2. Research Approaches

2.1 Aims

We do not know enough to arrange efficient flexible learning, especially concerning collaborative learning using forum systems (Moore & Kearsley, 1996). This might cause students to spend their time on courses without learning what they were promised, and teachers working hard without producing acceptable results. It also implies that the forum systems we use are not quite as effective as they could be.
   The aims of the research reported in this thesis are
  1. to review and supplement the knowledge about collaborative learning via forum systems by
  2. finding and describing factors influencing effective collaborative learning via forum systems, and especially
  3. identifying the teachers´ role and behaviour and need for system support and
  4. to use this knowledge to suggest effective teacher behaviour and ICT support functions.

2.2 Research Questions

When I started research within the area of flexible learning, the first questions that came to my mind were:
   Q1 What new possibilities do teaching and learning at a distance offer?
   Q2 What new problems are introduced?
   The best way to find out was to try. So I gave a course together with a colleague and made a research study on it at the same time (Study A). We found, among other things, that in this type of learning environment:
   R1 It was difficult for us as teachers to activate inactive groups.
   R2 The teachers´ facilitating work was hard and took a lot of time.
   R3 It was possible to learn when groups collaborated through a forum system but students were not familiar with communicating electronically.
   R4 Students appreciated the freedom of time and place, and the confidence we showed them to choose methods.
   Based on these results, I formulated the following questions:
   Q 3 Is it possible to support the teachers by adding functions to the forum system?
   Q4 Is it possible to help students collaborate by giving them an introduction to human communication?
   Based on these questions, I designed support functions to be added to the forum system, and studied the use of these in a course. In the same course, I made an experiment by giving an introduction to human communication to some of the students (Study B).
   This study gave the following results:
   R5 It was possible to support the teachers with the functions tested, but those were maybe not ideal and more support might be helpful.
   R6 The experiment did not prove any noteworthy impact from the introduction about human communication.
   Thus, I decided not to continue along the line about the introduction, but to elaborate on the teachers role and need for support. I also decided to keep the focus on the online discussion as learning format. The resulting questions were:
   Q5 How can we take advantage of possibilities and reduce problems, introduced by the use of computer-mediated communication (CMC)?
   Q6 What help is available from theories about learning, collaboration, and communication, to understand processes and influencing factors?
   Q7 What help is available from practitioners within the field?
   Next step was to establish a working group around the theme: Creative Teaching of Electronic Collaborative Learning Groups. I co-chaired this group during the conference ITiCSE'99 and it resulted in a report (Study C).
   I have also experimented informally with different kinds of online discussions as components of a campus course I am responsible for. During and after the studies, I have read literature about theories and practice within the field. I used this knowledge to interpret results from the studies and to draw the conclusions reported in this thesis.

2.3 Approaches and Perspectives

Alternating between practical and theoretical studies, I have found new problems to solve and new possible solutions to try. Integrating my own results with theories and other researchers' results, I have synthesized the results reported in this thesis (see Figure 2.1.) By describing this process and my findings, I hope to stimulate other researchers and practitioners to continue the experimenting and analysis, shaping a good environment for growing knowledge.


Figure 2.1. Flow chart illustrating the research process


I have alternated between sharing the responsibility for a whole course and doing research on courses given by others or by myself. That means that my perspective has changed from at first having a focus on a course overview, to studying the single discussion contributions from individual students, and from being the observer, to experiencing being a teacher. In my interviews and dialogue analyses, I have tried to understand what it means to be a student or a teacher at a distance, and how the learning process is impacted by the teachers´ behaviour and the system used.

Describing the overall research process like this, it looks like a hermeneutic approach (see Figure 2.2.). There is an alternation between the whole and the parts, between pre-understanding and understanding, interpretation of meaning, the use of "growing" as a metaphor, the integration between theory and practice (Alvesson & Sköldberg, 1994).

