Sunday, December 29, 2013

Writing Chapter 4

Writing Chapter 4 is usually one of the most grueling and tiresome points of the dissertation process, but for once timing was on my side as I am enjoying the opportunity to write my first draft of Chapter 4 in warm and sunny Florida.  I have been looking forward to this opportunity for over a month when I began to feel prepared to write the chapter.

Since Thanksgiving weekend, I have not had the opportunity to write Chapter 4 because I was either busy preparing or enjoying holiday events, visits to or from family and friends, or writing member checks for my participants.  During my downtime in December, however, I was plotting and outlining Chapter 4 in my mind.  I never got the chance to sit down and write any of it until last week.

My family and I have only been in Florida for 6 days and yet I have already written 55 pages of Chapter 4.  Two of these 6 days don't count because one was the day of arrival and the other was Christmas day, so I really wrote 55 pages in 4 days.  Today marks the point in which I completed writing the major components of my grounded theory, so I am quite satisfied now.

Looking over those rough 55 pages, there is plenty of room to flesh out my weaker arguments and transitions.  I also have a lot of ideas of how to reorganize the content of those pages.  I believe in my second edit of these pages, I would probably remove as much as or a little less than I would add.

My goal is to hit 100 pages before we leave Florida, and I'm making good progress on that.  With the bulk of my argument already covered, I am worried less about quantity and worried more about the quality.  With the 55 pages covering the biggest cross-case patterns, I anticipate the next 10-15 pages will cover the significant single-case findings that directly answer my research questions.  So I plan on spending Monday and maybe Tuesday (the remainder of 2013) writing that section.

After writing the significant single-case findings, I plan to tie my grounded theory with at least two established models, one in the field of acculturation psychology and the other in the field of English language education.  I'm trying to imagine a graphic organizer to represent how I plan to combine my grounded theory with these two models.  For now, this imagined graphic organizer is too complicated and messy, although the main ideas seem clear.  I need to create a graphic organizer that is clear, if not clearer than, the main ideas.  I believe that writing out these combined theories and models (and representing the data graphically) will take me (close) to the target of 100 pages. 

Once I have written the end of Chapter 4, then I will need to rewrite the introduction of Chapter 4 to better mirror the ending.  I believe that this will definitely take me to 100 pages if I haven't already reached that.

Why is 100 pages my target?  The model dissertation that I am following has a 95-page dissertation, and that dissertation is considered one of the longer dissertations in the recent history of my department.  I believe in the philosophy of writing more pages for the first draft than than the anticipated final draft because it is easier to remove than to add.  I had this same philosophy for writing Chapter 2, where I had 50 pages for the target for my first draft, then my co-chairs requested another section to be added, so I added 72 more pages for the second draft.  And in the final draft I ended cutting out whole chunks of sections to reduce the page limit to 47.  I hope that this pattern is not repeated for Chapter 4.

I am trying very hard to limit my writing so far in Chapter 4.  It is difficult to constrain myself when making a point because I want to provide evidence, and it's difficult for me to judge whether I'm providing my readers with too much or too little evidence.  I have purposely played with doing both at times to see where my co-chairs' preferences are.  At this point, I have no idea if my impression of too much or too little evidence is the same as theirs.  Reflecting upon the writing process so far, I feel as though I have practiced restraint for the most part.  In some instances the restraint feels effective in others it feels like I might be giving off the impression that I don't have enough evidence when I do.  That's one practical reason for chairs and dissertation committees, although I don't know how similar my dissertation committee is to a peer review panel of a research journal.

Anyway, the feeling today is satisfaction.  I am glad I'm producing a substantial amount of pages and getting my ideas out on "paper."  My organization is comprehensible but rough and I believe a second edit will provide a more coherent and cohesive presentation of my grounded theory.

Sunday, December 15, 2013

Participant Summaries

This weekend I spent several hours writing individual summaries for each of my participants for two purposes: 1) to send to the respective participants for the purpose of member checking, and 2) to use in Chapter 4.  This writing process amazed me by how little and how much I could write.  Because these were summaries of the participants and not any theory, I tried my best to describe them without hypothesizing or forming a ground theory.  In this sense, I wrote little in the way of my findings based on any previous conceptual models or theories. 

