative meetings. The typology described different forms of collaboration and the dimensions of the collaboration (Strøm, 2010).
Preconception in Brief
At the time of the study, LMCs and their activities comprised a new practice field on which little research had been conducted. The centres lacked an explicit theoretical perspective on user involvement and collaboration. The fieldwork was not based on any specific theories of collaboration. At the beginning of the study, there was little documentation internationally on user involvement in the planning and implementation of patient education (Crawford et al., 2002). International research was clearer regarding empowerment and liberating pedagogical methods in patient education (e.g., Cooper, Booth, Fear, & Gill, 2001; Roter, Stashefsky-Margalit, & Rudd, 2001). Our initial understanding included knowledge of these theoretical approaches (e.g., Freire, 2005; Gutièrrez, 1990, 1995; Wallerstein, 1992). Furthermore, we wrote a paper that discussed possible interpretations of juxtaposing user knowledge and professional knowledge from an epistemological viewpoint (Strøm, Kvernbekk, & Fagermoen, 2011).
Samples and Methods of Data Collection
A field study on collaboration in LMCs was conducted. Data were collected in three different hospitals. Two types of samples were included—collaboration groups and individuals who took part in these groups. We observed collaboration in 17 groups of approximately 100 diverse members. The field notes were based on non-participating observations in 46 collaboration meetings. The notes were detailed and included physical surroundings, who was involved, what was said and done, procedures and actions performed, continuous activities (events), sequences of events and interactions over time, what some were trying to achieve, and expressions and indications of feelings involved. The field notes were descriptive but also contained reflections. After the observation, the notes were separated into observation notes and methodological notes, as recommended by Schatzman and Strauss (1973). Ten users, eleven professionals, and three LMC leaders were interviewed. The interviews were held after the observation period at each centre. Total interview time was 13 hours. The interviews were open and were governed by an interview guide embodying keywords on the following themes: experience of participating in collaboration and of offering personal experience and competence, and perception of own role as a new role. The questions were quite open. For example, “How will you describe your participation in this group?” All the interviews were recorded and transcribed with no repetitions or incomprehensible sequences or sounds. Reflective notes were made after each interview.
The sources of inspiration for conducting the interviews were Kvale (1997) and Fog (2004). The abundant literature on methodology made us familiar with analytical techniques, such as alternating between open reading and theme-focused reading (Album, 1996; Aronson, 1994; Coffey & Atkinson, 1996; Fangen, 2004; Kvale, 1997; Kvale & Brinkmann, 2009; Schatzman & Strauss, 1973).
Analytic Processes Toward an Integrated Analysis
The data material amounted to a total of 440 pages of text of different types–field notes and transcribed interviews. First field notes and then interview texts were each analyzed separately. A short description is given below.
Separate Analysis of the Field Note and Interview Data Using a Well-Known Methodological Approach
The field notes were given precedence, in the sense that analytical processes started on the basis of the observations. We considered observational precedence to be important, inspired by the literature and in line with our perception of observations as the more complete data and larger units of data. The analysis of the field notes started with what is termed management of data for analysis (Kvale & Brinkmann, 2009, pp. 189-199; Polit & Beck, 2004, p. 570). Because this was one field study on three sites, the field notes from all three LMCs were generated in one document. The text was then read to identify characteristics of the collaboration. The point at this phase of analysis was to conduct an open analysis and acquire an overview of the field notes as a whole, including what struck us immediately and what main themes emerged. Two main themes were revealed in this open analysis: (a) participation and (b) interplay of knowledge.
After the analysis of the field notes, the interview transcriptions were subjected to a thematic analysis and meaning condensation, that is, extraction of meaning (Kvale & Brinkmann, 2009). All interviews were first read in the same order as they were conducted. Then we read all the interviews with user representatives; first with health professionals and then with leaders. Detailed reflective notes were taken during this process. The research question on the characteristics of the collaboration was focused on in this analysis of the interview transcriptions. In addition, this analysis was based on the themes from the interview guide and the main themes from the analysis of the field notes. The overall result of this analysis was a large number of subthemes.
