This case study examines the development of a general analytical inductive approach to qualitative research. It assesses the research design and analytical processes for developing a framework for understanding the development of collaborative partnerships between business schools and industry. I adopted an interpretivist approach which is about understanding how people make sense of their world. The outcomes demonstrate how an analytical inductive analysis, involving detailed readings and interpretations of raw data, can be used to identify concepts, themes, and models. I gathered data from both a university business school and firms; I transcribed and analysed 24 interviews, and produced a list of common features and accommodated deviant features either by linking them with other common features or by generating a new category. Eventually, I performed cross-case analysis within and between the groups at the business school and firms. The themes that emerged formed the basis of proposed models for initiating collaboration and initiating trust. Through practical application of the research design, data collection and analytical approach, this case study demonstrates the credibility of a general analytical inductive research strategy based on a qualitative research methodology. The benefits include facilitating the development of effective business relationships between universities and domestic firms.
By the end of this case, you should be able to
- Understand the application of the research design, data collection and analytical approach in relation to theory building
- Study the credibility of a general analytical inductive research strategy based on a qualitative research methodology
- Advance the knowledge on how general analytical inductive strategy promotes the linkage of theory and practice
- Comprehend the step-by-step process of qualitative data analysis and theory building grounded in data
This case study examines the development of a general analytical inductive approach to qualitative research. It critically assesses the research design and analytical processes that facilitate the development of a framework for understanding the development of collaborative partnerships between university business schools (UBS) and small- and medium-sized enterprises (SMEs).
Salter, Tartari, D'Este, and Neely's (2010) research argues that corporate demands are increasingly leading organizations to engage in ‘partnerships’ for the advancement of ‘collaborative advantage’ and the higher education sector is no exception in seeing this as a key focus for its business activities. I believe that the continuing need to develop practice-oriented theory into the management of inter-organizational collaboration has led many researchers to focus on exploratory research. This case study assesses the contribution that an interpretive research methodology, based on analytical inductive principles, can make to praxis. Johnson (1998) claims that the interpretive research methodological approach embodies analytic induction, the principles of which focus on the generation of theory from the observation of the empirical world of the participants. Thus, it facilitates the research need to develop a practice-based theory or model concerning the role of universities in building knowledge transfer (KT) between the two business sectors (see Darabi & Clark, 2012).
The existing literature highlights the importance of relationship management in inter-organizational collaboration. The purpose of my research was to understand how the relationship is initiated, formed and managed in the context of a UBS and SMEs, with the primacy of trust being seen as a key factor for collaborative development. Therefore, this case study explains how the appropriate research method was designed for this purpose.
My research was focused on making suggestions for building strong business links and also suggested a practical model of collaboration or partnership between the business school and SMEs. Therefore, I conducted an investigation into ‘what is going on’ and ‘how things take shape’ in the relationship between the two sectors. Understanding how the parties in each sector make sense of their relationships was key, so the research questions were designed as follows:
- how much do UBS and the SMEs understand about each other?
- what are the perceived benefits for UBS and the SMEs to improve their relationship? What motivates SMEs to engage with the business school?
- how were relationships and communication initiated, formed and managed between UBS and SMEs, from the point of view of both sectors?
- what are the barriers in the relationship between the UBS and SMEs, and how can the UBS overcome these barriers?
- what is the role of trust in such relationships? Is it a factor which accelerates the relationship between the two sectors? How can it be built in practice in such a relationship?
To explore the answers to the research questions, I will explain my research design in the following order: (1) the interpretive approach followed by Blaikie (2007) as the underlying philosophical assumption of the research; (2) the research method including, for example, interview technique and data collection technique; (3) data analysis approach – general analytic induction – followed by Johnson (1998); (4) step-by-step data analysis procedure followed by Thomas (2006); (5) the process of data reduction and emerging themes; and (6) the model developed from the data. At the end of the case study, I will explain the problems encountered and ethical issues.