    Figure 2.2. The "hermeneutic spiral" (inspired by
    Alvesson & Sköldberg, 1994, p. 174)

Hopefully, my knowledge has grown during these alternations, which indicates that I have followed a spiral going up instead of just going around in the circle.
   However, I did not use hermeneutical methods like source critique or text interpretation in my studies. My approach in the single studies can be described as inductive ethnography, largely built on observation of recorded behaviours. It is a primarily qualitative approach, studying phenomena and action in naturalistic environments with quantitative data collected when considered feasible and relevant. The qualitative approach requires a close and durable contact with the studied group or community in its natural environment, narrowing the gap between the observer and the observed. Often, one starts with a rather wide perspective, narrowing it as one finds interesting questions or details to focus on. It is important to be open-minded, flexible, and not restricted by theories or presumptions (Alvesson & Sköldberg, 1994).

2.4 Methods

Mainly qualitative methods were used in the studies. The reason was that my main goal was to create a deep understanding of factors influencing collaborative learning via CMC.
   One alternative approach would have been to compare a course given via CMC with the same course given in a traditional setting. Another alternative or complement could be to examine a great number of similar CMC courses and make a statistical analysis of factors that might influence the learning process. If a comparison between two courses should be meaningful, you have to control most variables. In a course situation, there is a huge amount of variables, and one of the most important is the student. It was impossible to conduct an experiment with the same students in two courses, and very difficult to experiment with so many students that the differences can be reliably eliminated by random assignment to experimental and control groups. Both in the studies reported and in the alternative approach mentioned above, it could have been interesting to measure the knowledge, resulting from the courses. However, the kind of knowledge aimed at here is very difficult to measure, and if you measure, you influence the knowledge. E.g., for the course examined in Study B, one of the knowledge goals is:
   To support the development of an analytical understanding and to awaken a lifelong interest in the social aspects of computerization.
   Thus, I have not tried to measure the knowledge resulting from the courses. Instead, I have judged the implications on learning outcomes from the students' self-assessments.
   Below, you will find a structured description of methods used in the three studies.

2.4.1 Methods of Data Collection

a) Data from relatively natural and spontaneous processes
   In Study A, behaviours indicating the learning and group processes were observed, and the experiences as active teachers were noted. Produced reports and the electronic communication, recorded automatically by the forum system, were saved.
   In Study B, the electronic communication was saved and the group processes were observed, without the present author taking part in the course.
b) Data produced on request or by the researcher
   In Study A, the students were asked to write diaries about their learning process. They also got three questionnaires each, and we made participant observations, as teachers.
   In Study B, an experiment was conducted and the students got three questionnaires each. Some students and the main teacher were interviewed.
   In Study C, there was a group discussion among researchers, documenting experiences and opinions in a report.

2.4.2 Methods of Data Analysis

a) Quantitative Processing
   In Study A, the electronic communication during one month was analyzed. Contributions were categorized and counted, in order to identify and compare different communication strategies. The questionnaires contained some quantifiable data, e.g., data about background and former experiences and these were processed.
   In Study B, the quantifiable data in the questionnaires were processed.
b) Identifying interesting cases
   In both Study A and Study B, examples of utterances and interaction were identified that illustrate different phenomena from the point of view of communication theory, activity theory, and theory about group processes.
c) Narratives
   Answers to open questions in questionnaires and interviews, experiences from the participation and observations in Study A and Study B were reported as narratives.

2.4.3 Methods of Inference

a) Induction
   In Study A and Study B, induction was used to generalize that some of the conclusions might be true also concerning other courses.
b) Theory-guided generalization/abduction
   In Study B, student behaviour in the CMC situation were found that seemed to agree with a corresponding behaviour in a face-to-face situation, according to a group-process theory. Thus, it was generalized that this behaviour might be typical also in other CMC situations.
c) Situated conclusion
   In Study A, some conclusions were drawn about the studied groups without generalizing to other groups.

2.5 Foundations for this Thesis

To conclude, this thesis is based on the following: