Introductory Comment
What I present here is a group discussion methodology which is much more complex and requires much more effort than the usual methods. However, if one is sincere about wanting to capture the input from participants, this methodology is well worth the extra effort.
In my previous post, I described how contributions in discussion groups are almost always filtered and distorted as they are recorded by scribes and repeated by reporters. The wisdom, knowledge, passion, and depth of the discussion participants is often lost in the translations. "Final reports" are usually poor representations of what actually transpired in the discussion.
I have used a different technique based on the assumption that every word in every statement by every participant is important.
I will describe how the process worked in a typical project. I will first describe the setting of a hypothetical discussion event:
A community service organization is about to launch into a strategic planning process and they want to do a comprehensive assessment of their current performance. They decide that they need input from all of their stakeholders: Board members, staff, clients, funders, partner agencies, suppliers, etc. They develop a list of key questions for which they would like answers. They also want to give each group a chance for open comments that they feel are important.
The Process
Focus groups are arranged for each of these groups. In this example, all participants are gathered in the same place at the same time. Each focus group is limited to about 6 to 8 people in order to allow for plenty of input opportunity from each person. Each group is homogeneous; that is, all members are from the same group: board members, staff, clients, etc. (There can be other situations where you want mixed groups to develop mutual understanding. That is not the objective in this scenario.)
Each group gathers around a small table. There is a discussion leader for each table to keep the discussion flowing, on topic and to watch the time, All members have the list of questions that will be discussed. In the middle of the table is a recording device capable of receiving voices from all directions (or a hand mike which can be passed around). The discussion proceeds through all the questions and open discussions. The complete record of the discussion has been captured at each table. The recording identifies the group involved. Maintaining identities of each speaker is optional, but usually not necessary.
The crucial part of this methodology comes into play at this point. here are the key functions that need to take place now.
The analyst identifies the list of entities that will be used to define each part of the audio file. Here are the entities that I normally used:
Group (Client, Staff, Board, etc.)
Question being discussed (1,2,3,4....)
Response Type: Positive, Negative, or neutral
Subject (This would include a long list of subjects which can be added to as new subjects show up. Some examples to illustrate this field: funding, cost control, staffing levels, leadership, strategic planning, budget control, legislation, community involvement, transportation, staff training, etc.)
Each statement will be analysed so that it is fully defined. It will be known to what question it refers, what group it came from, whether it was a positive, negative, or neutral point of view, and what aspect of the subject was mentioned. (Usually the person making the statement is not identified, but there are valid exceptions to this.)
An important design factor needs to be mentioned here. Statements in many cases will be complex and multi faceted. One statement could refer to two or more subjects, could be partly positive and partly negative. Because of this, you really need a well designed relational database to manage the data. A simple spreadsheet would not be an efficient way to process this material. Each statement may relate to one or many subjects. Each subject may be referenced by one or many statement. These are known as :many-to-many relationships which are best handled by a well-designed relational database.
1. Each audio file is transcribed, verbatim.
2. A separate copy of the transcription is then parsed, i.e. broken down into sentences or parts of sentences that deal with different subjects.
3. Each statement is assigned the attributes that apply to it. Group, question, Responses type (positive, negative, or neutral), subject or subjects alluded to.
4. Then all of the responses from all of the groups are combined in the same table structure in order to facilitate a very comprehensive analysis. We often describe this process as slicing and dicing the data.
The end product of a survey such as this can generate literally hundreds of thousands of observations obviously more that anyone would have time to consider, absorb, and react to. The analyst will conduct a large number of queries in order to identify the most descriptive and revealing snapshots of the data, and builds these into a report.
General Reports
Here are some typical high level reports that might be presented:
Percentage of positive, negative and neutral statements.
- same for each Group (staff, clients, board members, etc.)
- same for each subject dealt with in the survey
- same for each question on the survey
Percentage of positive, negative and neutral statements on each of the questions discussed.
Listing of all the personal statements from each group on each topic; then the same for all the negative statements.
Comparisons of all groups on a spectrum of issues arising within the discussions. Special note can be made of major differences among various groups - not to decide who is right or wrong, but to understand the different perspectives within each part of the organization.
Normally the recipients of the reports will find that many of the observations confirm their own perceptions, but at the same time they find many surprises - surprises that come about when the perspectives of some groups in the organization differ from theirs. We often find major divides in perceptions from people at different levels of the organization (executives. middle management, front line staff, and clients) and this kind of open participatory discussion is very effective in providing full understanding of why the perspectives differ.
It is not at all unusual when this kind of report is presented to the organization for people to present additional questions. This style of study makes it easy for us to drill down into the data to answer specific queries. For example, if someone were to ask if there is a correlation between those who expressed concern with the leadership in the organization and those who were positive about the need for strategic planning, it would easy to make a query to answer that question.
It is fairly difficult to describe this whole process. I hope that the above is sufficient to give you a sense of how this approach works.
I am having difficulty to enable comments but readers on this blog format, but if you have questions, I am in the phone book. On Facebook look for "brockest", or write me at Box 492, Blyth Ontario N0M 1H0, and I will answer you.
Blog address is https://allaboutblyth.blogspot.com
