All 16 entries tagged Mixedmethods
No other Warwick Blogs use the tag Mixedmethods on entries | View entries tagged Mixedmethods at Technorati | There are no images tagged Mixedmethods on this blog
April 08, 2018
Continuation of the Qualitative Thread: Current Thoughts on Axial Coding
In the previous blog post I talked about the role and function of Open Coding, which is to label data segments with meaningful codes that summarise the content, features, characteristics, events and activities within that data segment and from these codes, develop categories and their properties and dimensions. Remember that categories are a collection of similar codes, with data segment characteristics represented as properties and dimensions. Open Coding from what I can understand is essentially descriptive where it attempts to describe the features and characteristics of data through coding and category development, and as argued by some authors, carries realist assumptions based on its use of constant comparative analysis. I did ask questions about where to go next following Open Coding, and now I think I have the answer.
I had my doubts about Axial Coding initially simply because of the challenges and criticisms against Axial Coding from various authors, who shall be engaged with on here at some point and especially in the thesis. But you should never adopt or reject an approach just because others have criticised it: you should instead adopt or reject an approach based on its relevance and suitability for your research project. As long as you justify and reason why (and why not) you have used (or have not used) particular design components and that they are aligned more generally with your philosophical and theoretical (if appropriate) assumptions then you are within your right to use any coding form.
I have now come to the idea that Axial Coding is the most sensible next step level of coding for my research. Open Coding then is that descriptive approach to developing categories; Axial Coding, therefore, is a more abstract means of coding that involves linking or relating categories together in order to better understand a process not through the views and experiences of those experiencing a process, but through exploring the process itself. Axial Coding is beginning to be understood therefore as a means of developing relationships between categories, and of developing relationships between a category and its own properties and dimensions.
During the process of redeveloping my understanding of Open Coding, conceptions of categories were formed: what a category is, what information should best be part of a category, and the guiding questions I have when developing a category further in terms of its properties and dimensions. I have noticed upon further reading that some of the questions I ask of a category, some of the questions align with the purpose of Axial Coding but that’s fine as some authors have stated that the thinking about relationships between categories and between a category and its own properties and dimensions occur during the Open Coding stage. As I recode the data I shall have additional questions though, and are based on the development of the relationships and they include: what forms a relationship? How can I identify a relationship? What is the content of this relationship? What are the features and characteristics of this relationship? What is the influence and impact of the context of the situation upon the relationship? Axial Coding, then, not only establishes relationships but also appears to acknowledge and consider the context of the relationship. For example, a relationship between two categories might differ between different contexts and this is important when exploring learning phenomena.
Although Axial Coding can establish and identify relationships between categories, properties and dimensions, it does not, as far as I can currently understand, produce an actual network of activities and events relating to the sustainability and on-going nature of social learning situations but it can provide the foundational understanding of what is occurring within a discussion through categories, dimensions and properties. However, it might be possible that a grounded theory’s relationship identification process and network diagrams and associated analysis attains a better understanding of certain social learning processes.
This is simple a try it out and find out approach, but from what I have drawn out in a presentation that I am producing for the topic it appears that this is the limitation of grounded theory and hence the introduction of a network analysis method and the interest in quantitatively analysing relationships but this is for another blog post.
In all, Axial Coding makes more sense to my research now if I view it as a means of relating categories, and to relate categories with their properties and dimensions. This makes sense to me because clearly defining relationships between categories and their dimensions and properties shall assist with understanding the complexity and highly nuanced existence of certain learning phenomena and provide a basis upon which I can build complex networks and be able to quantitatively analyse the relationships between these categories.
That’s the picture of Axial Coding for now!
‘till next time!
April 06, 2018
Some Current Thoughts on the Qualitative Strand and Open Coding
There has to be a sense of emergence where codes are derived from data and categories are derived from codes. This idea of emergence makes sense as I have not been able to identify an existing framework that is suitable and relevant for the type of data I am using and the way that I am now exploring the phenomenon of interest, and therefore, theoretical constructs, relationships and hypotheses must emerge from the data.
I read an interesting paper the other day where an author aligned the idea of an emergent approach with the realist ontology: truth emerges from the data after a continuous cycle of coding and recoding, but this brings about a couple of problems. First of all, what is defined as truth and of being true? How can it be measured? How can I know that something is true even after going through continuous cycles of coding and recoding? How can I know that this truth emerges from the data and not simply a reflection of my interpretations of the data? Is it absolute truth that emerges from the data or is it that with each coding and recoding I could come closer to the truth without completely attaining it? How do I know either way? Is there some set criteria for truth? If so, then would this criteria itself represent truth if it’s simply been constructed by another human being? Would it therefore be better to consider the set criteria is bringing one closer to the truth rather than mirroring the location of absolute truth?
One thing I do know is when I think about the qualitative strand, the purpose that it brings to the research, and what I currently would like to achieve with the strand, allowing the data to speak for itself; to enable this “voice” to emerge naturally and to code in accordance to what is believed to be occurring within the data makes sense. And this is where it is interesting because some authors suggest that enabling the data to speak for itself (please note that data do not literally speak!) and to therefore let understanding and meaning emerge from the data, but it is clear that there is an interpretation process happening. We as researchers interpret what we are observing in the data and attach to chunks of data what actions and events we believe are occurring within that data segment. Question here therefore is what is the relationship between truth and meaning? Is meaning objective and already exist within the “voice” of the data? Or, do we define meaning and apply it to what we perceive or interpret to be happening within the data? There are techniques within grounded theory such as theoretical sampling and constant comparisons that provide some answers to these questions but to what extent is truth realised by just grounded theory alone? Can ultimate truth really be attained?