Meaning condensation entails providing a clear meaning of participant utterances. The purpose of the process is to make colloquial language more formal, to protect the interviewees from any recognition of language particularities, and to help avoid assessing the statements as naive (Kvale & Brinkmann, 2009, pp. 205-207). It was, therefore, decided that the texts that were identified as subthemes should be condensed. Table 1 provides an example within one subtheme: “holding several roles in the collaboration.” The statement in the table below was made by a person who is both a professional and an experienced user. The left-hand column contains a natural unit of meaning (an original statement). In the right-hand column the meaning is condensed.
An Example of Meaning Condensation Within One Subtheme: Holding Several Roles in the Collaboration
After thorough separate analysis of the field notes and the interview transcriptions, the status of the research was as follows: main themes had been identified in the field notes, as well as a considerable number of subthemes from the interview transcriptions, and the essential content of key interview passages had been condensed. Nevertheless, we had the impression that it might be possible to uncover more aspects of collaboration through a more integrated analysis. These preliminary results were placed in two different documents—one for field notes and one for interview texts. Two residual documents were created for notes and leftover interview text passages, that is, small talk, irrelevant matters, and privacy.
Systematic Data Integration—A Method for Combined Analyses
The study's interactional focus was on groups that collaborated over a period of time, with a defined objective. The meetings in the collaboration groups involved planning, implementing, and evaluating the course; the same participants took part in several collaboration situations over time. This sequential interaction we named a pathway, and we proceeded to look more closely at the collaboration process. We sought a manner of developing an analysis that would allow original field notes and interview texts, main themes, and subthemes to be integrated.
The result was two types of processes: data management combined with data analyses, both supporting each other. These processes enabled us to simultaneously analyze both types of texts. Hence, a new method for Systematic Data Integration was developed that involved interweaving observation data and interview data from the pathways of the group interactions. In this process, both main themes from the field note analysis and subthemes from the interview analysis were applied as structuring elements for further integration. Both original field notes and interview texts were linked to these structuring pivots. In our material, six pathways were suitable for integrated analysis.
General Description of the Method
The method includes two interweaving processes, and a short version of these will now be described. Below, the method is written in a manner so it can be used in other studies that explore what is going on during interactions, that is, in a study group or in a parent-child relationship. The systematic integration analysis is conducted in two phases. The first interweaving process is a variation of the Event-Oriented Data Integration method (Vatne & Fagermoen, 2008), in which field data and interview data from the same single event are interwoven. In our study, work was first done one by one on each pathway. The combination of field notes and interview transcriptions was thereafter expanded, by interweaving all pathways to form a comprehensive and new text variant. The interweaving processes can be illustrated in six steps.
Illustration of the Interweaving Processes.
For each pathway, that is, a sequence of interactions, raw data from field notes are copied into a document with three columns and thereafter subjected to open thematic analysis. Field notes are placed in the left column, the identified themes in the middle column, and matching interview sequences in the right column. Matching is an analytic process to identify and select interview passages, and involves a thorough analysis of the transcribed interviews. The result of this first interweaving process will be a reconstructed and integrated text from each pathway. Next, each reconstructed text is read vertically and horizontally to explore what is going on during sequences of interactions. By reading vertically one can follow the interaction processes in more detail, both in the field notes and the interview transcriptions. In our study, by reading horizontally the sequences of collaborations were analyzed in this order: from planning, to implementation, to evaluation of the course. Analysis was primarily carried out in this order to clarify whether, and how, our observations and reflections corresponded to the participants' experiences and opinions and vice versa. During these readings, patterns of different interaction processes might emerge and can be related to the participants' attitudes and contributions. The second interweaving process involves bringing together all the reconstructed pathways in a large comprehensive text. The integrated analysis of this text is guided by the primary research question. During this analysis the patterns of typical and contrasting interaction processes may be formed.
Illustrations From Our Study
With regard to step 1, the pathway in our study consisted of meetings for planning, implementing, and evaluating a learning programme.
