Managing the data

Now that I have completed my dissertation, there are some aspects of the research process I want to write about. One topic is related to questions I received at my defense and issues that have been brought up on #phdchat.

One of the first issues was to use a software like NVIVO or come up with my own process/technology. In the end, after downloading NVIVO and trying it out, I felt that it was too much work for the benefit. I have trouble with using software that does not fit my own style, writing and thinking styles included.

So I opted for index cards and work processing software. The index cards allowed me to move ideas around, making connections between ideas, reorganizing so related ideas could be reused. I also did a lot of hand diagrams. Like NVIVO, I have yet to find a diagramming software that matches the way I like to think.

I was doing a qualitative study, so there was a lot of data that needed to be coded. I ended up writing the emerging codes (in pencil) on index cards, then I would discard or revise the codes (by hand) as I coded electronically. I used the comment function to code data and highlighted key words (which I eventually began to write down to identify data later on).

Using Charmaz's constructed grounded theory process, I answered each research question using the index cards with each of the codes. Not all index cards/codes were used as they did not answer the research questions. I laid out each of the codes and began to create answers to the research questions, moving the index cards around when it appeared they were related. After each "category" of codes, I would fill out an index card with the theme. I then wrote out each code that fell under the theme.

I'm not sure I could have achieved this part of the process if I had used a computer. The advantage of the index card was that I was able to see many ideas at one time and manipulate the cards to reflect the various relationship between the ideas.

Once I had the themes, I then went back into the data and identified passages that supported the emerging themes using the key words I had identified (on the coded index cards) and highlighting the font different colors. I then cut and paste each of those passages, identifying the source and page number (i.e. Ronda, interview 1 as the section title, p. 4 {passage}. I also identified when there was no data from participants for a specific theme. At the end of this I had about 40 pages of data arranged by theme and further arranged by participant and/or document.

Finally, I documented the relationship between the data over the study period, developing a visual of how the data changed over the course of the study.

Once I had this, I began to write up my findings by answering the research questions. I would then go into the data and develop my conclusions/findings. I continued to go back into the master list of "passages" trying to find examples that would support my conclusions or reading through all of the passages to help fine tune my findings. This was especially important during my revisions and preparation of defense. It also helped when I was asked to give more detail on my research process as I had all the cards, hand created models, and printouts of passages I used.

I have written this post in the hopes that those with limited resources can use this process to help expedite their research process. This was a relatively low tech, yet thorough process that anyone could use.

Intuitive hunches in doing primary research

There was a new study that came out in the Journal of Psychology supporting that there might be something to ESP (Extra Sensory Perception). They found that some people have predictive powers that may be more than just pattern recognition which then helps to make a "best choice." They also found that people that have this ability also tend to be risk takers. It could be that they are risk takers because they trust their instincts (although the commentator suggested that those with a risk taking personality are more apt to have the "gift"...i.e. ESP goes hand in hand with risk taking personalities).

At any rate, as I am working on writing up my findings for my dissertation today, I keep coming up against a road block. While I can write on my blog those patterns I see without having to support them with others ideas, and scads of documentation, I do need to do so with my dissertation. As a result, I often find myself with writer's block. My family often complains that my explanations are convoluted and long winded, and I blame my academic training for this. Rather than going with intuition and a hunch, I have to be careful that I can support every sentence of every paragraph of every section that I write. As a result, there are many ideas that I can't include in the dissertation. This makes the process even longer as I try to skillfully craft rhetoric which will support my findings, and present it in such a way that ANYONE can understand it.

Now you know why I have had a headache on and off for the past 2 years!

Knowledge creation in distributed group collaborative writing projects: Methodology

As promised, below is an outline of the methodology I used to study indepth a distributed group as they worked on two collaborative writing projects.

METHODOLOGY

The original data collection was based on the research questions: How do individual members interpret the experience of a collaborative writing process within distrubted workplace groups? What factors influence the interpretation of experience that individuals take with them outside of the group as a result of participating in the collaborative writing process? The purpose in answering these questions was to gain a deeper understanding of what the individual perspective is and the various factors, including group dynamics, context, the role of writing and technology, culture, organizational structure (including agency and power structures), and types of learning, that influence the individual in creating his or her own knowledge, contributing to the group’s shared knowledge, and contributing to organizational knowledge and learning.

