All 3 entries tagged Open Coding
No other Warwick Blogs use the tag Open Coding on entries | View entries tagged Open Coding at Technorati | There are no images tagged Open Coding on this blog
April 08, 2018
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!
December 14, 2017
The three key tasks that I have been engaged with as part of the early stage of grounded theory application (besides continuously developing the relationship between my philosophical beliefs and grounded theory, as explained in the previous blog post) are engaging with the literature, using open or initial coding, and writing memos.
Engagement with Literature
I have adopted a two-fold approach. As can be imagined, there is much debate among all key authors regarding the role of literature and the position of the literature review within the thesis. This shall be discussed more during the upcoming blog series on Grounded Theory, but at the moment it suffices to state that Strauss and Corbin along with Charmaz and Bryant provide the foundations upon which my approaches have been constructed.
The first function of the literature is to provide the grounding, critiques, focus, background and basis of the investigation. I am situating my ideas and approaches within existing literature, which is establishing the social and economic justifications of the need for my research and the philosophical, pedagogical and methodological need for further explorations of the phenomena of interest. As has been discussed in a previous blog post, these discussions, justifications, debates, critiques, analysis etc. of literature are being constructed across three related literature review chapters placed at the beginning of the thesis.
The second function of the literature is to act as a means of validating and verifying the emerging theory. Specifically, the theory’s content (categories, dimensions, properties, and relationships between categories, dimensions and properties) can not only be verified and validated by continuous comparisons with the data, but also comparisons with findings of existing literature relevant to the substantive area.
I am aiming to integrate the literature findings within sections of the discussion chapter related to each category and the constructs of each category, and the relationships that exist between categories and constructs. This way, I can classify literature in relation to each of the theory’s category although bearing in mind that some literature might refer to more than a single category. Either way, I think integrating findings within the discussion chapter within the relevant category section is the common sense way to manage extant literature.
Open or Initial Coding
My approach to open coding appears to have a strong pragmatic base and therefore, appears to be influenced more by Bryant’s pragmatism grounded theory than Charmaz’s constructivist grounded theory, Glaser’s apparent positivist grounded theory and Strauss’ symbolic interactionist grounded theory. Again there is much debate about the way in which the researcher should work with the data, and shall be a topic I shall cover in next year’s blog series on grounded theory.
There are, therefore, various approaches that can be used to analyse the data using grounded theory, but I have adopted what I shall term segment by segment analysis. What determines an individual segment is the meaning that it holds, or the meaning that I perceive it to hold. A data segment can be a whole line, a sentence, or half a line or half a sentence. A segment therefore holds some sort of complete meaning, which could be an idea, a stand-alone part of an argument, a justification, a reason, or anything that I can identify as having some meaningful existence. This meaning or meaningful existence must be naturally occurring (if you want, you could call it a natural relationship between researcher and the data), and therefore a criticism I have of line by line and sentence by sentence analysis is the possibility of forcing meaning and therefore identify concepts that do not actually exist. Segment by segment analysis enables you, in my opinion, to openly detect what the data is saying. If you find or perceive meaning within a whole sentence, then give that sentence a label that represents that meaning. If you find half a sentence represents a full meaning, then give that half a sentence a label. I used these labels, therefore, to describe or summarise the meaning that I had perceived to exist within that segment. As I was coding the data, I became consciously aware of similar data segments. These similar data segments share similar characteristics or roles within the context of their existence. Similar data segments were compared with each other, and if considered similar enough they were coded with the same label. What was occurring therefore was a series of data segments that shared the same characteristics, but were presenting themselves differently (e.g., negative and positive reasons).
Multiple similar data segments and related codes shall eventually combine to create abstract classes or groupings known as categories, which enable theoretical development. What is important to remember here from my current understanding is that codes derive from our interpretation of what is happening in the data, and categories form from similar codes that can be grouped. Categories are considered abstract exactly because they are not derived from the data. Each category consists of properties and dimensions, and these are discovered through significant and comprehensive comparative analysis of coded data segments that belong to that category. It is the variation among all data segments that contribute towards a category’s properties and dimensions, and it is this variation that shall be playing a major part of theoretical sampling when I develop criteria for the next set of data to collect in the new year. Some category development has taken place, in fact more development that I had expected, but there is much more to be done. The groundwork has been laid for further development through the writing of memos, which shall be described a bit later.