What is the purpose of the qualitative strand within a mixed methods approach? From what I have been rereading, mixed methods can be used to build and test a theory, theoretical constructs, relationships and hypotheses. Their development occurs in the qualitative strand and then tested in the quantitative strand, and therefore adding an extra dimension of richness, integrity, authenticity, verifiability and validity to the research design.
A question I am working on at the moment given that Grounded Theory is part of the qualitative strand is to what extent do I use grounded theory? I have now more or less worked out the initial phase of qualitative data analysis, and this initial phase shall consist of Open Coding also known as Initial Coding. I think this is more or less a definite because it is through Open Coding or Initial Coding that meaningful data segments are labelled with suitable codes that describe what is happening; where a technique known as constant comparison is used to identify similarities, differences and variances, and where (in broad terms) these similarities, differences and variances contribute towards categorical development. I have recoded my data a few times and so far I have lots of codes, and some initial categories developing but shall now have to recode the data since developing new ideas about the research design and about the way I want to explore the phenomenon of interest. And also because I understand the data more now. This leads me to an interesting thought: not only do our theoretical understanding of what is occurring in the data develops over time along with the need for particular research design elements (assuming emergent research design), but also understanding of the data itself emerges from the way that we perceive and interpret what is going on. There is an interesting relationship going on here between our own perceptions and interpretations, the development of these perceptions, and the data itself. What role does the data play in this relationship?
I can begin to observe what I had not previously observed and I can understand the grounded theory techniques better than before. I have started to draw out the steps and phases of the new research design with the current focus on the qualitative strand. I understand more now about categorical development and have outlined more questions I want to ask about the data as I proceed with recoding the data and continue to develop categories.
Aligned with my philosophical beliefs, I believe that there is a truth out there behind the process of the phenomenon of investigation but whether or not this real truth can occur only from coding and recoding for the context of my research is doubtful. But a mixed methods design perhaps could lead me closer to that ontological truth without actually reaching absolute truth. Aligned with my epistemological beliefs, the logical process (abductive) that underlies my use of grounded theory (develop hypotheses inductively from the data and use deductive methods to test the hypotheses against the data) aligns with my beliefs that knowledge is not certain and absolute. We need to continuously think about the data, think about what is happening in the data, think about how we interpret the data and how we know what we know to be true or perceive to be truth (meta-Philosophy) as long as everything is grounded in the data. All hypotheses, ideas, observations, and thoughts must be grounded in the data. We need to question our own biases and acknowledge them. All this while we maintain our sanity long enough to do so!
A big question that I have next is: when I have all the codes, and have developed all the categories and identified relationships between each category and the relevant properties and dimensions, what then? Grounded theorists talk about bringing everything together to form a theory whilst other grounded theorists discuss the idea of linking categories together to identify relationships in a process known as Axial Coding. I think I am currently leaning towards axial coding or some sort of coding technique that enables me to relate categories, because it is through the relation of categories and really understanding the way that categories interact with each other could I then begin to understand the way that the particular learning phenomenon of interest progresses from start to conclusion. This is challenging and whilst I shall try to work it all out for the sake of the diagrams I am drawing out as plans, the only way I think I am going to know for sure what I shall do is to simply do the coding. But the way I am viewing this at the moment is whilst the categories in themselves explain what is happening with certain parts of the phenomenon, by themselves they do not explain the process. There needs to be that extra step that identifies the process and the relationships therefore between elements of this process in order to better explain the phenomenon.
Once I have developed the ideas of the way I am going to approach the qualitative strand I shall then deal with the quantitative strand, fit everything within a mixed methods scenario if proven to be the most appropriate strategy, as well as a case study methodology if necessary, and then actually test my ideas against the data and remodify accordingly after receiving feedback from the supervisor.
‘till next time!
April 04, 2018
Mixed Methods Research Design: Should This Now Be Considered Suitable?
This past weekend has encouraged me to re-evaluate and re-explore the value of using both quantitative and qualitative data within my research project. This is an ongoing task that demands careful and reflective thought, and currently constructing diagrams that illustrate aspects of the design and the way in which these different aspects relate to each other, and the way in which the research shall now progress. Once I have completed these diagrams I shall be sending them to my supervisor for further feedback and confirmation of the design’s suitability. There are, not surprisingly, many thoughts, questions and ideas that I have about the emerging research design. As mentioned, going through all these thoughts, questions and ideas is an ongoing process but there are some key questions and ideas that I am focussing on at the moment with regards to the characteristics and aspects of the research design.
Firstly, and probably most importantly, should I reemploy a mixed methods approach? Is a mixed methods approach actually possible given the data collection context? Instead of collecting qualitative and quantitative data separately as is typically found in most mixed methods research, I have collected qualitative data and from this data set, both qualitative and quantitative data analysis shall be applied. I have a vision about what qualitative and quantitative data I want, but I am working through how this is going to be precisely and exactly realised particularly the quantitative aspect. I realise therefore I am not using mixed methods at the data collection level, but there appears to be a mixed approach at the data analysis methods level. This has implications at the methodology level: should mixed methods be confirmed as the appropriate approach to the research, then grounded theory becomes the qualitative method and not a methodology, whilst network analysis or some form of it becomes the quantitative method.