Due to the complexity of each of these factors that might influence the individual, it was necessary to have a deeper understanding of the individual, group, and organizational processes and contexts in which collaborative writing takes place. Therefore, qualitative methods, specifically grounded theory and ethnomethodology, were used (Charmaz, 2006; Clancey, 2006; Garfinkel, 1967; Patten, 2002).

After the data was collected, it was apparent that there were many directions in which the data could be analyzed. There were three areas that could be analyzed based on the literature review: the effect of group communication, the impact of organizational structures, and/or the effects and affects of written formats on the collaborative writing process. As a result, as often happens in grounded theory research (Charmaz, 2006), the research questions were rewritten to address the emerging themes and concepts found in the data. I decided to focus on the impact of the collaborative writing process from an organizational learning perspective. The questions still were used to gain a deeper understanding of what the individual perspective is and the various factors, including group dynamics, context, the role of writing and technology, culture, organizational structure (including agency and power structures), and types of learning, that influence the individual in creating his or her own knowledge, contributing to the group’s shared knowledge, and contributing to organizational knowledge and learning. However, the questions were more focused on the methods of common understanding created through individual and group practices located within the organizational social and knowledge structures (Garfinkel, 1967).

The questions were then changed to: What knowledge do members of a distributed workplace group identify as being important when creating a group product? What factors influence the choice of what knowledge is important?
The subquestions included:

• How do individuals define “knowledge”?

• What process or processes does a distributed group in the workplace use to create shared meaning and understanding during collaborative writing projects? What factors do they identify as shaping that process or processes?

• What patterns of work activity are maintained and changed at the individual, group, and organizational level within a distributed group? Who do workers identify with in maintaining or changing work patterns in different contexts?

Methodology Framework

While there are many qualitative methods from which to choose, ethnomethodology allows researchers to look at how the individual and social system (the collaborative writing group, organization, clients, and sponsers in this case) influence each other (Brandt, 2005; Schneider, 2002). Ethnomethodology, with its roots in sociology, looks at the effect that a social system has on individual behavior, which behavior in turn informs the development of the social system either by modifying the structure or reinforcing the social dynamics, assumptions, and power structures (Brandt; Schnieder). Researchers using ethnomethodology (as opposed to ethnography) look at everyday practices as coconstructed social activities based on the perspective of the individual (Clancey, 2006; Garfinkel, 1967). Thus, data is analyzed through the lens of both the individual and the organization through the interaction ritual chains (Hilbert, 1992).

According to Brandt, who has looked at workplace literacy practices using ethnomethodology, this approach not only looks at how an individual’s writing is influenced by the social context in which the writing takes place, but also how the individual then becomes the part of the social context by justifying his or her choices and helping to reinforce organizational writing formats. The context and the individual cannot be separated since the internal processes of the individual helps to create the social context.

Therefore, ethnomethodology will require that I look at the individual as a co-creator of the context for the group and the organization. However, rather than looking at it from an organizational or group perspective (thus separating the individual from the context), ethnomethodology allows me to look at it from the individual’s perspective through the interaction with the social context, however complex that might be (Clancey, 2006). In using this methodology, therefore, the influences on individual and distributed group outcomes in a collaborative writing project in the workplace may be both internal and external to the individual, but their location will not be as important as the impact on the individual’s perception of the collaborative writing process and outcomes.

It is important, therefore, that this study includes:
• a complete description of the perceived contexts by group members at the organizational, individual, and group level;

• individual member epistemologies;

• the perceived social structures and discourse communities in which individuals and the group as a whole work;

• a description of the process the group uses to achieve their task and understanding of each other’s position in order to describe the context of their work;

• individual perceptions of what they should and what they actually bring into the collaborative writing process;

• individual perceptions of the effect members have on the collaborative writing process and their group members, and how they themselves are affected by the collaborative writing process and the other group members;

• perceived learning and knowledge creation due to the collaborative writing process; and

• perceived value of the collaboration at the individual, group, and organizational level.

While ethnomethodology will help inform the type of data that is collected and impact how data is analyzed, grounded theory, especially constructed grounded theory, will guide the research process. Constructed grounded theory uses the identified themes emerging from the data to construct theory. It differs from classic grounded theory in that it interprets the data in developing theory rather than looking for explicit codes initially (Charmaz, 2006).