However, I produced a total of about fifteen pages that could be inserted into Chapter 4.  This is a bit alarming.  These fifteen pages mostly serve as single-case analyses.  There is no cross-case analysis, which is where most of the grounded theory is emerging.  I fear that I leave this fifteen pages as they are, which I most likely will not, the scholarly reader may want me to get to the point. 

Writing these fifteen pages or so of participant summaries show how I could have written small case studies on each.  There were a few participants that seemed to have provided me enough data for a single case study, and there were a couple that did not.  In this sense, the writing process has helped me reflect on how I succeeded and failed in the data collection process.  For example, there was one participant who answered interview questions very succinctly.  Even during the time of the interview, I was not satisfied with the quality of his answers, but I felt like I would be pestering him by asking him to elaborate on nearly every question.  Even though I was able to get him to elaborate on a few key questions, this writing processes demonstrated that I should have been more aggressive.  However, this may have jeopardized any rapport I had with him.  Perhaps this is evidence of my newness to conducting qualitative interview.  One of my principles as a researcher is to respect my participants so they feel like active contributors to the research process.  I don't want them to feel like they are just regurgitating data for my benefit.

Another surprising element in this writing process is that I was able to identify at least three themes for each participant and describe them with bits of evidence from their blogs and interviews within a few pages.  I was quite happy that I was able to do this for the purposes of publishing in a journal with very strict limitations on word count.  However, I was a bit upset because I felt like that I could have written at least triple the amount of pages for most of the participants.  If I had done this, the single-case study reports would be around fifty pages.  That's fifty pages of description with little or no grounded theory and no cross-case analysis.

That reflection helped me to develop a strategy for tackling the monstrosity that can be Chapter 4.  The most important point of Chapter 4 is to describe my grounded theory in the cross-case analysis clearly with enough support and evidence from the single-case analyses.  I do not want to misdirect my report towards the unique themes that arose from each case.  There are enough cross-case patterns that answer my research questions that I need to focus on report on those to make my own case.

Finally, writing a summary for each of my participants is another example of how often I revisit my participants data.  Looking back at the minimum number of past visits after the first blog reading or interview, here is a rough count of the number of times I have read through the data for each participant:
  1. Skimming the blog data to make sure the participant met the criteria
  2. Reading through the blog data after the participant consented for the interview
  3. Selecting relevant parts of the blog to bring in to parts of the interview
  4. Copying, pasting, and organizing the blog data onto MS Word documents
  5. Copying, pasting, and organizing the interview survey data onto MS Word documents
  6. Transcribing and organizing the online video conferencing interview data onto MS Word documents
  7. Reading through the first part of interview answers to identify areas to follow up on in the second stage
  8. Reading through the second stage of interview answers to identify areas to follow up on in the third stage
  9. Reading through the third stage of interview answers to identify areas to follow up in the closing
  10. Reading though all of the blog and interview data for the initial stage of coding
  11. Reorganizing all of the blog and interview data to prepare for the second stage of coding
  12. Reading through the reorganized and shortened blog and interview data to prepare for analysis
  13. Organizing the blog and interview data into major themes
  14. Reading through the blog and interview data now divided into themes instead of participants for cross-case analysis
  15. Reorganizing the data back to participants with the themes identified in order of strength
  16. Reading through the reorganized data to form the participant summaries
  17. Reading through the participant summaries to check if they made more sense to my participants than my dissertation committee for the purpose of member checking
  18. Preparing the same summaries for Chapter 4 with purpose of using them for reporting on a grounded theory
I went through this quite fast, but it gives you an idea of how many times I had to pour over the same data over and over again, although some data was removed from further analysis at #12.  Even though I have looked over this data so much, I know I may have missed or misinterpreted something in one or more of these steps.  And that is why the member check is an important step for this study. 

As a side note, this posting may provide an example of how research can encourage obsessive-compulsive behavior.  If I weren't doing research, this behavior would be truly upsetting to my family, my participants, and myself. 

Sunday, December 8, 2013

Grounded in Grounded Theory

Today I have begun my cognitive transition from data analysis to reflecting upon my research methods, particularly while reading through Constructing Grounded Theory by Kathy Charmaz.  The chapters that require my attention the most are Chapters 3-5.  Chapter 3 is about coding in grounded theory practice, something which I believe I have completed.  Chapter 4 is about memo-writing, which I believe I have been doing and continue to do.  Chapter 5 is about theoretical sampling, saturation, and sorting, which I believe is my next step.