But here’s something to think about, and forms my second current thought and question: what comes between grounded theory and network analysis? What acts as the bridge that enables qualitative data to cross over into the quantitative realm? I think the answer lies in visualisation. In my understanding, a network is a visual representation or diagram of what is happening. A phenomenon can be understood through its aspects, features, events or activities and these can be represented as a network of nodes and connections. What I am attempting to do here is convert the concepts, categories and their relationships, products of grounded theory analysis, into a network. I am slowly working through how these grounded theory concepts can be converted into aspects of a network and this is going to take some time, but currently I am thinking that concepts and categories can be represented by nodes, and the relationships between categories can be represented by connections between nodes. What I am also interested in is exploring the relationships between these nodes because it is at these points where interesting observations and values can be obtained, but I’ve yet to figure out the way this can be fully considered. I’m thinking at the moment these relationships shall be related to the hypotheses that shall be developed as well as the properties and dimensions of categories and might also might be involved with quantitative analysis. The quantitative analysis shall be used to analyse these relationships to determine the strength between different types of nodes within different contexts, but the exact relationships and hypotheses that are to be explored are undetermined at this time and shall be until the qualitative data analysis section has been completed. This in a sense brings me to a third concern I am working on.
If my research is to adopt a mixed methods methodology again, what type of mixed methods should it be? My previous approach to mixed methods was a sequential exploratory type where qualitative data were to be collected and analysed first followed by the collection and analysis of quantitative data. This was therefore sequential in nature but I am not sure at this time whether my mixed methods approach now would be sequential or transformative: sequential because qualitative analysis will come after qualitative analysis, or transformative because it might be that some aspects of the qualitative data might be transformed into quantitative data. Is this even possible? It is in some context but I’m not sure if my qualitative data will be able to transform into quantitative and I am probably unable to know this till the qualitative analysis phase is complete and I begin to really look at the findings. At a push at the moment I'd say sequential exploratory: might be best to design both types just in case!
There are many other concerns that I now have that I shall be exploring further as my thinking and experimenting of the potential mixed methods approach progresses: in what way should I now present my research questions? The research questions shall have to change to better represent a potential mixed methods approach as the questions cannot be purely qualitative: a question must be qualitative and another must be quantitative but derived from an overarching question that brings both together.
Also, what are the implications on the use of literature and the roles of the literature reviews? At the moment I cannot imagine there being too many changes because of the important role that grounded theory shall continue to play in terms of identifying the nodes and connections of a network, which shall subsequently have some form or forms of quantitative analysis placed onto it (is this really network analysis, or something else?) although I shall have to double check the role of literature within mixed methods research.
What about the product of or the outcomes of the research? What is the nature of theoretical development within mixed methods research? A key role of mixed methods as described in some of the methodological literature is to both build and test a theory and / or a set of hypotheses. The qualitative aspect builds theoretical constructs and hypotheses and the quantitative strand tests these theoretical constructs and hypotheses.
What shall be or should be the extent to which grounded theory is used? Should I use grounded theory to the extent that a general theme of the learning phenomenon can be established and use that as the basis of the network construction and exploration? Or, should I use grounded theory to the extent that categories, relationships and hypotheses can emerge from the data, but use an existing overarching theoretical framework to guide their use in the network construction, and use quantitative analysis to test the identified relationships and hypotheses that come from the qualitative stage? I am not sure at this time.
What about the case study methodology? Should I return to thinking about the value of a case study methodology with mixed methods approach encased within? There is some debate about whether or not a mixed methods approach really is a methodology and not just a strategy of the way in which methods are to be sequenced or arranged. I shall have to revisit this debate area.
I have so many questions at this time, so many more than answers but I have a plan to work through all these different questions and issues that I have discussed here and more besides. I shall probably be writing on here on a regular basis now if only to document this challenging yet exciting journey and therefore to help me reflect upon my ideas and their development.
Thanks for reading! If you’re on your Easter holidays still, continue to have fun!
Research Design: A Reflection and Possible New Direction
Not only have I been stuffing my face full of Easter goodness (hot crossed buns and Easter eggs) but given that the newly added methods to my research design have been confirmed and accepted as being appropriate as a result of coming to know the phenomenon of interest in a way I had not previously considered, I have been rethinking the structure and process of my research design. This is particularly since this past weekend where I had the revelation that perhaps I should return to and re-evaluate the value, worth, role and purpose of combining qualitative and quantitative data within my project. A current task is therefore to think very diligently, carefully, strategically, and comprehensively about how qualitative and quantitative methods can analyse the data, and how qualitative and quantitative data can be combined or utilised in a way that can comprehensively describe and explain the phenomenon of interest unachievable by a single approach.
An Emergent Research Design?
What has struck me recently is that my research design can be characterised as emergent. The newly added methods and the possible re-evaluation of the methodological approach has emerged from further understanding of the data, further understanding of existing literature, and further understanding of the different types, structures, processes and outcomes of the phenomenon of interest. Further, these sources appear to triangulate to provide some sort of justification for what has emerged e.g., what I have observed in the data and the need to explore these observations further can be backed by existing literature, and both give rise to the need of the additional data analysis methods and perhaps a rethink of the methodology and research questions. This idea of an emergent research design appears to be a characteristic not just of grounded theory but qualitative research design more generally.
