All 4 entries tagged Casestudy

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!

Alex Darracott : 04 Apr 2018 21:34 |  Tags: Casestudy Design Groundedtheory Methodology Methods Mixedmethods Qualitative Quantitative |  Comments (0) |  Close comments |  Report a problem

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.

Theoretical incompatibility

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.

Sampling Incompatibility

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.



Non-Triangulated research

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

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.

Well then!

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.

References

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………

Alex Darracott : 25 Sep 2016 20:05 |  Tags: Casestudy Design Groundedtheory Methodology Mixedmethods |  Comments (0) |  Close comments |  Report a problem

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?

Case-based?

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!

Alex Darracott : 25 Sep 2016 14:45 |  Tags: Casestudy Design Groundedtheory Mixedmethods |  Comments (0) |  Close comments |  Report a problem

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.

Case Study

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.

Summary thoughts

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!

Alex Darracott : 21 Aug 2016 20:40 |  Tags: Casestudy Design Groundedtheory Mixedmethods |  Comments (0) |  Close comments |  Report a problem