An Interpretive Approach in Developing Business Relationships
Creswell (2003) and Guba (1990) argue that most of the research paradigms share three fundamental elements: ontology, epistemology, and methodology. Ontology concerns the nature and form of knowledge in the physical world, while epistemology concerns the source of knowledge, or the ways of knowing knowledge. The actors in the two sectors were the source of knowledge for my study. The way in which I gained access to the knowledge or social reality was through interviewing the actors. The third fundamental element is methodology, which according to Creswell (2003), and Guba and Lincoln (1994), concerns the rationales behind the procedures used to research what it is believed it is possible to be known. Because the researcher's philosophical position can shape the research design, knowing how I as a researcher explore the reality of the relationship between UBS and SMEs was important in order to adopt the appropriate method to study the phenomenon.
Research can be undertaken through a number of research paradigms that shape the research design. It was possible to take a positivist, postpositivist, interpretive, critical theory, constructionist or a postmodernist approach to study a business relationship between UBS and regional SMEs in my research. The inductive or interpretive approach was undertaken because I believe that the way that people behave is based on their interpretation of a phenomenon.
I was trying to understand how actors make sense of their relationships. Gill, Johnson, and Clark (2010) argue that the aim of interpretivist approaches is to understand (verstehen) how people make sense of their worlds. Interpretivists such as Blaikie argue that the study of social phenomena requires an understanding of the social world that people have constructed and which they reproduce through their continuing activities. I share the same belief with Blaikie that people are constantly involved in interpreting and reinterpreting their world – social situations, other people's actions, their own actions, and natural and humanly created objects. Hence, I thought individuals in SMEs and universities would likely view the phenomenon of collaboration differently, which was my main reason for adopting an interpretivist analysis for my doctoral research.
The interview technique was the main method I used for collecting data. The type of interview was semi-structured, in-depth and face-to-face: I was interested in getting a detailed and deep picture of each interviewee's perceptions and understanding about the interactive process of the collaboration. Semi-structured interviews are widely recognized as being able to facilitate this aim, and as Silverman (2005) stresses, ‘qualitative interviews provide us with a means to explore the points of view of our research subjects’.
Through the interview sessions, my focus was on finding the history of SME relationships with the university. The advantages of this method, according to Creswell (2003), are participants can provide historical information and it allows researcher ‘control’ over the line of questioning. It allowed me to fully explore the topic from the respondent's perspective. As Denzin and Lincoln (2005) argue, the interview is a conversation; it is not a neutral tool, for at least two people create the reality of the interview situation. In total, I conducted 24 semi-structured interviews, with 13 managers in the SMEs and with 11 academic managers in the business school.
As a result, I gained historical and in-depth information about the successful and unsuccessful examples of working between the two sectors by asking actors in both sectors about the stories of their current and past relationships. The process of data collection started in January 2010 and finished in January 2011.
Purposive Sampling Technique
I followed the purposive sampling procedure in the selection of the organizations and the participants in the research. Creswell (2003) argues that the idea behind interpretive research is that the researcher purposefully selects participants who are viewed as most likely to help the researcher understand the problem and the research question. Thus, I was aware that purposive sampling, as Silverman (2005) claims, demands that the researcher think critically about the parameters of the population under study, and I chose the sample cases carefully on this basis. In this manner, I identified a number of SMEs which were already interacting with the business school; then I selected the organization sample to represent a range of businesses based on the type of the services that they received from the business school.
As a result of using a purposive sampling approach, four more interviewees were identified as potential participants. For example, I interviewed two sales managers from two companies because they were recommended by their managing directors to participate in the research due to their interactions with the business school.
The participants were chosen from strategic and operational levels, for example, managing directors and operational directors in the SMEs and senior management and project managers at the university and business school. It was important for my research to gain the views of senior strategic management at the business school and within the SMEs with respect to the development of effective relationships because these people were the decision-makers in their organizations. In a similar way, it was important for me to gain access to the perceptions of project managers and operational directors as these could give key pictures of the relationships.