Essentially and I shall be writing more about this in the future, the research design emerges as the data analysis progresses with further readings as necessary to support the need for any emergent research design aspect. Where I am now with the research design and the inclusion of network analysis as a method has come from what I have observed in the data. In other words, the need for such a method has emerged from understanding the data, from observing particular patterns and trends, thinking carefully about the way these trends and patterns could be explored more comprehensively, and the potential value and worth their explorations might offer to the research.
Let’s take a brief journey in time to reflect on where I have been with the research design
The Journey of the Research Design so far
The Ph.D. research began prior to the Upgrade process as a mixed methods project, where mixed methods approach was introduced at the data collection level where the idea was to collect qualitative data from observations of the learning phenomenon and quantitative data from surveys. After a series of doubts started to creep in following the submission of the original Upgrade paper about the data collection methods and the context of the quantitative data collection and analysis aspect, and after discussions with the Upgrade member panel and the supervisor, the approach was dropped. The qualitative aspect was kept and therefore, grounded theory became the sole focus of the research design. Grounded theory became the methodology and its coding package became the methods of data analysis.
For many months after I began to downplay the relevance of mixed methods approach in my research and began to focus exclusively on learning about Grounded Theory and the way that I can utilise Grounded Theory within my research context, which again has been documented extensively throughout the previous year. I also began to realise and became aware of the complexity of my philosophical beliefs both at the ontological and epistemological levels though had not travelled down to the methodological and methods level because of my continued denial of the value of a mixed approach to understanding the phenomenon of interest. I did, however, later in the year and earlier this year seriously began to challenge the theoretical orientation of grounded theory and began to really believe that symbolic interactionism (the most common theoretical framework of grounded theory) was not compatible with the research context and began to search for other possible frameworks. Again this has been documented in earlier blog posts. I also began, through reading through more existing literature and the draft writing of earlier thesis chapters, to challenge my own understanding of the phenomenon of interest: the way I perceived it, the way I approached its exploration, and the way I could define it.
This led then to me challenging the way I had used grounded theory previously to analyse the data and I came across a startling thought: grounded theory could be used to recognise a central theme of the phenomenon of interest and theorise about the phenomenon around this theme, but I began to doubt grounded theory’s ability to theorise or hypothesise about the progress and process of the phenomenon of interest over a period of time. It was not, so I came to eventually realise, the central theme of the learning phenomenon that was the only product of the research that is of interest to me: it’s the way in which the learning phenomenon initiates and is sustained over a period of time. This I think is an area that is not addressed by grounded theory.
Where am I now with the Research Design?
Grounded Theory is still of interest and of importance to the research in terms of, from what I can currently understand, identifying a central theme to the phenomenon of interest, and to theorise about the phenomenon in accordance with this key theme. However, in what way do I explore the progress of the phenomenon of interest and the way in which this learning process can be sustained over time? This is where network analysis comes into play. But here is something else: I have always created diagrams and “networks,” if you will, about what is occurring in the data in order to help me understand what is going on in the data but I had not considered these diagrams as being somewhat of an independent data analysis method in their own right as I always thought of them as part of the grounded theory. But as I drew out more of these diagrams I began to realise that I was making observations and identifying trends that perhaps grounded theory on its own might not be able to explore to a substantial extent. At least, not to the extent that I am now interested in.
More significantly, I’ve very recently began to think about the way in which I could use these diagrams to further explore the phenomenon of interest through network analysis and the inclusion of quantitative analysis to test hypotheses and theoretical constructs that have and shall continue to emerge through grounded theory analysis. And therefore, a reintroduction of an old idea: the mixed methods approach!
And that shall be the topic of the next blog post!
September 25, 2016
All Case Study Aspects Dropped: the case, pun intended, for this
Ok, after much deliberation I have decided to drop all case study elements from my research. Debates and discussions regarding the inclusion of case study elements within a mixed methods context utalising grounded theory vary widely. But for me the case study approach goes against the nature of the research and the research intentions even if it were used for only framing the research questions and the data collection and analysis procedures, because quite frankly relevant approaches from critical realism, mixed methods and grounded theory appear to encompass all that is required, making the case-based study or strategy rather redundant. Despite numerous reasons for being initially attracted to case study aspects e.g., exploration of phenomenon in its natural setting and the carrying out of an intensive and detailed study on a phenomenon, the five key deciders for dropping any mention of a case study are discussed. Note that what is discussed has come not from actually carrying out case study research, but from logic and reason based on my current understanding.
Intentions of my research are to develop theory from the grounded theory data within the quantitative strand, test the theory using the quantitative strand, and then use the quantitative data to refine the theory. Case study emphasises not the development of theory, according to key authors Eisenhardt and Yin, but the testing of a theoretical framework, either existing or developed through the analysis of literature, before commencing any case study research.
The emphasis on the theory development through literature and prior to carrying out the research is incompatible with grounded theory, which suggests that an existing theoretical framework should not be forced onto the data but emerge and develop from the data. I accept however that there are theses out there that have not developed a theory prior to carrying out a mixed methods case study, but for me and the intentions of my research that approach would not work.
Additionally, I am unsure of case study’s stance on theory refinement. Plentiful literature describes it as an effective strategy or methodology, depending on the way it is used, for theory development, but nothing on actual theory refinement.
