Presented at Advances in Qualitative Methods, February 18-20, 1999,
West Edmonton Mall, Edmonton, Alberta, Canada.

Tomofumi Oka
Department of Social Work, Sophia University, Tokyo, Japan.

Click the figures, and you will go back to the top.

1. Main Points
2. Research Context
3. Focus Group Interviews
4. Main Questions of Interviews
5. "Formal" and "Informal" Problems
6. Anticipated Pitfalls
7. Procedure for Interviewing
8. Three Goals of Questioning
9. Coding Procedures
10. Counting Coded Lines of Transcripts
11. Codes on Multiple Lines
12. Group A "Kindness"
13. Group B "Friends"
14. Group C "With-A-Patient"
15. Group D "Men"
16. The Four Graphs
17. Averaged Numbers of Lines
18. The Leadership Problem
19. Degree of Shortage 1
20. Degree of Shortage 2
21. Degree of Shortage 3
22. Why Are There So Few Leaders?
23. Causes of Shortage of Leaders
24. Numbers of Participants Pointing Out Causes of Leader Shortage
25. Numbers of Lines Pointing Out Causes of Leader Shortage
26. Combinations of Causes 1
27. Combinations of Causes 2
28. Two Types of Cause and Feeling
29. Types of Cause and Reactions to Them
30. A Hypothesis on the Shortage of Leaders
31. Conclusions

It is a great pleasure for me to be here and to be able to speak at this session. My purpose today is to talk about how I used data obtained from focus group interviews to identify and analyze the management problems of self-help groups for parents of sick children.
	I would now like to show the four main points of this presentation. First, I will outline a moderating strategy I used to avoid group-think, and the results achieved with this strategy. Second, I will talk about a counting strategy on my transcripts of the focus group interviews. Third, I will show how I tried a Boolean search on the data patterns. Fourth, I will give an example of how I used a combination of two strategies: counting transcript lines and interpreting non-verbal communication to analyze the problem of the shortage of voluntary leaders confronting some self-help groups.
	First, I will talk about my research context. The participants in our focus group interviews were the leaders of self-help groups for parents of sick children. I asked twenty-one self-help groups to send me participants; sixteen of the groups did so. I had four focus groups, each with six participants. I arranged it so that the participants in each focus group belonged to different parent groups. In the participatory research project, the research team consisted of six leaders of self-help groups, a social worker employed by a self-help group, and myself. The research team was established as a subgroup of an association of parent groups which represents twenty-one different parent groups. When I formulated this interview, I had been working with these groups for about four years, and had completed a series of unstructured individual interviews with thirteen groups. By then, I had developed a rough background knowledge of these parent groups, which I used to formulate the focus group interviews.
	Each interview consisted of a two-hour discussion which focused on the management problems of self-help groups. My role was that of moderator. I also used an assistant moderator - a graduate student of nursing who had completed another research project on the same groups. During a debriefing period after each interview, the assistant moderator and I discussed the non-verbal communications of the participants. The interviews and discussions were tape-recorded and transcribed. I could not share the interview transcripts with the other members of the research team because they were group leaders, and so the participants' discussions was directly related to them. I needed to keep confidentiality. However, I did show my analysis and summaries of the transcripts to the research team and consequently my analysis was improved by the feedback of the participants and other group leaders.
	I asked two main questions in this interview: The first was: "What problems do your groups have?" During the one-hour discussion on problems, I let the participants suggest and discuss as many problems as possible, and then asked them to choose the main problems to be discussed in the next part of the interview. In the remaining one-hour discussion, I wanted to find out about the solutions used by the groups so I asked: "How have you solved these problems?" Seeing that time is limited in this presentation, I would like to focus on the first question: "What problems do your groups have?" because as anticipated, difficulties did arise when the participants discussed the "management problems" of their groups.
	I would like to divide the groups' problems into "formal" problems and "informal" problems. "Formal" problems are problems which the groups are dealing with "formally". For example, medical problems, the children's problems at school, employment problems, and any problems caused by illness. These problems are often discussed in the groups' newsletters and group leaders are often asked to speak about them in public. They are used to talking about them to the mass media: to newspapers, radio, etc. and to the people in control of various administrative organizations, and of course to the public. However, what I wanted to ask about in this interview was the groups' management problems. Management problems are "informal" problems in the sense that they are not mentioned in any of the groups' formal documents. Therefore, members are not used to speaking about them outside their groups. Such problems might be a source of "shame" to some groups and so I anticipated that the participants would be reluctant to talk about them. Another difficulty that arises in asking about such problems is the word "management." Management sounds so business-like or bureaucratic that some leaders might think that "management" did not apply to their voluntary groups.