The broad range of channels of interaction – such as ‘joint research projects’, ‘contract research’, ‘consultancy’, ‘training of firm employees’, ‘postgraduate training in the company’, ‘recruitment of recent graduates or postgraduates’ and ‘student placements’ – were examined by D'Este and Patel in 2007, and Bruneel, D'Este, and Salter in 2010. Building on D'Este and Patel's research, my research considered ‘knowledge transfer projects’ and ‘consultancy projects’ as channels of interaction between UBS and SMEs.
Drawing on the themes derived from the literature (e.g. the organizational processes as barriers to developing collaboration in Marzo-Navarro, Pedraja-Iglesias, and Rivera-Torres's (2009) research, and the need to create a knowledge sharing culture based on trust in Vangen and Huxham's (2003) study), I designed a list of interview questions. This list was based on two main themes – relationship management and collaboration in business relationships – to explore how individuals make sense of their business relationships so as to give me a deep insight of each theme.
I asked the participants to state their role, the history of their involvement, and some of their experiences working within the partnership. Therefore, the questioning attempted to develop a picture of the story behind the relationships, for example, ‘How did you start the relationship with ____?’ ‘Do you have any successful or unsuccessful experience of working with ____ ?’ ‘Why do you want to develop the relationship?’ ‘What are the perceived advantages of working with ____?’ ‘How can things be improved?’
I did not follow a rigid sequence in asking the interview questions. Instead, the sequence was varied in order to pick up on specific points made by the interviewee; however, I ensured that all the questions were answered by the interviewees. The same questions were asked with all the interviewees. Through this type of semi-structured interview, interviewees were encouraged to talk about what they were experiencing and what they thought were key issues. It gave me a chance to clarify the questions and answers and also ask new questions as follow-up to an interviewee's replies to get a rich picture of each interviewee.
Data Analysis Approach and Procedure
In social science research, there is a need to focus analysis on which explanations of human action are generated inductively during data collection to develop an understanding of the interpretations deployed by the actors who are being studied (see Denzin & Lincoln, 2000; Giddens, 1976; Shotter, 1975). The aim of my research was to contribute to the understanding of how UBS initiate, develop and manage their inter-organizational relations with SMEs; and what, if any, are the mutual advantages for business schools and SMEs to work collaboratively. As the research progressed, it became clear that the data suggested that the key issue was how to initially start the relationship, and trust became the central phenomenon of interest. Therefore, I decided to adopt a general inductive analytical process for the interpretation of the data.
General Analytic Induction
The method of induction, according to Locke (2007), is the process of proceeding from particulars to the general – universals. Its process starts with an observation or something that is a puzzle and needs exploration (e.g. a general question) and ends up with a new theoretical conceptualization of the issues. According to this approach in social research, researchers attempt to generate theory at the end of the research. Generalization is questionable in this approach because, according to Bryman (2004), the scope of the findings of qualitative investigation is restricted, and it is impossible to know how the findings can be generalized to other settings. Can just one or two cases be representative of all cases? The answer is no, but the findings of qualitative research are intended to generalize to theory rather than the population. It is the quality of the theoretical inferences that are made out of the qualitative data that is crucial to the assessment of generalization.
However, Gill and Johnson (2006) argue that human beings are able to attach meaning to the events and phenomena that surround them. Therefore, examining people at SMEs and UBS could result in different understandings and different views of their relationships because they are from different organizational contexts and have different experiences of relationship with each other. It can also provide good contrasts and comparison and thereby confront the emergent theory with the patterning of social events under different circumstances, according to Johnson (1998).
As Johnson (1998) claims, analytic induction is a set of methodological procedures that tries to generate theory grounded in the observation. This approach shaped my thoughts in applying an analytic induction approach for the data analysis. Figure 1 illustrates the processes.
Figure 1. Data analysis approach – analytic induction.