Group Selection

A major criticism of research on group dynamics and processes in the past is that studies created groups and group tasks in an artificial environment, therefore, minimizing the complexities of group work. More recently, research on groups have studied naturally occurring groups in their own environment, so as to look at the relationship between members in a more authentic environment and capture the dynamics that are the result of organizational structure, shared culture, organizational politics, and shared past experience (Gersick, 1988; Jehn & Mannix, 2001; Jehn et al., 1999; McGrath et al., 2000). As I am trying to understand the external, as well as the internal factors that influence an individual’s experience, it was important that the group I study is a naturally occurring group that would normally collaborate together on a written document.

The first step in recruiting a natural occurring distributed work group was to identify organizations that would allow the study to be conducted, as required by research protocols laid out by the University’s Institutional Review Board. Groups were chosen from organizations dealing in “knowledge” as part of their business. Brandt (2005) identifies this knowledge as the intangible tacit knowledge that adds to the value of a product. Because a large part of services are the intangible product, capturing knowledge and codifying it through writing to make it more tangible is very important to service organizations. There may also be strong external pressures to a group collaborative writing process such as government regulations, organizational quality control to standardize processes, and the demand for knowledge in a tangible format (such as a book, webpage, prototype, or software). As a result, sites were chosen from organizations whose products do not fall into the first three categories of the North American Industry Classification System (NAICS product codes for: 1. Agriculture, Forestry, Fishing and Hunting, 2. Mining, and 3. Manufacturing). Possible service industries included government organizations, economic development, advertising, education, consulting, financial services, transportation and travel, retail, healthcare services, legal services, or software development (United States Census Bureau, 2002).

Organizations were identified through trade organizations (i.e. Capital District Trade Organization), listservs (Communication Faculty listserv), and other professional groups and contacts. I made contact with management within the organization, gave them an overview of the project (based on the application made to the Institutional Review Board) and identified two organizations that worked with distributed work teams.

I then needed to work with an organization in identifying pre-existing distributed groups that would be interested in participating in this study. The group as a whole would need to agree to the study, meaning that if one member did not want to participate, the group would be excluded, rather than replacing that member with someone new that would agree to participate.

In identifying possible groups, potential teams needed to meet the following criteria:

• the group or team should be distributed during some part of the process,

• use technology to support their collaboration,

• and have a core group of between 3 and 7 members.

Groups that are “distributed” do not work within the same physical space on a daily basis. They may be distributed by department, by location (e.g. field office and home office, US based and foreign based) or by speciality (e.g. sales, support, management). The group or team may include face-to-face interaction, but part of the collaboration process should include interaction supported by technology (i.e. telecommunications, internet or e-mail, video conferencing, or shareware). This would help inform the study as to how enforced structure (including the limitations technology might place on interaction) and conflicting contexts (local and group due to distributed team) might affect the collaborative writing process. Finally, research has identified the optimal size of groups at 3-7 members (Moreland & Levine, 2001).

In order to limit the effect that group size might have on the collaborative process, optimal size of the study group would have been 3-5 members. However, in many naturally occurring groups, there are peripheral members that would also have an effect on the group process and at any given time more than 5 members may have been working on the project. It was important, therefore, to include these peripheral members in the study. As a result, the group studied was a core group of 4-5 members working on the collaborative writing project, with the input of 2-3 additional members at any given time during the study, all of which gave their consent for the study.

The first site organization was chosen because of its international distributed work teams. Two groups were identified as potential study participants and I contacted each group member directly with an overview of the project. It soon became clear that because of cultural differences and work pressures, that international members of the group were reluctant to participate. So I changed the site (and organization) of the study. The recruitment of the participants for the second site was modified somewhat. I gave a presentation that was an overview of the study to potential participants. They then gave feedback as to whether they would be willing to participate in the project. Based on this, groups were identified as potential study subjects.

I chose the group that met the criteria for the study. Many of the groups had been working together for a while. The final group chosen had never worked together, was at the beginning of their project, and were distributed in three different locations. In addition, the group used email and a project management software, Basecamp, to communicate.