From my understanding, everything needs to be designed, developed and explored relevant to the case or a series of cases. Whilst case study research employs a form of purposive sampling of cases, it appears to me that all participants of a particular case must be included in the research with no "outsiders". Whilst this is fine if that is the intention, the mixed methods approach being developed for my research requires different population samples from outside of the cases that shall be explored, and this does not appear to fit within the use of sampling for a case study. If I were to use a sequential explanatory mixed methods this would not be an issue, but because I am developing a sequential exploratory, an amended version, this would be a problem. A problem would involve the fact that the theory would be tested on a population sample different to the participants of the cases explored through grounded theory. The fact the samples shall differ between qualitative and quantitative makes case study incompatible.
Case study actively encourages triangulation of research findings, meaning that the findings come from different research methods for a variety of purposes including corroborating data and improving research validity. The concurrent triangulation variation of mixed methods was going to be used until it was realised that this would have led to difficulties in the research design and therefore render it unreliable, therefore it was switched to a sequential exploratory approach. Concurrent triangulation would have achieved the triangulation objective of the case study approach, but the sequential exploratory does not: at least, not the in the way it is being used in this research to develop a theory.
Replication logic is what gives case study a mode of generalisability or in other words the ability to generalise identified events and activities across a series of cases. Replication appears on two levels: literal replication if few cases are explored and theoretical replication if several are selected. The former is used for predicting similar results across cases whilst the latter is used for predicting contrasting results across cases but for reasons that can be anticipated. Yin’s description of replication logic is akin to experimental designs: the focus is on replicating findings in some way, and therefore highlights a positivist approach to research, which would in my opinion oppose the general philosophical stance of grounded theory. Grounded theory is a mode of interpreting data and is therefore not a mode of enforcing a particular theoretical framework upon data in order to find some sort of replication. There is a form of replication that can be found within grounded theory, but this does not come from an enforcement of a theoretical perspective but is allowed to emerge naturally from the data relative to the perspectives and interpretations of the researcher.
Therefore, replication logic appears to be based on replication based on pre-existing theoretical frameworks and assumptions. This is unlikely to work in my research.
No, no, I am not going to say that the sequential exploratory mixed methods using grounded theory and questionnaire (more than likely: depends on the findings of the grounded theory) underpinned by critical realism shall be the research design because I might change my mind, but it’s not likely though I have said that before! But that’s the beauty of research: you can never really be certain or absolute of anything.
All of my latest ideas about the research design is to be confirmed as appropriate by the supervisor.
Those of you interested:
Robert Yin’s book on case study methods: Case Study Research: Design and methods. The fourth edition is available on Google books, and all University libraries! Though a bit difficult to get hold of from a University library if you are not a registered student or researcher at that University……
Plus, Kathleen Eisenhardt’s research paper Building Theories From Case Study Research available from The Academy of Management Review journal.
Plus, before any person comments, I realise that is not the formal way to reference materials! Have to adore Harvard referencing………
Case study? Case–based strategy? Not a case–based strategy? What is it then?
Remember the time I posted up a post that began with remember the time when I said that my research design is complete? I’m saying it again: remember the time when I said that my research design is complete? Well, earlier in the day a thought literally struck my ideas of a case study down, stamped all over them, and performed some sort of war dance over them chanting in some intelligible language. So, I will not dance about and sing claiming that I have found my research design because, given that I am a critical realist, to claim that I have found an absolute research design would be complete and utter nonsense.
Regardless, the intention was never to implement a full case study design as the research uses a particular variant of the mixed methods methodology. Whilst mixed methods can be used within a case study design, I have come to realise that the type of mixed methods that I am proposing (an amended version of Creswell’s sequential exploratory to reflect better the theory refinement phase) and the types of methods being used is making me question the use and role of a case study approach.
So I began thinking about the idea of calling case implementation a case-based strategy for framing the research questions and acting as a general guide for guiding the mixed methods data collection and analysis sequences, but it was realised that this didn’t really make much sense because it would only apply to the qualitative aspect and not the quantitative aspect, and also from what I understand all phases of a sequential mixed methods approach would have to study the same participants because it is about studying a single case e.g., a group of people, an organisation and so on. Sequential exploratory calls for different population samples, though sharing the same basic characteristics, for both qualitative and quantitative strands.
A key question that has been playing on my mind is, is it really the right way to call something a case study or even a case study strategy if some of the key aspects of such an approach are not going to be utalised? Can you really mould and combine bits of different methodologies and methods and call it by a particular name e.g., is it really right to call my study an exploratory mixed methods case study on X phenomenon if most of the key functions of a case study shall not be used?
If not, then what can I call it? What does it all even mean?
Could my research be case-based? There are clear examples of cases existing, with a case being defined as simply an instance of a phenomenon of interest e.g., I am exploring a particular process of learning so a case would simply be defined as a separate but related instance of this phenomenon. In terms of Education think of a “case” as a classroom, or an individual student, or a group of learners.