As Figure 1 shows, in developing the analytical approach, data from both sectors were gathered and the interview transcripts were analysed, producing a provisional list of common features and identifying deviant cases. Then, similarities between categories were established. Deviant features were accommodated either by linking them with other common features or by generating a new category with unique features. Eventually, I performed cross-case analysis within the groups and between groups at the business school and SMEs. A number of themes emerged from the data and tentative models of initiating collaboration and initiating trust building that is linked with Vangen and Huxham's ‘trust building loop’ suggested.
Data Analysis Procedure and Data Reduction Process
Thomas (2006) argues that many of the underlying assumptions and procedures associated with qualitative data analysis are related to specific approaches or traditions, such as Strauss and Corbin's (1998) grounded theory, Potter and Wetherell's (1994) discourse analysis and Lieblich's (1998) narrative analysis. However, a much-used strategy in qualitative data analysis is the general inductive approach by Bryman and Burgess (1999) and Dey (1993). Thomas states that the inductive approach is a systematic procedure for analysing qualitative data and is one in which the analysis is guided by specific evaluation objectives. It refers to detailed readings of the raw data, and the researcher's interpretations of these data drive the identification of concepts, themes or models. The researcher begins with an area of study and allows the theory to emerge from the data, thus building an understanding of data analysis and theory in a manner that is consistent with Strauss and Corbin's grounded theory methodology. I adopted this approach in my doctoral research. I was following a systematic procedure with the aim of reducing the mass of raw data, through coding and categorizing it, such that clear links between the research objectives and the findings could be derived while ensuring that these links were both transparent and defensible.
As a part of the data analysis, I transcribed audio file interviews from both sectors, that is, the university and SMEs, into Microsoft Word files. Data analysis followed an iterative, constant comparative approach until data saturation occurred, or according to Glaser and Strauss (1967, p. 61) ‘theoretical saturation’ happened, that is, ‘where no additional data was found whereby the [researcher] could develop properties of the category’.
In other words, I was faced with repetition in the answers to the interview questions and no new relevant data were presented by the participants. Lincoln and Guba (2000) state that the process of coding can be finalized when the categories are saturated, incidents can be readily classified, and sufficient repetition occurs in the data. Thus, new data did not add anything to the development of the categories and the created models. For example, when participants at the business school responded to ‘Why do you need this type of relationship with SMEs’, the answers were almost the same, so there was no need to alter the categories which had already been created. Therefore, data collection ceased and the last interview had been conducted.
Silverman (2005) argues that transcription is a form of data analysis. Therefore, audio files were transcribed word for word and typed into Microsoft Word files. The process of transcription was time consuming; however, it enabled me to get very close and familiar to the content of the data. I read through each transcription several times. When reading the raw data, I had two options for managing and analysing the data, either manually or electronically (i.e. using specialist software).
The first trial was manual, that is, I read through the first transcription a few times to understand what each quotation, whether sentence, phrase or even a paragraph, was about and then wrote each on a Post-it Note and gave each a code based on my interpretation. These were then stuck to flipcharts, and the Post-it Notes moved around the flipcharts as necessary to form constructs, whereby Post-it Notes with similar context were clustered together. As the nature of qualitative data is bulky, it soon became hard to find enough space to hang the flipcharts and it was also too messy. Therefore, I decided to use NVivo software to help manage data. I imported the transcriptions into NVivo software while data collection was still in process. Figures 2–6 illustrate the step-by-step data analysis procedure using NVivo software.
Figure 2. Examples of SMEs and UBS folders in NVivo.
Figure 3. Examples of free nodes in NVivo.
Figure 4. Examples of tree nodes in NVivo.
Figure 5. Examples of categories and sub-categories in NVivo.
Figure 6. Examples of emerged themes in NVivo.
During the data analysis procedure, I read each transcription several times and labelled and coded every sentence, phrase or paragraph based on my interpretation of the raw data. Each sentence or phrase was separated from the body of the transcription and the initial letters of the participant's first name and surname were added at the beginning of each sentence or phrase followed either by the letter U or I. The letter U indicated that the quotations were from the university sector, and the letter I meant that the quotations were from the industry sector.