In thinking about the segment by segment approach, it appears that I have adopted the wider analytical approach known as micro analysis. My current understanding tells me that, at least within sociological terms, macro level analysis addresses wider arguably post-modern issues such as social stability, social injustices, social change and power authorities. Micro analysis explores deeply and comprehensively social and collaborative structures such as learning, interaction, group construction and group dynamics. The logical approach of this grounded theory research is abductive reasoning, which has been strongly advocated by Bryant’s Pragmatist views of grounded theory.
Briefly, this form of logic uses a mixture of deduction and induction where concepts and codes are inductively derived from the data, and deductively derived from any hypotheses and further abstract interpretations and predictions tested against further collected data. There is more to abductive reasoning than this rather basic definition, but that’s beyond the purpose of this blog post. I shall write more about abductive reasoning in the blog post series of Grounded Theory, and as I progress through the grounded theory coding stages.
Writing of Memos
Memos have been advocated by all grounded theory authors as a significant part of the grounded theory process and of theoretical development. In general, memos capture your thoughts, hypotheses, ideas, predictions, conceptions, perspectives, reasons, explanations and anything else related to the codes that you use and the categories you are developing, and the eventual wider theory. Memos, therefore, carry many functions, features, aims and objectives throughout a grounded theory process. Memos are therefore versatile and can be written in any grounded theory context, and therefore you are in control of the structure and content of memos, and why you need to write them.
I have been writing two different types of memos that are contributing towards theoretical development: phenomena memos, and code memos.
The phenomena memos relate to the ways in which learning takes place in relation to the specific phenomena of interest. These memos refer to my discussions, thoughts, observations, propositions, predictions, hypotheses and critiques of the events, happenings, consequences, conditions, situations, contexts and causes of the phenomena. Various phenomena memos have been written, with each memo exploring each instance of the phenomena, and where necessary and appropriate comparisons have been made between each memo in order to identify similarities and differences, which can lead to hypotheses development as part of abductive logic. Comparisons of memos can actually lead to more memos that document these comparisons.
The code memos are related to the created codes and categories. Many memos have been written, with each memo relating to each code. Each code memo explains each data segment associated with that particular code, with the explanation involving the segment’s context, function, purpose within the context, and its interpreted or observed meaning.
Each data segment is included within the memo in accordance with its uniqueness. E.g., if the data segment can be observed to be different or similar in some way, or consists of some characteristic or property that warrants explanation and discussion, or in any way adds to any existing explanation. Each relevant data segment is then included and explained, and where necessary compared and contrasted with other data segments within the memo. This is a method also of assisting with categorical development, and categories and categorical development also comes with their own sets of memos. Few categorical and categorical development memos have been written at this time however, but is set to progress forward significantly in the new year.
The first set of data has been coded, and now ready to further the development of categories. However, upon advice from the supervisor and from my schedule I’m taking a break from grounded theory but shall return with recharged batteries and reread every memo and the coded transcript in the new year!
November 27, 2017
Initial Stage of Grounded Theory Coding
Coding the data using the initial stage of the Grounded Theory process, known as Open Coding or Initial Coding, has progressed substantially since the previous update. In fact, I’ve actually completed the task of coding through the first set of data during the previous week, which I had not expected but has put me ahead of schedule!
Just as a brief reminder, Open Coding or Initial Coding refers to identifying concepts within the data and the use of codes that summaries or describes the meaning or characteristics of that particular data segment, and therefore identifies these concepts. You could say that coding gives data segments an identity that you can refer to time and time again as you progress through your coding, depending on the characteristics that you identify and interpret within each data segment. You are essentially making practical, empirical observations of the data, and interpreting that data to mean something that is of value or in some way contributes towards characterising the phenomenon of interest that you are exploring. Whilst all your codes and code-data segment matching is an interpretive process, it is also objective as all codes are grounded in the data, particularly with the process of comparisons between data segments for similarities of characteristics. I shall be talking more about this in my future short blog series of Grounded Theory from next week.