Initially, I monitored the group’s work through Basecamp, after receiving their signed consent to participate forms. I identified two collaborative writing projects that the core group was working on simultaneously and decided to include both projects in the study. Because of the complexity of the phenomenon of collaborative writing, it is important that in-depth data be collected on both individual perceptions and group perceptions. Since one of the things that I would like to understand is how the dynamics of individuals and groups affect the creation of knowledge at the individual level, the question of agency and enforced structure (from the context) is very important. In order to understand the contexts, group dynamics, and individual perceptions, I studied the one group in-depth before, during, and after collaborating on these common documents. Looking at the two documents gave two contexts in which to study the same group.

Data Collection

This study was divided into three phases. The first phase, pre-task and collaboration consisted of a group of interviews looking at what an individual perceives he or she brings into the collaborative writing process and the perceived context in which he or she was collaborating. In addition to the interviews, data was collected on group interaction in Basecamp, meetings through meeting minutes, and additional documents such as email, document drafts, and planning documents.

The second set of individual interviews were conducted 3 months after the first set of interviews and the group interview. This had the unanticipated advantage of being able to collect examples of other documents, work, and group processes that were the result of those documents created during the collaborative process in the first set of interviews. In addition, it allowed me to explore group member perceptions of the organization, project, group, and knowledge creation over a longer time period.

I personally transcribed all interviews. The average length of these interviews was 45 minutes, the shortest being 38 minutes and the longest being 72 minutes long. In total, 15 interviews were conducted. With one exception, all transcriptions were verbatim from the audio taped recording. The one exception, due to malfunction of the recording equipment, had to be pieced together with partial recordings and notes from the interview. In addition, 40 documents were collected, including background information on the project, postings by group members on basecamp, drafts and final copies of documents studied, and additional documents that were the result of the documents studied.

Data Analysis

Using Charmaz’s (2006) process for coding in constructed grounded theory, each transcribed interview was initially coded at the line level using action words. Starting with an interview from each department, initial codes interpreting the interviews were developed using action words. After the fourth interview, codes were reviewed and common terms were written on index cards. Using constant comparison within the interviews and between interviews some codes were combined, others were dropped, and new ones were created (Glasser & Strass as quoted in Charmaz, 2006, p. 54) based on the analysis of the remaining interview transcripts. Once it was obvious that there was a theoretical saturation in coding the data (Charmaz, Patten, 2002), these codes were then used to develop themes.

Looking at each question, codes were combined that developed themes/concepts or “relational statements” (Glasser & Strass as quoted in Patten, 2002, p. 490) that could be used to analyze the data and begin to develop theory.

References

Brandt, D. (2005). Writing for a living: Literacy and the knowledge economy. Written Communication, 22(2), 166-197.

Charmaz, K. (2006). Constructing Grounded Theory: A Practical Guide Through Qualitative Analysis. London: Sage Publications.

Clancy, W. (2006) Observation of Work Practices in Natural Settings. A. Ericsson, N. Charness, P. Feltovich & R. Hoffman (Eds.), Cambridge Handbook on Expertise and Expert Performance. New York: Cambridge University Press, pp. 127-145. Available at http://homepage.mac.com/wjclancey/~WJClancey/
WJCWorkPracticeObsCUP.pdf

Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall, Inc.

Gersick, C. (1988). Time and transition in work teams: Toward a new model of group development. Academy of Management Journal, 31(1), 9-31.

Hilbert, R. (1992). The classical roots of Ethnomethodology. Chapel Hill: The University of North Carolina Press.

Jehn, K., & Mannix, E. (2001). The dynamic nature of conflict: A longitudinal study of intragroup conflict and group performance. Academy of Management Journal, 44(2), 238-251.

Jehn, K., Northcraft, G., & Neale, M. (1999). Why differences make a difference: A field study of diversity, conflict, and performance in workgroups. Administrative Science Quarterly, 44, 741-763.

McGrath, J., Arrow, H., & Berdahl, J. (2000). The study of groups: Past, present, and future. Personality and Social Psychology Review, 4(1), 95-105.

Moreland, R., & Levine, J. (2001). Socialization in organizations and work groups. In M. E. Turner (Ed.), Groups at work: Theory and research (pp. 69-112). Mahwah, NJ: Lawrence Erbaum Associates.