This differentiation was helpful for the last stages of data analysis, that is, comparing and contrasting the views in cross-case analysis within and across the groups. This enabled me to distinguish the quotations from university and industry quickly and to find the original text easily if needed. Each sentence, phrase or paragraph was considered as a free node. Free nodes were coded mainly from the words mentioned by the participants. Therefore, a list of 248 free nodes was established (see Figure 3).
At the second stage of the coding process in inductive analysis, I tried to find the similarities between the nodes. Miles and Huberman (1994) argue that pattern coding is a way of grouping codes into sets, themes or constructs, thus reducing large amounts of data into a smaller number of analytic units. Some of the free nodes had something in common in terms of meaning; therefore, they were merged together.
I opened every single free node before merging them to make sure that they had something in common; then I labelled them, creating 58 categories as tree nodes, that is, each tree node with few child nodes which were related to a category. The categories therefore developed from coding. Thus, Thomas (2006) argues that the label of each category carries inherent meanings that may reflect the specific features of the category (see Figure 4). Each tree node was considered as a category. Some of the free nodes stayed alone, as they were not linked to or fitted into any category.
At the third stage, there was some overlap among categories. Some of the categories which had a link or relation with other categories were merged together into a hierarchical category system and labelled with a broader heading. These links may point to superordinate, parallel and subordinate categories, for example, ‘advantages’ under the main category of ‘Purpose of Involvement’ (see Figure 5).
At this stage, the number of categories decreased to 16 in accordance with the theory of data reduction which expects a reduction in the number of categories.
At the fourth stage, I decided to select the most important categories to merge together to convey the core theme because some of the text was not relevant to the objectives of the research. Thus, three main themes and 16 categories emerged to create a model incorporating the most important categories. At this stage, as the amount of data was still too large, some of the categories were not assigned to the main themes and left aside to be used in suggested further research (see Figure 6).
As a result of this analysis, three main themes – relationship management, collaborative opportunities and challenges, and the role of trust – emerged from the data.
The amount of qualitative data presented methodological challenges. It was challenging to manage the data and to use the most appropriate data for analysis. I overcame this challenge by focusing on the objective of the research and selecting the information which was very much related to the objectives of the research.
The Process of Data Reduction and Emerging Themes
In this section, I explain the process of data reduction and the emergence of themes. Through this process, the research data were reduced into sub-categories and then categories and finally themes.
My analysis of the overall findings showed that a large part of the data shares the concerns related to relationship management. This led to the first emerging theme of the research, Relationship Management. Figure 7 shows the creation of Theme 1 by showing data reduction from sub-categories to categories and from categories to Theme 1.
Figure 7. The creation of Theme 1.
The data suggested some driving factors that facilitated initiating and developing the collaboration process. For example, from the academics' point of view, SME leaders with higher education are more likely to approach the university for any business issue. Previous experience and relationships are also drivers in initiating and developing a business relationship. From the practitioners' point of view, academics with commercial experience and knowledge can be an encouragement to SMEs in initiating a relationship with the business school. In addition, personal characteristics such as academics' interest in a particular area of business can also influence SMEs to initiate collaboration. These issues led the analysis to create a category under ‘motivation’ with sub-categories such as ‘leaders with higher education at SMEs’, ‘previous experience’, ‘personal characteristics’, and' academics with commercial experience'.
Figure 8 shows the creation of Theme 2 by showing data reduction from sub-categories to categories and from categories to a theme.
Figure 8. The creation of Theme 2.
The findings show that one of the approaches in initiating honesty and reliability is ‘delivering promises and competencies’. Moreover, ‘managing the expectations’ by ‘understanding the needs of SMEs’ and having ‘deliverable objectives’ are different ways of initiating trust with the other party. So a category which addresses ‘initiating trust-building approach’ was created to cover this area of the findings. Therefore another theme, The Role of Trust, emerged from the data (see Figure 9).
Figure 9. The creation of Theme 3.
The outputs of my research are two practice-based models of initiating collaboration and initiating trust in inter-organizational relationship. The models inductively developed from the data grounded in observation (see Figures 10 and 11).