	In this slide, I highlight two problems that I anticipated would arise in the interviews. The first was that the participants might confuse their groups' management problems with family problems that are caused by the children's illness. To avoid this confusion, my main strategy was to let the participants compare their problems. The participants in each of the focus groups belonged to various groups representing the families of children with different diseases. These diseases were so different that the problems caused by them were also very different. Therefore, I thought by allowing the participants to compare the problems of all the groups and to identify the ones that were most common, those problems caused by specific diseases would be dropped from the list of problems to be discussed later. Shortly I will demonstrate how this strategy worked very well. The second pitfall I anticipated was that "group-think" would hamper the development of the discussions. We needed a questioning strategy that would ensure all alternative opinions and suggestions were fully explored.
	This was my procedure for questioning. First I asked: "What problems does your group have? Write them down and explain them." I provided all the participants with large sheets of thick paper and magic makers. I asked them to write their answers on the paper so the other group members could see them. Next, I collected the papers and hung them on the wall. Then I said, "Compare all the problems that you have compiled and find out which are the most common." Then after some discussion, I hung up a list of problems I have obtained through my previous research. I added these extra problems in case there were problems that the participants might have missed. I rewrote this list after each focus group interview, adding problems identified by the previous focus groups. While showing this list and all of the participants' answers, I said, "Choose some common problems for us to discuss later and to try to find solutions for." However, some groups found common problems that they did not want to discuss, because they said it seemed almost impossible to find any solutions to them.
	By using this questioning procedure, I wanted to attain three goals: First, to acquire a diversity of points of views, by letting the participants write down their answers. Because their answers would be already written on the card, they could not be modified in subsequent discussions, no matter which direction the discussions took. Second, to promote the participants' analysis of the problems, by allowing them to compare their answers. As was shown later, this discussion procedure helped participants focus on their common management problems and leave aside problems associated with any particular diseases. Third, to clarify the priority of topics, by letting the participants choose which problems to discuss later. This gave them a chance to decide which problems were more important. In the next few slides, I will demonstrate whether or not this procedure was of any help to me.
	Before discussing the results of the interviews, I would like to explain my coding procedures. My first step was "open coding." I made a list of thirteen codes which I then put on the transcripts. Although I had a research team, I did all the coding of the interview transcripts myself, because I had promised the participants that I would ensure complete confidentiality throughout the research, so that no-one would know who had said about which groups' shortcomings. Some members of the research team belonged to the same groups as some of the participants, so I could not let anyone else read the transcripts. Consequently, I could not use "cross-checking" by another researcher, which is one of major defects of my research.
	After coding, I counted the coded lines with NUD*IST 4, and totaled the number of lines of code. About eight percent of the lines had plural codes, so I counted these in duplicate. I then changed the totals into percentages, because the transcripts of the four focus group interviews were different in length. After that, I averaged the total number of percentages of code dedicated to each of the 13 topics, across all four transcripts. The results will be shown on the next slide.
	Of the total of 13 codes, these 7 were found on multiple transcript lines. The figures in the column: "Lines" show how often the topics were discussed. This column works as a sort of index. The column "Persons" displays the number of participants who talked about these problems and the column "Groups" shows the number of focus groups which discussed these problems. Let me explain these problems briefly. The first problem concerns leaders. Leadership problems include difficulties finding voluntary leaders and conflicts between old and young leaders. The second is money. Money problems include not having enough funds and members who do not pay their membership fees. The third is counseling. Many problems associated with counseling come about because group leaders have many chances to participate in peer counseling. The fourth problem consists of difficulties arising from the children's illness. The fifth problem involves finding suitable places to meet. Most of my participants work around Tokyo, which is one of the most crowded areas in the world, so they have difficulty finding good meeting places. The sixth problem concerns social actions. Many group members have few skills and little knowledge about how to organize social actions. The seventh is the problem of adult patients. Some parent groups have very significant problems with adult patients, but for other groups this problem is irrelevant because their children inevitably die young. I will now focus on the top 4 problems and demonstrate how we can analyze the way in which they were discussed in the focus groups by counting the number of coded lines each problem was given.
	This diagram shows what occurred during the first group's discussion. The graph shows the changes in the number of coded lines given to each problem while the participants were answering my questions. First, when I asked them to list their problems and explain them, this group talked about problems associated with their leaders (see the yellow area), and with counseling (the orange area). When they compared their answers, they also talked about money problems (the light blue area). One participant said that her group had enough money, and she then explained ways they had raised funds, however although money problems were talked about a lot, they were not chosen as a topic to be discussed later. When the group was required to choose topics to discuss further and find solutions for, they talked about counseling again. Finally, they chose leadership problems and counseling problems. I named this group "Kindness," because the participants were very interested in ensuring counseling was provided for people who needed it.