A discussion on the academic site ResearcherGate included a comment suggesting that within a sequential mixed methods design, a case study can be used as part of the qualitative component but because my research design has been amended slightly to reflect theory refinement, it kind of goes beyond the limit of what a case study is supposed to achieve, in my opinion from what I can currently understand. Another debate revolves around if whether or not aspects of case study can really be used within a grounded theory exploration, and opinions on this appear to vary widely with some supporters of this idea whilst former Ph.D. candidates stated that their examiners were critical of this approach. An interesting comment suggests that if an approach uses an existing theoretical framework upon the data then it is difficult to call this grounded theory, because all versions of grounded theory does not work strictly within a theoretical framework. The commentator suggests that if a project contains more case study principles than grounded theory principles then it would be incorrect to call it a grounded theory study, which makes sense. If on the other hand, as another commentator alluded, and as my research appears to be leaning towards, the project contains more grounded theory elements than case study elements then it would be wrong to call that something like a case study grounded theory project, because it would not be adhering to case study principles.
So perhaps I have to think of my research in terms of it being case-based, and not actually call it a case study or even a case strategy. Interestingly I have just read a research paper that states that studying cases is not exclusive to case study research, but can be present in all approaches that adhere to qualitative assumptions and this includes grounded theory.
Ultimately, is there really a need for any reference to case study or case strategy if all what I need in reference to studying instances (cases?) qualitatively resides within a grounded theory approach?
Something to think about!
September 20, 2016
Current Reflections On Mixed Methods Methodology
Right at the beginning of the Ph.D. all researchers are presented with an introduction to various research philosophies and methodologies, and the many ways in which these have been defined and applied. The research introduction is not extensive: it is designed to initiate your thinking about what philosophical perspectives, methodological approaches and methods might be suitable for your research. Usually it also introduces you to some of the general advantages and disadvantages of and some of the more general arguments for and against each approach. Basically, it acts as a platform upon which you jump off into a much wider arena of discussion, debate and application.
At this point, given that the state of the research is in the upgrade process, my experience of mixed methods has revolved around designing and developing mixed methods as a methodology guided by a case study strategy, which itself can be applied to a research design in many ways but this will be discussed in another blog post, underpinned by a critical realist Philosophy. Explaining the way all this fits together is beyond the purpose of this blog post, but specific to mixed methods and thinking about the way that mixed methods can work with my research design has been an extremely interesting adventure and continues to be so.
When I began thinking about mixed methods methodology I found various typologies that explain the way in which different components of a mixed methods methodology fit together, and their general functionality. Creswell is a prolific writer of mixed methods research, and has come up with various typologies of mixed methods designs: sequential, concurrent, embedded, and transformative. When I first began reading through the various types and their applications, I initially chose concurrent parallel variety but then changed to sequential exploratory. Read previous blog posts to read all about why the transition took place.
Initially, I thought these typologies were fixed with no flexibility with their amendments to fit particular contexts. I did struggle with this somewhat as the context calls for some amendments to take place, but had decided to put these amendment ideas aside for the time being. Through the trialling of grounded theory (to be discussed further another time) and through careful reading so far of relevant grounded theory and case study literature, I found that methodologies and methods are continuously shifting and amending based on current discussions of research methodology. Following this, literature was discovered from various authors criticising typologies from Creswell and other mixed methods authors, stating that these typologies are not meant to be used in an absolutely unchangeable way, but should be used as a guide and therefore amendable relative to the context.
Reading this has been somewhat of a relief. Since the research is based on theory development I wanted some process or element of the design to deal with theory refinement following the development and testing phases. I am now amending the sequential exploratory approach developed by Creswell to reflect an iterative, extra phase of theory refinement.
Reading methodologies, methods and even philosophies are continuously shifting concepts adds a new layer of complexity to understanding, designing, developing, testing and applying a research design. Not only are there actual philosophical perspectives and methodological approaches, and methods, but these are shifting positions all the time with researching continuously adding, for example, different philosophical arguments for or against different methodological approaches. Research can combine philosophies, methodologies and methods in various different ways relative to the research problem, the research questions, and the skills and experiences of the researcher. This would lead to not just simply applying a particularly defined approach to exploring reality, but also gives the researcher a chance to remedy criticisms and push the boundaries of methodological knowledge.
When a researcher begins to realise this, a new chapter begins. A new layer of understanding begins as you explore to the most detailed and in-depth level what research philosophies, methodologies and methods are about and to locate opportunities to extend and push current knowledge about these different philosophical and methodological approaches.
I have really only just begun to realise and push for developing new philosophical and methodological arguments and perspectives relative to the research problem and the phenomenon of interest. This is risky, I’m not sure if it will work, but I feel that what I am doing is right. Sometimes you just have to take that academic risk to push forward with what you want to achieve, because knowledge cannot push forward to new arenas if no person is willing to saddle up and ride the horse of knowledge! I’ll be talking about this adventure more in future blog posts.
‘till next time: keep playing with your designs!
August 21, 2016
Slightly Amended Research Design
Since the research methodology chapter of the thesis is going to be around ten thousand to fifteen thousand words, it’s impossible to really explain all aspects of the research design in a blog post of less than a thousand words. Firstly because I’ll probably bore people to the point they would rather listen to One Direction, and secondly because understanding the relationship between all the components is a continuous and ongoing task.
The mixed methods variety has changed and the research is now encapsulated within a case study approach. The research design is now a critical realist multiple-case based sequential exploratory mixed methods design. So, why the change? The change came about through a conversation with the supervisor and the progress of the trial study.