Figure 10. Model of initiating collaboration.
Figure 11. Model of initiating trust.
Further discussion on the first model, initiating collaboration, is accessible on Darabi and Clarks, 2012 (http://www.emeraldinsight.com.lcproxy.shu.ac.uk/journals.htm?articleid=17036613).
Problems Encountered and Recommended Solutions
Although my research had a well-designed methodology, I encountered a number of problems during the process of data collection and data analysis. Therefore, this section provides some of the practical lessons I learned from my experience.
The first problem was related to using an audio file recorder. In one case, I did not press the record button properly and, as a result, missed a 1-h interview and needed to reschedule another appointment with the interviewee. In that case, it took 2 months to reschedule the appointment, as the interviewee was one of the senior managers at the UBS and it was hard to find free time to repeat the interview; also, the interviewee understandably did not have the same interest in the questions as she did during the first interview. Therefore, I required technical skill and attention to detail. This was a lesson for me to double-check the recording device even during the interview process for the rest of the interviews.
The second challenge was generally related to making appointments with the participants. At one interview with one of the senior managers at the business school, the interviewee agreed on a 1-h meeting. While in the middle of the interview, the interviewee received a call and apologized that she needed to answer the call and apologized again that she needed to leave the session in 15 min, thus cutting the interview to half an hour. At that point, I tried to ask the most crucial questions, as the view of that interviewee was important to the research. In other cases, my interpersonal skill and my network facilitated making appointments within the business school and with practitioners at the SMEs.
In another case, I would have liked to have had the view of one of the KT champions in another faculty to compare with the business school's KT champion; however, the potential participants from the other faculty refused to participate in the research. Getting access to the companies who work with the university was a major struggle too. The communication with some departments or people was smooth and helpful, but in some cases, it was very difficult to convince a department of the business school to contact their SME clients and ask them whether they would like to participate in the research. For example, the University Enterprise Centre, a gateway to external business enquiries, appeared to be extremely reluctant to link me with the companies as it kept passing me from one person to another, making the data collection process longer than expected. However, I took advantage of my own network and liaised with different departments in the business school to overcome this problem.
The third issue was related to transcribing interviews which were conducted on the interviewees' premises. Two interviewees were from manufacturing companies, and during their interviews, there was noise from machinery and from staff talking among themselves in the background. This caused some difficulty for me while I was transcribing the audios. However, I managed to get help from the notes that I made during the interview session, and after listening a few times to the audio, the problem was resolved.
Specifically, this case study focused on the design of a general qualitative inductive research approach. The conceptual justification for the qualitative approach adopted was based on a theoretical perspective premised on a commitment to verstehen. This represents approaches to management research where it is believed we can understand how others make sense of the social world and where the collected qualitative empirical data are used for the inductive generation of theory ‘grounded’ in observation.
Through a practical application of the research design, data collection and analytical approach, this case study demonstrates the credibility of a general analytical inductive research strategy as a qualitative research methodology, a strategy that promotes the linking of theory and practice. This research design was highly appreciated verbally at the Viva session by the examining team of my thesis.
I believe the main benefit from my research is facilitating the development of effective business relationships between local universities and SMEs in their regions. In this manner, key benefits for companies from collaborative partnerships can be gained through enhancing management development activities in order to change business behaviours.
The models that I have developed from the data can be useful to the business schools executives, for example, with strategies for developing relationship with external stakeholders.
Exercises and Discussion Questions
- In my research, I adopted general analytical induction as a data analysis approach. What are the advantages and disadvantages of this approach from a researcher point of view? Discuss the pros and cons.
- In my study, I followed purposive sampling technique for the research data collection. Critique this method and suggest alternative methods of sampling.
- I have used several figures to present my data reduction procedure. Evaluate and discuss the process of data analysis and presentation, and suggest alternative ways of presenting the data analysis procedure.
- Reflect on the method discussed in this case study and highlight four main learning points. Discuss how you could use these learning points in your current or future research.