Patten, M. (2002). Qualitative Research & Evaluation Methods, 3rd Edition. Thousand Oaks: Sage Publications.

Schneider, B. (2002). Theorizing structure and agency in workplace writing: An ethnomethodological approach. Journal of Business and Technical Communication, 16, 196-214.

United States Census Bureau. (2002). North American industry classification system (NAICS): United States Government.

Study on knowledge creation in distributed groups

I currently am in the process of identifying emerging themes for my dissertation. However, recently, the blogsphere has been ripe with discussion on communities of practice and working with groups. This might be the result of the publication of Wenger, White, and Smith's book, "Digital Habitats".

In reading these posts and comments, I realized that my dissertation topic may in fact be relevant! Sometimes I think it is just a matter of finding the correct community in which to place research and ideas. With that in mind, I have decided to make my work in progress available on this blog. As this is a grounded theory based study, I will write my research memos (Charmaz, 2006; Maxwell, 2005) on a blog (research blog posts?). I hope I can get some feedback as I am finding the process of writing my dissertation very lonely.

My next post will outline the research methodology.

Charmaz, K. (2006) Constructing Grounded Theory. Sage: Thousand Oaks, CA.
Maxwell, j (2005) Qualitative Research Design. Sage: Thousand Oaks, CA.

Does any one know anything about open source qualitative research tools?

As I am in the midst of my dissertation data analysis, I keep thinking that there has to be an easier way to do the data analysis than to cut and paste coded data (using the tracking and comment function of word) to do my data analysis. The problem is that I need to maintain a certain degree of privacy in my data analysis as per the IRB (Institutional Research Board) at our school. While I have access to data analysis programs at the university, I can't use them securely.

So after reading Michael Hanley's series on open source elearning, I began to wonder if perhaps there are some open source free research tools. Why do qualitative research tools cost so much? And why aren't universities trying to develop free research tools for others to improve the quality of research?

Is constructing knowledge based on analysis a work literacy skill?

I have been reading "Constructing Grounded Theory" by Kathy Charmaz for my dissertation. As I was reading it, I began to think that many of the points she brings up could very well be used for the workplace.

One problem I see with all of the data and communication tools available to students is the lack of training on how to problem solve and conduct an analysis. In recent years, I have had to dedicate at least one entire class in how to conduct and analysis, and give a lot of support to students as they do their own analysis.

The idea came to me, as I was reading, that I could use the framework presented in the book to help develop analytical skills in my students and to help them to learn to construct knowledge from data by:

1. learning how to collect relevant data. This requires skills in:

• interviewing
• document retrieval
  
• story telling
• interpersonal communication skills
• searching skills
• creating links between content and perception
• the ability to see the other person's perception

2. Coding and identifying trends in the data. There are two different types of coding in grounded theory: initial and focused. In initial coding, a researcher uses action words to describe the data (categories). After the initial coding, the researcher then goes through the data to see if any of the data supports those initial categories. This process allows gaps in the analysis to be more obvious. These gaps then require further data gathering.

Since initial coding requires constant interaction with the data as it is collected, it is a better mechanism to identify gaps in the data early and to construct theory that is less biased (less likely to look for data that supports the theory rather than reading the data to create/construct theory which is then supported by the data).

Imagine giving workers this tool in knowledge work. Attach labels to information (perhaps using programs like delicious?) that is the initial categories. Then going back and reviewing the data to see if it fits into the categories or if there are gaps in the data. From this, their analysis was begin to emerge. The book includes other types of coding (which I am still reading about) that would help to structure an analysis.

3. Memo writing: I haven't gotten to this chapter yet. However, just by skimming the chapter, I perceive this as a way to reflect and document the knowledge creation process. By putting the analysis in words, the individual is creating knowledge for both themselves and others, making their learning transparent. I could see using a blog for this portion.

What struck me about reading this book, is how useful qualitative research methods could be in helping to prepare the 21st century knowledge worker. There still is a focus in most business schools and training programs on the quantitative data (hard facts) while qualitative research methods are considered "soft". However, much has been developed over the last 20 years in the field of qualitative research methods that makes qualitative research as valuable as quantitative research, as long as the research uses a systematic approach. In addition, the deeper thinking skills needed in good qualitative research could create opportunities for the construction of new knowledge and more complex ideas.