Sequential Exploratory Mixed Methods
The trial study, as indicated in the previous blog post, has been and still is a serious, reflective exercise of the research activity and the research design. The previous preferred mixed methods flavour was concurrent triangulation meaning that the questionnaire data collection and analysis and grounded theory analysis would have been carried out at the same time, or concurrently.
This proved to be problematic during the trial study as it became apparent that the emerging theory’s constructs would not have been testable or explored further using the quantitative instrument. Why is this? Because the concurrent nature would have meant the construction of the questionnaire occurring before quantitative and qualitative data collection and analysis, leaving no room for further amendments to it. This would have led to the research being susceptible to validity problems and potential gaps in the theory therefore the theory’s integrity, relevance, completeness and even generalisability would have been questioned. After returning to the research methodological literature (Ph.D. potentials reading this will come to know the iterative nature of the Ph.D.) I found that the sequential exploratory approach to be more suitable.
The sequential exploratory variety of mixed methods involves the qualitative phase occurring first therefore the grounded theory method shall be used to generate an emerging, developing theory from the data. The constructs that emerge from the data shall then be converted to a format suitable for development of a quantitative instrument (likely to be a questionnaire) and then the third phase shall be the quantitative phase where aspects of the theory shall be tested for generalisability across different contexts. This solves the identified problem therefore the quantitative instrument shall take into account constructs found in the literature along with constructs built from the grounded theory data, with in some cases both sets being used to support each other with regards to evidencing the need to explore certain constructs further.
A case study has been defined in so many ways and can be used in various ways such as either a strategy or as a methodology. A case study enables the researcher to explore a particular case, or multiple cases, in substantial detail hence it is suitable for mixed methods research, in order to generate significant understanding of the phenomenon of interest. It is the depth of detail the case study brings to the research that makes it an attractive option. It is not usually famed for allowing generalisability of data but this is determined by the type of case study that is being used. Multiple-case study using mixed methods enables more of a generalizable approach to be considered.
For this research, case study shall be used as a general strategy for guiding the quantitative and qualitative phases of the mixed methods and their respective methods, whilst the case study strategy is being guided by critical realism.
This research shall adopt the variety known as a multiple-case study, which means that multiple cases or instances of the similar activity shall be explored using grounded theory (hence, a relationship between case study and grounded theory) in order to increase the validity of the findings, with the selection of the cases being determined by the findings of the first case. The multiple-case study has been deemed as appropriate given the nature of grounded theory’s approach towards sampling of data, and given the depth of detail that case study allows.
The blog post doesn’t do the research design justice, as for example I have only lightly touched on the relationship between critical realism, case study, mixed methods, and grounded theory. Understanding this relationship between the different components is ongoing and progressing and will be explained more in the upgrade paper and fully elaborated in the thesis. But it is enough on here to say that I am convinced that these amendments to the research are correct and the way that the design should be. But I must not have the mindset where I am absolutely sure that the research design is correct as it has to pass the upgrade assessment panel later this year.
It’s been challenging working towards the research design, but it’s been a rewarding experience. A key activity at this time is to further understand the relationship between critical realism, case study, mixed methods and grounded theory, and that is no easy feat.
So, the research design shall not change! Till next time, that is!
August 20, 2016
Been quite a while since I have written a blog post on here as there have been plenty of things going on to distract me, but now is time for a series of blog updates! The following is a brief overview of the key areas of activity, with further blog posts following that explore these in more detail.
Update on the Trial study
The trial study is based on trialling a collaborative activity and trialling a grounded theory approach adapted from the ideas of Strauss and Corbin. I say adapted because there are a variety of authors suggesting Strauss and Corbin’s version is too regimented and difficult to follow, but Strauss and Corbin did respond by suggesting their procedures for coding do not have to be followed through strictly, therefore researchers should logically and constructively apply grounded theory procedures relevant to the research context. I might alter aspects of the approach as the trail and the main thesis study continues.
The trial appears to be coming along well. As reported in earlier posts I managed to successfully carry out the trail activity so it is now the case of continuing to use grounded theory to learn about it more, and to learn about the data that I have collected. I am fascinated with grounded theory! Lots of information on grounded theory online but for now it suffices to say that grounded theory is a key methodology (or method: it can be viewed both ways depending on its purpose and position within the research design) in developing new conceptualisations, a new theory, or new theorising of the phenomenon of interest. This is challenging and I have the belief that in order to really be successful at this you have to be quite creative and able to think about data, events, observations and aspects of reality at an abstract level. You begin with exploring the data but you have to think about the data at a higher, abstract level. It’s about interpreting the data based on your observations, though the validity of these interpretations can be increased depending on the research design and the way in which existing literature is used. But despite it being challenging it is an exciting approach to exploring data.
Trial Study as the Thesis Study
The most significant feedback and conversations during the past month or so has been the idea of using the trail study as a thesis study, or acting as a case within the thesis study. This was and still is huge, because it effectively means that when I am exploring and coding the data not only am I exploring data to benefit the production of the upgrade paper but also to benefit the thesis and the research itself. This has been deemed feasible therefore I shall be using the activity as part of the thesis and also explore other activities, known in the thesis as cases, in order to locate differences and similarities, which shall increase the validity of the theorising and theory development.
Change of Research Design
This change has actually come about through working through the trial study. Previously I had mentioned about adopting a critical realist, concurrent triangulation variety of mixed methods where the quantitative and qualitative aspects of the study would be carried out at the same time. In other words, I would have been carrying out grounded theory analysis of the activity whilst collecting thoughts and beliefs of the learners at the same time and then integrate the findings and discussions.