With regard to step 2, the analysis processes revealed themes such as collaboration, participation, and knowledge in interplay. The field notes from each pathway were analyzed anew with the open question: “What themes, apart from collaboration, is this text really about?” One result of the new open thematic analysis was that collaboration was specified in a subtheme, such as leadership of the group, planning of user input and professional input, apparent lack of planned pedagogical approach, humour, the art of conversation, and enthusiasm.
With regard to step 3, the interviews from the six pathways were already analyzed, the subthemes identified, and the interview transcriptions condensed. With the themes from the first and second analyses of the field notes as guidelines in the middle column, relevant interview transcriptions on the theme were selected. Thus, themes that corresponded with each other in terms of content in the two types of text were matched. Obviously, the matching of the field notes and the interview extracts involved both subjectivity and reflexivity, which are recognized as conditions connected to “researcher-as-instrument” (Hammersley & Atkinson, 1995). Identifying texts and themes that correspond with each other was a creative process that represented far more than a linguistic pursuit to match synonymous themes. At the same time, it required precision in reading the rest of the documents to make sure that no important data were missing. Both parties in the research team were involved in this reading. This process was repeated for all six pathways identified.
Below are two extracts from extensive examples of how the texts were interwoven. The first subtheme is “Unstructured route to form and content in the learning programme” and the second is “Art of conversation and humour.”
With regard to step 4, when the text was organized in the above manner, it could be read both vertically and horizontally. Having the material “side-by-side” paved the way for other questions such as: Do they do what they say they are planning? Have they planned what they are doing? How do they evaluate afterwards? Reading horizontally was also important as a form of validation between the field notes and the interview transcriptions.
With regard to step 5, we restructured all the pathway documents into a new text variant of the entire material to make a new overall analysis. Now, the research question of the study was divided into subquestions: (a) “What characterizes the planning of the program?”—We wanted to find out who was present, how did people make themselves heard, and who talked to whom? Moreover, we tried to discover how the processes were led, what the participants thought was important to give priority to, and how decisions were made, (b) “What characterizes the implementation of the programme?”—We wanted to find out who was actively involved in the learning programme, what happened when user input took place, what happened during and after the professional input, when humour had emerged, and whether the education programme was formed with an explicit purpose, and (c) “What characterizes the evaluation?”—We wanted to find out how the evaluation was made, what was evaluated, and what the consequences of the evaluation were. This restructuring occurred in light of the three subquestions. All interwoven texts (as illustrated in Table 2) were positioned in relation to the research question to which it belonged. The restructured text comprised 45 pages.
Two Extracts From an Extensive Text Illustrating Interwoven Texts
With regard to step 6, this second interweaving process of the texts demanded both time and work, but it was essential for creating an overall perspective and acquiring integrated material for further analysis. In our study, these analyses revealed that we were faced with at least two typical and contrasting types of collaboration. One group was characterized by equivalent collaboration, proactive participants, and mutual attentiveness. In the other group, collaboration was characterized by domination of the professionals with the users given less time, importance, and influence in the processes. Hence, two contrasting cases of collaboration were identified.
Discussion and Conclusion
The Systematic Data Integration method helps to elucidate and utilize the complementarity of the various types of data with the interweaving of texts as an essential step. We have been inspired by good methods for analyzing data from various data sources (Vatne & Fagermoen, 2008) and hold that sharing examples of analytic processes and procedures are important in the development of qualitative research. There are different views on using “recipes” in qualitative data analysis. Miles and Huberman (1984) offered the term “rules of thumb for the analysis of matrix data” (pp. 213-214) at an early stage, and included information on horizontal and vertical reading of data. In its original form, Grounded Theory (GT) emphasized that data should speak to the researcher without a schema indicating a path for the analysis of the material (Glaser & Strauss, 1967). Terms such as matrix and axis coding, however, have been included as types of analysis schema in revised versions of GT (Midré, 2009; Strauss & Corbin, 2008). Abstractions based on the participants' actions and their meanings are linked to the construction of theoretical concepts and models (Polit & Beck, 2008, p. 230). In our vertical and horizontal analysis, more abstract concepts such as user representative were developed into more empirically close concepts such as novice, advanced, and professional user.
