Case study 1
An evaluation of an action learning
This is a resource file which supports the regular public program "areol" (action research and evaluation on line) offered twice a year beginning in mid-February and mid-July. For details email Bob Dick.
... in which an action-research-based evaluation study is described. The emphasis is on the use of triangulation to achieve rigour
A Czech translation by Valeria Aleksandrova is available at http://www.autoersatzteile.de/blog/vyhodnoceni-akcniho-uceni-program.
- Group interviews
- Other sources of data
- Comparing participants and other stakeholders
- Meeting with providers
- Other reading
The purpose of action research is to achieve both action (that is, change) and research (that is, understanding). In my view, effective action research tries to work towards effective action through good processes and appropriate participation. It tries also to collect adequate data, and interpret it well.
At its best, action research is done so that the action and the research enhance each other.
This file presents the first of two very different case studies. Both, in different ways, illustrate some aspects of the simultaneous pursuit of change and understanding.
Below you will find an overview of a particular evaluation study which used an action-research approach. The evaluation was of an action learning program. We were approached by the people conducting the program to evaluate it.
The case study illustrates some of the features of action research applied to evaluation. (I address them further in later sessions of areol). You will read:
o a brief background to the study,
o an account of the interviewing method used, and
o an overview of the comparison of the interviewer findings to other data.
There were constraints: limitations of time, money, and availability of people. So it was not done in a very participative manner. The participants took part mostly as informants.
The program providers were more directly involved in the study. Even here, though, their participation was not substantial. It was our perception that they were committed to the program, and keen to improve it. We judged, therefore, that the disadvantages of low participation might not be as important in this instance as they might often be.
(A second case study describes a more participative approach.)
For brevity, in this case study I'll emphasise the methods we used to check our data and interpretations.
We gave attention to building relationships and clarifying roles with participants and clients. I won't detail that here. For similar reasons I will describe only part of the evaluation in detail.
As you read the case study, you will note
- an open-ended start to each piece of data collection -- each interview or focus group
- the use of multiple and diverse data sources
- the use of a number of different data collection methods
- a step by step process in which the later steps could be designed to take account of what we learned from the early steps
- a continuing focus on challenging the data and interpretations already collected; in particular, when there were apparent agreements between informants we deliberately sought exceptions; when there were disagreements we sought explanations.
A colleague and I were approached by a staff member in the training and development unit of an organisation. We were asked if we would evaluate a project-based training program which the unit had set up for organisational members. We agreed.
To simplify the consultancy we decided to allocate the tasks. Each of us would carry out different parts of the evaluation. My colleague would collect data from the non-participant stakeholders. I would work with participants.
We could immediately see an advantage of this division of labour: if we did our early work independently, each of us would later be able to check our part of the evaluation against the conclusions drawn by the other.
I would be using an interview method which works best with two interviewers or more. I therefore negotiated for another colleague, Karyn Healy, to join me. This was readily approved.
The technique we used was convergent interviewing, described in more detail in the archived file iview.
We explained to the training unit that we preferred to use a "maximum diversity" sample of participants. This provides a greater range of data than a random sample. We asked if such a sample could be drawn up for us. This was done: we were given a list of names and telephone extensions of a sample of participants.
We then asked for a second sample, to take part in group interviews. This was also done. This sample, too, was as diverse as possible. There was no overlap between the two samples.
From time to time throughout the evaluation we talked to people who were involved in the delivery of the training. This included:
- early negotiation with the person who first approached us, clarifying the unit's expectations and our roles
- presenting results in progress to members of the unit, for their comments and feedback; this also gave the unit members a chance to provide us with background information, and to add some colour to our interpretation of the data we were collecting.
- contact with unit members just prior to writing the final report, to check that we had covered the information they required
- meeting with unit members to give them a further chance to react to, and challenge, the report.
We regarded this contact as important. Our final recommendations were more likely to be understood if we clarified our intentions along the way. They were more likely to be acted on if we addressed the concerns of unit members, and involved them in helping us interpret the data.
Karyn and I each interviewed half of the interview sample. We phoned each of them, explained who we were, and negotiated a time to meet them. In a few instances we were unable to find a suitable time, and instead arranged a telephone interview.
Each of us first carried out one interview with a different informant. At the start of each interview we explained our role in some detail. We made clear what use would be made of the data they provided to us. We also explained that we would take pains to protect their identity in our reports to the training unit.
To begin the interview proper, we said "Tell me about [the program]". We then used attentive listening, and other verbal and non-verbal signs of attention, to keep our informant talking for about 45 minutes. In this and similar studies, I have noticed that most informants appear to talk freely in response to this approach. I think that having someone listen carefully to your every word, and show every sign of being very interested in it, is an affirming experience for most people.
(In some similar studies I have been given information that, if I had revealed the identity of the person giving it, would have caused problems for them within the organisation. With some exceptions, it is usually not difficult to gather a lot of valuable information in a relatively short time.)
Towards the end of the interview we asked more specific questions. (The role of these questions will become apparent soon.)
During the interview we listened for important themes. At the end of each interview we asked our informant to summarise their interview for us. We mentally compared their summary to our recollection of the themes, as a check.
Finally, we restated the guarantees about anonymity, and thanked the informant enthusiastically.
After each pair of interviews, Karyn and I met to compare results. We made particular note of any themes mentioned by both informants. (In the later interviews we also noted themes mentioned by only one of the two informants, but which had come up in earlier interviews.)