	I named Group B "Friends" because they looked to be enjoying the discussion. As you can see, this yellow area is very large, which means that this group's discussion focused on problems associated with leaders. See the red area. This means that when they explained their written problems, they talked about illness-related problems. During the choosing phase, this group scored low in the words section. This does not mean that they had little to say while choosing their topics. On the contrary, I saw that they were enjoying talking about the leadership problems so I let them continue in order to gain interesting information from them. As a consequence, little time remained when it came to choosing problems for later discussion. Finally, this group chose to discuss leadership problem.
	This group is named "With-A-Patient," because one of the participants was an adult patient. Actually, I asked the association of the parent groups to send me interview participants who were leaders AND parents. In spite of this, one group sent me an adult patient, because she was an active group leader. In the debriefing discussion, the assistant moderator and I confirmed our shared impression that this group did not enjoy their discussion. As you can see, the topics discussed were not focused in the comparison phase. Some members of the research team, who are parents of sick children, said later that parents could not speak freely in front of patients. The parents in Group C seemed to feel ambivalent towards the adult patient, because as one participant said, "We have been criticized by adult patients who say, 'This group is for the benefit of the parents, not for their children." And the last word of the adult patient in this group was, "Remember that your children might outlive you, because parents die, too." She said this, not as a parent-group leader, but as a representative of sick children.
	The last group's nickname is "Men," because this was the only group where men were in the majority. Five of the six participants were men. In the other groups, four or five of the six participants were women. Concerning the streams of this discussion, this group succeeded in focusing on common topics. In the comparing phase, they talked about leadership problems, and while choosing topics for further discussion, they discussed money problems. Next, we will view the four graphs simultaneously.
	As you can see, only Group C has a dip in the line from writing to choosing. This means that the members of Group C were not so successful in making comments about the problems of other groups. This is consistent with what the assistant moderator and I both felt at the time. While I cannot discuss the details of these graphs, I can say that they reflect the streams of these group discussions very well. Therefore I would like to say that this graphic strategy is helpful for enabling us to get "broad ideas" about the results of focus group interviews.
	This is a graph of the averaged numbers of lines across all four groups. As you can see, while in the comparing phase, the participants discussed leadership problems most frequently. In the choosing phase, money problems and counseling problems were included. While writing and explaining their problems, participants talked about problems related to the children's illness, but when comparing problems, they stopped discussing illness altogether. This demonstrates clearly that counting the number of coded lines given to each topic is a simple tool for detecting changes in topics being discussed during group sessions.
	I would now like to proceed to the next topic in this presentation: leadership problems. As I have already mentioned, most of the participants named leaders as their group's most important problem area . The leadership problems can be classified into two groups: the shortage of voluntary leaders, and conflicts between leaders. Because we have little time available to discuss them, I would like to concentrate on the first one: the shortage of voluntary leaders.
	First of all, I picked out the transcript lines related to "the shortage of voluntary leaders." Then I made up a scale from "-2" to "+1". Lines mentioning the shortage of voluntary leaders and the lack of ways to reduce this shortage were marked as -2. Lines which mentioned the shortage, and some ways in which to reduced it were marked as -1. Lines which indicated that there was no shortage of voluntary leaders but a shortage of suitable candidates for chairperson were marked as 0. And lines indicating that there was no shortage at all were marked as +1. After that, I divided the participants according to this scale.
	Among the total of twenty-four participants, the comments of ten persons were marked -2, and four were marked -1. So, just over half of the participants said that their groups had a shortage of voluntary leaders. Next, we will see the ratios by group.
	This shows the number of members within each of the four groups who mentioned they had leadership shortage and the degree to which they were affected. Group A, B, and C had two participants who said nothing about the leadership shortage. But look at Group D. In this group, four scored -2, and the remaining two scored -1. So all the participants in Group D felt strongly about the shortage of leaders. Therefore, we would expect that this group would express most freely and frankly how they felt about the shortage of leaders. As I will demonstrate later, Group D had interesting comments to make about why they had a shortage of leaders. The participants who are marked -1 or -2 gave various reasons why they believed voluntary leaders were in short supply. I have picked out for further analysis all the responses that were given by at least two of the participants who are marked -1 or -2.
	These are the reasons that at least two participants gave for the shortage of voluntary leaders. As you can see, four reasons were given: The first: group members are dependent on their leaders and so they do not dare to become leaders. The second: group members know how heavy the leaders' burden is and so they are not confident that they will be good leaders. The third: their children's illness and the burden of caring for them prevent many members from committing to their group's activities. The fourth: group members are interested in other things besides their parents' groups. Next, I will show how much discussion took place about these reasons, and by how many participants in how many groups.
	The graph of the right side shows how many lines were used when discussing each of these reasons. I used NUD*IST, in which each line has 70 characters. So the bottom item on the graph: "Other interests" scored only thirteen lines. Only one participant in each of the three groups mentioned "Other interests" as a reason for the leadership shortage, and nobody else added similar comments. Therefore, I would like to leave "Other Interests" and focus on the top three reasons: dependency, the burden of being a leader, and the children's illness. These three reasons were given by six participants, remembering that some participants gave more than two reasons. I was also interested in the combination of reasons given, and discovered that they formed an interesting pattern.
	This table shows the number of participants per group who suggested reasons for the leadership shortage. We can see that Group D had a different pattern from the other three groups because they emphasized "Dependency" whereas the other three groups seem to put more emphasis on "Burden" and "Illness". Next, I will show how many lines were given to each reason by each group.
	This table shows the number of lines each group gave to each of the three reasons suggested. As you can see, Group D talked a great deal about "Dependency." On the other hand, Group A talked about each of the three reasons almost equally. Groups B and C did not discuss "Dependency" at all. From this table, we might think that "Dependency" is different from "Burden" and "Illness" in certain aspects, and that "Burden" and "Illness" have something in common. To confirm this, I made another table combining the relevant data.
	This is a table showing "truth" responses to the causes of the leadership shortage. What is interesting here is that there were only three occurrences where the causes appeared in combination. Let me explain: The first line of the table indicates that four participants in two groups gave "Dependency" alone as the reason for the leadership shortage. The second line showing causes in combination indicates that four participants gave both "Burden" and "Illness" as reasons. And the bottom line shows that two participants in one group gave all three causes. I will now show the causes given in combination according to group responses.
	After deleting the patterns that did not appear in combination, only three combinations remain. Only "Dependency" was mentioned by four participants in Group D. The combination of "Burden" and "Illness" was mentioned by one participant in each of the four groups. All three causes were given by two participants in Group A. So, we can say that the leader shortage results from two types of cause. What is most interesting is that the participants' feelings towards these two types of cause appear to be quite different. I will show you what I mean.
	By observing the participants' non-verbal communications and reading the transcripts, I found that while talking about the causes of the leadership shortage, the participants expressed different feelings towards the members who did not commit to their groups' activities. First, while pointing out member dependency, they seemed to blame or criticize the members. One participant said, "They come to us to get information. After enjoying the 'fruits' of our labors, they say 'Good-bye'. They only come back to us again when they are in difficulties." On the other hand, while referring to the burden of leadership and the children's illness, the participants seemed to accept and be reconciled to the passive attitudes of their members. Participants said, "They know how heavy the leaders' workload is," and "Their children are so young that they cannot go out."
	This table shows the causes mentioned and the reactions and locations of the participants who referred to these causes. Here, we can see a clear pattern: the participants who referred to "Dependency" only were located only in Group D. On the other hand, the participants who mentioned "Burden/Illness" only were scattered throughout the groups. What caused this difference? The answer is probably to do with the number of participants in each group who were affected by the shortage of voluntary leaders. In Group D, all the participants mentioned the shortage of voluntary leaders and so they found themselves sharing the same circumstances. This feeling of sharing the same experience probably enabled them to express more freely any negative and aggressive feelings they had towards the passive members in their groups. On the other hand, in the other groups half or less than half of the participants mentioned the shortage of voluntary leaders. Moreover, in doing so they demonstrated that their group's circumstances were different from that of the other groups. Being short of voluntary leaders is probably an indication that something is wrong with their group or its leaders. Would they therefore feel free to express any negative feelings towards the passive members in their group, in front of group leaders who were possibly more successful than they were? Probably, not. They had to keep face.
	So, I reached a hypothesis on the shortage of leaders. The participants feel ambivalence about the causes of the shortage of voluntary leaders, however the reasons they give for this shortage may be modified by conditions and influences within their focus groups. Finally, let me summarize.
	I have reached two conclusion: 1. Counting lines in interview transcripts can help detect data patterns. 2. Different combinations of participants in group interviews can produce different data, depending on the participants' ambivalence towards the topics under discussion. Lastly, I would like to add that this research was done under the supervision of Dr. Ian Shaw, Cardiff University, UK. Thank you very much for your attention.