The development of Systematic Data Integration resulted from an early decision when planning the analysis; we based the analysis on the observations because “That's where it happens.” Our research situation was such that we had open access to an unexplored field. As a result, we also felt ethically obligated to apply as much of the observation data as possible. We wanted the observations to have a legitimate place in both the analysis and presentation of the material. In our method, field notes are given precedence because observation data are the first to be analyzed, and because the original field notes are used once more when we start interweaving the different texts of each pathway. Instead of giving the interview transcriptions precedence for the choice of quotations, the matching of the two types of texts contributes to choosing quotations from both depending on which text best illustrates and emphasizes the findings.
Interweaving as a form of analysis provides transparency—for both the researcher and the reader. The analysis is described and can be followed in relative detail. The interweaving made the material clearer for specific choices of relevant theoretical perspectives. As an example, our initial understanding was that empowerment could be a suitable theoretical perspective. Through Systematic Data Integration, however, it became clear that new theory must be read and brought in. Theory on citizenship was shown to be more suitable for understanding user involvement and group collaboration.
Systematic Data Integration helps to create and clarify an audit trail of the analysis. The interweaving contributed to our choices of suitable theoretical perspectives. New theoretical concepts formed the basis for the development of a model for different types of collaboration in a practice field that was pervaded with “big” terms (such as user involvement), without the terms being clearly rooted in the applicable context. The interweaving prepared large amounts of data material for further analysis and for validating the analysis. The method made it possible to read across the various texts. In the analysis it enabled us to validate whether, and how, our observations and reflections corresponded with the participants' experiences and opinions, and vice versa. Transparent descriptions and validation of the analysis are essential according to Malterud (2001):
Declaring that qualitative analysis was done, or stating that categories emerged when the material had been read by one or more persons, is not sufficient to explain how and why patterns were noticed … the reader needs to know the principles and choices underlying pattern recognition and category foundation. (p. 486)
A second important validation of our method of analysis is included in the term communicative validity (Kvale & Brinkmann, 2009). During the analytic process, the analysis and presentation of the preliminary research findings were submitted to various qualified readers. These peers possess sound scientific competence in field research and also qualitative approaches in general. The purpose was to generate constructive disagreement that could help to create new ideas. Their reading and the discussions between them and us as authors contributed major critical input and positive feedback on the analysis. All in all, the peer assessment supported the fact that the method had made it possible to bring new inductively developed concepts to the field.
We consider that the Systematic Data Integration method is applicable to other studies that may require analytical work on an integrated body of material from at least two data sources. The method is suitable for research with a focus on interaction that continues over some time to enable identification of pathways. Hence, at least two collaborating parties are needed for the study—either dyads or groups. An example of dyads is a study with a focus on collaboration between one patient and individual professionals in an interdisciplinary team during a four-week rehabilitation period. The data material will include both field notes and transcriptions from individual interviews. This will generate x number of pathways of dyads, comprised of the patient and each of the different professionals. The various pathways are finally interwoven, for example, in view of a purpose to identify general characteristics of the collaboration in rehabilitation. Another example is a study of the supervision relationship between tutor and student nurses. Groups of student nurses have weekly supervision sessions with their tutor in an eight-week clinical practice designed to develop professional skills and skills in reflection on practice. In this example, the pathway will consist of eight supervision meetings, with data that includes, for example, observation notes and transcriptions from focus group interviews. The number of groups studied determines the number of pathways that will be restructured for final analysis.
In our study, the Systematic Data Integration method resulted from the experience that separate analyses of field notes and interview transcriptions were not adequate to allow all the data material collected to be utilized or to develop new knowledge about collaboration. We consider the interweaving of equally regarded and legitimate sources of data to be generally applicable in studies of interactions—both in dyads and in groups. We hope the method can also provide inspiration for further development of methods for analysis in field studies.
Footnotes
Acknowledgements:
This project has been financially supported by the Norwegian Extra Foundation for Health and Rehabilitation through EXTRA funds, and by Diakonhjemmet University College, Oslo, Norway.