At the time of writing this blog post, I think I have about twenty or more different codes that I have used across the whole data set, and this is actually a reduction on the amount produced during previous coding sessions. What I am increasingly discovering within the grounded theory approach is the influential impact and role of context on my interpretations, and perhaps the way that I should be interpreting and coding the data, and identifying the appropriateness of code-data segment matching. What is assigned a particular type of code in one context would be coded as something completely different in another context. This appears to be the nature of exploring learning processes and phenomena using grounded theory: the understanding and acquirement of knowledge regarding the development and process of learning differs between contexts. With collaborative learning for example, the collaborative activities, processes and communication shifts and moulds what is happening within the data as time progresses, and can illuminate different patterns at different times depending on the context; depending on what is being dealt with at the time. It is simply not a case of observing a particular process and thinking that it’s always universally understood because learning processes and phenomena have a nuanced existence that is shaped and moulded by events, happenings, actions and others within collaborative situations.
Therefore, as a grounded theory researcher, when you are exploring learning phenomena, the context that envelopes or provides the basis for the learning process is able to mould and shape this learning process over time, yet grounded theory enables you to identify the nuanced existence and subtle differences between the characteristics of similar concepts. Beyond reading the textbooks on grounded theory, the biggest learning curve and learning experience of my application of grounded theory has been trying to understand the importance of context and the way in which this really impacts my interpretations and observations of the data. I’m still learning now. I’m still wondering and questioning if I have really coded everything correctly even though I have checked through things several times during the past week and have altered the coding where I feel necessary.
Along with coding, I’ve also been writing plenty of memos. Memos is a technique of grounded theory that helps you to build your theory by capturing all of your thoughts about the development of your codes, what you have observed, the similarities and differences that you find between coded segments, and the comparisons between different coded segments e.g., their characteristics and contrasts between similar and different concepts and what makes those data segments really what they are.
Additionally, all this information contributes valuable insights and input into your theoretical sensitivity and theoretical awareness of the data, as well as developing theoretical sampling. Theoretical sampling is a qualitative sampling method that determines what to sample next (e.g., what information or data you need next) based on the emerging theory: the observations and questions derived from the data and the codes all guiding and directing the next set of data to pick up and analyse. I shall be talking about this more either during the short blog series of grounded theory or at some point early next year.
Focussing on rewriting the memos shall be the focus of the rest of the week in an attempt to communicate my ideas more clearly, to tidy them up a bit, and to reduce their number and organise them into something that makes a bit more sense. The set of memo writing sessions just completed involved writing a memo page (in some cases several pages) per code, within which each data segment coded with that respective code was explained and compared to previous segments in order to identify and locate subtle differences between each segment, leading in some cases to identification of potential categories (which are basically a combination of various codes and provides the core of the theory) and categorical properties and dimensions.
Writing a memo per code worked fine for a while, and the potential categories identified so far are suitable although these obviously need to be re-examined continuously (shall talk more about categories next month) but what I have realised is I have been taking these data segments out of their context and trying to explain them as standalone entities. As I went deeper into the data I began to realise that data segments can be logically connected, therefore trying to explain them independently of each other was becoming an increasingly difficult task. I found myself referring to these logically connected data segments in order to provide a contextual explanation for the data segments and their difference between other similarly coded data segments.
What I shall do next is rewrite the memos and add more details about the context. Instead of writing about each data segment as stand alone entities, I shall now write about complete units of logically connected data segments. This way, I can break the unit down into constituent segments and attempt to explain them individually and then discuss their relationship to each other as part of that unit. Doing it this way, I think I can then explain the meaning of individual segments without losing its contextual meaning and relationship with other segments. And, I can compare data segment to data segment, and data unit (a series of logically connected data segments) to data unit. It makes sense, um, well, currently in theory……..
What’s The Aim Then?
At the conclusion of the week I aim to have a complete coded first set of data (shall be rechecking again), a full set of rewritten memos and an updated theoretical framework. This will then, as far as I am currently aware of, bring grounded theory work to a conclusion for the year. I shall send everything off to the supervisor for feedback and guidance, and up to the Christmas holiday I shall work on the first literature review chapter, and write the blog series on Grounded Theory!
Plenty to come; watch this space (or just read the blog!)