The purpose of this posting is to share my thought processes, in which I am making sure I did the best that I could in terms of coding, that I am writing memos as appropriate to myself and my project, and that I am ready for theoretical sampling, saturation, and sorting and maybe even bits of Chapter 6, reconstructing theory into grounded theory studies.

Coding: What have I done?
In September, I began initial coding through my 3000+ pages of data.  The most obvious part of this coding was to identify the chunks of data that directly addressed my research questions, that indirectly addressed my research questions, and that did not address my research questions.  I have previous blogged about this step before.  An important point in Charmaz' book about this stage was to "move quickly through the data" (p.49).  I did not want to get bogged down in this tome of data.

On the next page, Charmaz describes word-by-word and line-by-line coding, both of which I could do for my interview data but would last months if not years if I did so for the blog data.  After I selected blog posts that directly and indirectly addressed my research questions, I was then able to do line-by-line and sometimes paragraph-by-paragraph coding.  This coding process was different than my coding practice for two qualitative methods courses that I took because the data in those courses were limited to either a certain number of pages or a certain amount of time (for interviews).  In those courses, 20 pages of data seemed daunting to most of us.

After my initial coding process, I shared the themes that emerged from the data with my dissertation committee co-chairs.  They helped me to discern a common heuristic that unified most of the themes I identified.  This heuristic, in term, helped initiate the second stage, which Charmaz called focused coding.  I spent the last few weeks engaged in this focused coding, which was faster than I anticipated for a number of reasons:
  1. I had already gone through the data twice, so I was much more familiar with the participants and their data.  I was getting a better feel for the flow of the blog and interview data.
  2. The initial codes were still in place, so I could easily look for them and determine how to focus on them.
  3. I did most of the focused coding over Thanksgiving break, in which I had the most free time since I started my full-time job.  I was able to completely immerse myself in the data.
 Before I end this section, Charmaz identifies two other types of focused coding that I had considered.  The first is axial coding, which helps the researcher visualize the connections between the codes.  Although this process may have been helpful, the guiding heuristic theme that my co-chairs helped identify served as a key connection, so I though axial coding may have been a bit redundant.

The second type is theoretical coding, which I thought of because the guiding heuristic is very close to a theoretical model.  I decided against invoking the theoretical model for theoretical coding as I wanted the data to emerge without a theory at first as I believe the codes already have a coherent relationships.

Memo-writing: What am I doing?
I have been writing memos ever since my pilot study data collection last year, before my research proposal was approved.  The most significant writing times were at various stages of data collection and all throughout the data analysis phase.  I discovered that my memo writing at the beginning of data collection revealed that I was too eager to find patterns.  I believe I see evidence of my maturity to let the patterns emerge from the data instead of forcing it.  I also found that I began to embrace the tenet of KISS (keep it simple, stupid) as I began to gather over a thousand pages of data.

Charmaz writes that there is no formal way or writing memos, and that it depends on the researcher's writing style, personality, and data.  While reading through chapter 4, I found many of the techniques to be similar to those from my creative writing courses in how to best organize your thoughts.  Although I prefer to be modest, I have discovered that this is one of my strengths.  I know how to take notes and communicate effectively to myself.

My favorite bit of advice is to write memos as soon as possible.  In one class, we practiced memo writing immediately after collecting data.  I found that memo writing after analyzing a selected chunk of data to be more helpful.  For data collection, my best memos were when I saw a pattern emerging after the midpoint of the data collection process.  Like I wrote earlier, the first part of my data collection memos is most garbage for the study, although it reveals a lot about me.

Theoretical Sampling, Saturation, and Sorting: Am I doing this?
The first question in Chapter 5 that stands out is, "Can you clearly define your themes?"  I can easily answer yes.  The next concern is that I may not have enough data for these themes, which I do not.  When I begin a cross-case analysis, each theme should have at least a full page of evidence; some may even have more than five pages.

Because I believe I have enough data for each category, I do not feel that I need to interview my participants more.  However, I believe that a member check with each of my participants may prompt one or more of them to clarify my interpretation of the data.