Here is an important feature of the technique... For each theme identified, we developed probe questions to explore the theme further in later interviews. Agreements and disagreements were differently treated:
- Sometime both informants mentioned the same topic and revealed the same attitude. When this occurred we devised one or more questions to seek out exceptions to that agreement
- Sometimes both informants mentioned the same topic, but with different attitudes. We then developed probe questions that sought information to explain the disagreement.
In short, we actively sought out exceptions to apparent agreements, and explanations for apparent disagreements.
All interviews began in the same open-ended way. We wanted to ensure that the information we collected was contributed freely by the informants. We didn't want it to be determined by the questions we asked. As the series of interviews progressed, the probes increased in number and detail.
In other words, we allowed the data, and the interpretations placed upon it by our informants, to lead us deeper into the study. You will notice, too, that we refined questions and answers over the series of interviews. Guided by the informants, we developed a more in-depth perception of the program we were evaluating.
When the series of individual interviews were nearing completion we conducted group interviews.
Karyn and I held a detailed planning session before the group interview. We reviewed the results so far, and noted any uncertainties we had. We then planned the group interview to begin again in an open-ended fashion, saving the more specific questions for later.
Our chosen process was a particular form of focus group. Compared to many focus groups it was modified to allow more systematic data collection and more interpretation of the data by participants.
You will recall that the people delivering the program assembled a sample for us. Before the focus group, each of these was also asked to fill in a brief questionnaire indicating which sessions they found most satisfying, for whatever reason, and which they found least satisfying.
There is a description of this form of structured focus group in another document. Here I will provide only a very broad overview. The main steps were:
- We introduced ourselves, described our role and the purpose of the group interview. We gave guarantees about anonymity, and explained what would be done with the information. We did this in a way that established rapport with the informants.
- We asked each person to think about their reaction to the training program, jotting down notes to themselves as they thought of an issue.
- Then followed an unstructured discussion in which everyone was encouraged to speak. We introduced this discussion by saying that we were keen to know what issues people agreed on; we were also interested in noting where different people had different experiences or different views. We explained that the discussion wasn't intended to provide an opportunity to argue for a point of view. We wanted them to note and explain the disagreements as well as the agreements. During this discussion we also asked people to note down the themes which emerged from the discussion. We explained that we would collect these (on an electronic whiteboard) after the discussion.
- The themes and issues were written on the electronic whiteboard, and then further discussed. Through further discussion the most important of these were then identified and marked.
- We asked, and recorded the answers to, the more specific questions we had which were not already answered earlier in the focus group.
Other data sources
From time to time during these processes we also gained access to other data sources. The most important of these were written accounts by participants of the projects they had conducted, and what they had learned from their experience. To complement this one of us was also able to attend a one-day workshop at which each of the participant teams gave a spoken presentation on their project. We used this information to refine the interpretations based on individual and group interviews.
After the interviews and focus groups were completed we again reviewed the data we had. We agreed on a set of themes which captured the main advantages and disadvantages of the program we were evaluating.
It was time to compare our results, from participants, to those our other colleague had collected from non-participants involved in the program.
Comparing participants and other stakeholders
In comparing our data and interpretations to those provided by the other evaluator, we followed a similar approach to that used in other comparisons. That is, we particularly noted agreements and disagreements.
In practice, there was high agreement on the main themes. Where there were differences, these most often reflected the differing roles and interests of those providing the information.
Following this, we agreed with our colleague on the way in which we would report the results. We decided that we would report the strengths of the program as such. The disadvantages we decided to report in the form of recommendations for improvement. Our intention was to make it as easy as possible for those delivering the program to absorb and understand the data and interpretations we were to report.
Meeting with providers
A meeting was then set up. Those attending consisted of two of us, and the team delivering the program. In addition to the program leader, the delivery team included program designers, project facilitators, and presenters. Lasting several hours, this was conducted as an initial report and then an open discussion.
We began by presenting a summary of results, and inviting comments. More detailed evidence was presented when it was appropriate during the discussion. From our (the evaluators') point of view, there were three main purposes:
- providing information to those who could make use of it; we wanted this to be interactive so that there were more opportunities for clarification and inquiry than with written presentations
- obtaining information which challenged or refined the data and interpretations obtained by the evaluation so far
- gauging how to present the information so that threat to the program team was minimised; this, we thought, would increase understanding and subsequent action by the program team.
Finally, a formal report was written and presented.
Minichiello, V., Aroni, R., Timewell, E. and Alexander, L. (1990) In-depth interviewing: researching people. Melbourne: Longman Cheshire.
Dick, B. (1990) Convergent interviewing, version 3. Brisbane,Australia: Interchange.
Dick, B. (1998) Convergent interviewing: a technique for qualitative data collection [On line]. Available at
On focus groups:
Dick, B. (1998) Structured focus groups [On line]. Available at
Morgan, D.L. (1988) Focus groups as qualitative research. Newbury Park: Sage.
Stewart, D.W. and Shamdasani, P.N. (1990) Focus groups: theory and practice. Newbury Park: Sage.
Copyright (c) Bob Dick 1997-2013. This document may be copied if it is not included in documents sold at a profit, and this and the following notice are included.
This document can be cited as follows:
Dick, B. (1997) Case study 1: an evaluation of an action learning program [On line]. Available at http://www.aral.com.au/resources/case1.html
Maintained by Bob Dick; this version 1.06w last revised 20131221