The only aspect that has changed is the mixed methods variety: I’ve ditched the concurrent triangulation and adopted the sequential, exploratory variety. Briefly, during the trail study I was struggling with the way in which data from the grounded theory findings and questionnaire findings could merge due to the abstractness of grounded theory. There would have been a chance of new concepts and new conceptualisations emerging from the data but exploring them via quantitative measures would not have been possible because the quantitative phase would have started at the same time, or in other words concurrently.
Using a sequential exploratory approach enables the constructions and conceptualisations, the theory, to emerge first and then test these out using the questionnaire and generalise across a wider population. This makes a lot more sense than to carry out both threads concurrently given the potential for extra insights that would not have been possible to explore had the mixed methods approach remained unchanged.
I’ve also encapsulated the research into a multiple case study, therefore the research design is now a critical realist multiple case based sequential exploratory mixed methods design.
Progress on the Upgrade Paper
This has been rewritten more times than I care to remember and will be reread and edited in various places before actually handing it in. When I rewrite each section I ask myself why did I not think about things in a particular way and communicate in a particular way before. But that is the beauty of re editing: it makes you think and contemplate more as you go through the document and think if you can explain anything in a more concise, or more sophisticated way without dropping the meaning behind the text. I know this, but still, sometimes it takes me by surprise at the differences that are found between each modified version of a particular document.
Most of the sections are now complete. Only sections that need to be completed are the trial study section, and talk a bit more about potential problems and possible outcomes.
That’s the brief overview of the main things that have been going on during the past month or so!
March 18, 2016
Mixed Methods: Critical Realism Considered Most Suitable!
Critical realism deals with ontology! Yes!
Critical realism has been developed by the British Philosopher Ray Bhaskar as a result of combining separate philosophies: transcendental realism, which is a philosophy of science, and critical naturalism, which is a philosophy of the social sciences. It is not the aim of this blog post to explain either of them. Critical realism does not assume reality to be a single, observable, measurable, determinable layer whose actions and events are independent of the mind nor a single layer that is understandable through exploring experiences and perspectives. Critical realism assumes reality to have multiple layers containing structures and mechanisms that influence the observable and what can be experienced. It is the exploration of these structures and mechanism that provide the basis for exploration of reality using critical realism.
Unlike pragmatism, which is considered to be the most adopted philosophical perspective of mixed methods, critical realism contains ontological assumptions which are spread across three domains: the empirical, the actual, and the real. The empirical domain refers to aspects of reality that exists and can be observed or experienced directly or indirectly, the actual refers to aspects of reality that exists but might not be observed or experienced in some way, and the real refers to the structures and mechanisms that causes or influences what is observed or experienced. These structures and mechanisms are beyond the realm of human observation and experiences; they cannot be detected, known, or perceived, but can be, as defined by McVoy and Richards (2006), inferred through a research design consisting of both deductive (empirical investigation) and inductive (theory construction) processes. Where critical realism differs from all the other middle ground philosophies therefore, and what acts as the central reasoning for adoption in this mixed methods research, is that it places a focus on further understanding and explanations of these structures and mechanisms.
Relating Critical Realism To Research Context
Critical realism is a complicated middle ground philosophy probably the most complex of them all along with complexity theory, but it is a middle ground philosophy that makes the most sense for my research and for the aims of the research. The context of the Ph.D. research is not to explore research phenomena using only quantitative or qualitative methodologies; the problem area identified and developed does not assume that answers can be found in a single methodology or a single philosophical perspective such as absolutism or relativism. The problem area assumes that answers can be found through an integrated approach that involves both quantitative and qualitative approaches. So, with that, and with critical realism addressing the ontological level, it can be assumed that critical realism goes beyond the research question and places the research problem as central to the research project. It assumes, it can be proposed, that it is the identified problem area that can lead to the development of philosophical assumptions about reality, which then lead onto the development of research questions, which then lead onto the selection of the methodology and research approaches. A question here however is whether or not the philosophical perspective leads onto the development of the research question sequentially, or if the research questions and philosophical perspectives are identified and developed concurrently. That is something to be thought about and perhaps discussed another time.
Summary Of Thoughts Regarding Critical Realism
What has been discussed, briefly, is what makes critical realism distinctive and more suitable for my research than other middle ground philosophies. Post positivism focusses too much on the quantitative at the methodological level whilst pragmatism focusses too much on changes that are made at the practical level. Critical realism suggests that both quantitative and qualitative approaches are important to use in a single research project in order to fully explore and understand the structures and mechanisms of what can be observed and experienced.
There is much more to learn and understand about critical realism: its concepts, its use, its history, and the way in which critical realism can be fully integrated into a mixed methods research and the specific context of my research. Reading shall be continuous, but at the moment I am just pleased that I have been able to identify the most appropriate middle ground philosophy and start to fit the whole design around the principles of critical realism.
Fun stuff! The Upgrade Paper shall be used to introduce critical realism and the way in which its concepts have been applied throughout the research design whilst a full elaboration of critical realism including its application and possible solutions to problems of critical realism shall be provided in the thesis.
McEvoy, P., Richards, D (2009): A critical realist rationale for using a combination of quantitative and qualitative method, Journal Of Research In Nursing, 11 (1), 66 – 78