Reading further in Chapter 6, I realize that, for some of my participants, I have clearly saturated the data because they have conclusions to the experiences I am investigating.  Those participants who are still going through these experiences may have a change of perspective by the time of the member check, but I had to decide upon an arbitrary cut point in time or I would have a difficult time ending the project.  Besides, I already have 3000+ pages of data, which seems like a good indicator of saturation.

It is clear to me that most theoretical sampling is done and that I have most likely saturated the data for each major category, but I have not physically sorted the categories yet.  I have done this to some extent in my head, but that's not reliable enough.  I need to write down and visualize theoretical sorting.  Once this is done, then I seem to be ready for the next chapter in the textbook, "Reconstructing Theory in Grounded Theory Studies."

The first step is here is to make sure I have a consistent worldview when I use the word "theory" and apply it in my report.  Reading through this section of Chapter 7, I am surprised to find myself more on the positivist end than the interpretive end of defining theory.  Perhaps this is so because, in the larger scheme of things, I would prefer a mixed-methods approach to the study.  Although I'm not measuring anything in this project, I hope to open the door for measuring variables if this study is compelling enough for further investigation.

In the next section of Chapter 7, Charmaz compares constructivist grounded theory to objectivist grounded theory, and I believe I share more the of constructivist's perspective, especially in that one participant's narrative differs slightly between interview data and blog data, mostly in tone.  I am a bit surprised to discover that some withhold information from the interview that was revealed (earlier) in the blog.

Since I haven't developed my grounded theory yet, I will stop with a reflection on "Theorizing in Grounded Theory" on page 133.  Charmaz shares how researchers view or define the theory of grounded theory:
  • An empirical generalization - I think I may go down this path.  To what extent, I don't know.
  • A category - I believe I created a category (a group of English language teachers or sojourners) before grounded theory, but I'll need to investigate this category with the my theory.
  • A predisposition - This is similar to "an empirical generalization," and I'm unsure at this point which of the two is stronger.
  • An explication of a process - This was my intention for the project, but I believe my data may only reveal a partial explication
  • A relationship between variables - There is a good possibility for my grounded theory to do this.
  • An explanation - Of course my grounded theory will explain something.  It will most likely be a partial explanation.  To declare a whole explanation is impossible for an exploratory study like mine.
  • An abstract understanding - I hope not.  With so much data, I believe I can analyze certain categories more deeply for a grounded theory that is an abstract understanding.
  • A description - I hope this is the by-product of my grounded theory.  I want to shed more light on the experiences of my target population.
Just reading through the list above gives off the impression that I'm still biting off more than I can chew.  I agree with this perception, and my mission now is to give the analysis a sharper focus so I can give a clear report of the findings.

Saturday, December 7, 2013

Single-Case Data Analysis Completed

Although my data analysis is not completed, I believe I have gone deep enough analyzing the interview and blog data for each of my participants.  Consistent patterns have emerged enough that I am ready to begin forming a grounded theory.  This is when I take all the pieces I have found and arranging them into patterns.

Because the most complex pattern in my data is about social relationships, I will use the data to create graphic organizers (undecided on which type at this time) to visualize the social relationships and how they help answer and address my research questions.  I believe I have some firm grounding based in the literature I reviewed, most notably one model that address the host culture complex.  I believe I can connect this model with the graphic organizers I will be creating for each of my participants.  After I can visualize the social relationships in connection to my research questions and one or more models and theories for each of my participants, then I believe I can begin a cross-case analysis.

This morning I completed this last stage of data analysis for my "last" participant.  I consider her the last one only because she was the last participant I interviewed.  I wrote in my memos that this analysis affected me more than the other in terms of questioning my research methods.  Since this just happened, I'm unsure if I should include this part of self-reflection and critical analysis elsewhere than in my memos.  I believe this critical analysis was raised in at least one of my works cited in the literature review, however it is not related to my research questions directly.  It raises more philosophical concerns for analyzing qualitative data, making me more aware of my own biases than previously, or at least in the recent past.  I am thankful that this last exercise in data analysis has raised this awareness before I write about my findings. 

I am now at the point where I need to review the key chapters and sections of my grounded theory texts to make sure I'm ready for the next phase, to better identify the next phase, and to better equip me with the mindset and vocabulary to discuss the next analytical process.