Projects and problems are being increasingly used to provide meaningful contexts for learning (Blumenfeld et al., 1991; Collins, Brown, & Newman, 1989). These problems are often complex, necessitating that students work in collaborative groups in order to distribute the cognitive load. In addition, collaborative groups have the potential to provide opportunities for reflection as the group engages in sustained dialogues. Reflection on problem-solving and project work is critical if students are to learn from these activities, rather than simply completing tasks (Collins, Brown, & Newman, 1989).
But effective collaborative learning does not just happen, as we learned in several early attempts to support student collaboration and reflection with CaMILE, the Collaborative And Multimedia Interactive Learning Environment. We found that students often focus on completing their course-related tasks and fail to learn what they might from their collaborators (Newstetter & Hmelo, 1996). This paper will report the evolution of CaMILE and the strategies that have developed to make this into a valuable forum for reflection and collaboration.
CaMILE was developed to help students collaborate and to encourage useful reflection. This environment provides a forum similar to a newsgroup but allows students attach and link to multimedia documents and provides scaffolding for collaboration and reflection (Guzdial, Turns, Rappin, & Carlson, 1995). CaMILE supports collaborative discussions as students create threaded notes. The software provides prompts (procedural facilitations) to help the learners consider the role of their notes in establishing and maintaining an ongoing dialogue. Students reflect on their thinking as they choose the type of note they enter. This is an approach to support metacognition about the role of the student's note in the discussion (Scardamalia, Bereiter, McLean, Swallow, & Woodruff, 1989).
Students use this environment as a tool at times when they need to collaborate to complete a project and when it is integral to an assignment. They can share documents by attaching them to the their notes and can create links to the a variety of multimedia documents and to resources on the World Wide Web. When students click on the attached documents the document and its application is opened allowing students to work collaboratively on shared documents. By encouraging reflection as students collaborate, the students are more likely to construct usable knowledge and to transfer what they have learned to other problem-solving situations (Salomon & Perkins, 1989).
We first tried to use CaMILE in two courses: a mechanical engineering design course and a course in sustainable development and technology. These initial attempts to use CaMILE in two engineering courses were not very successful (Hmelo, Vanegas, Realff, Bras, Shikano, & Guzdial, 1995; Turns, Guzdial, Mistree, Allen, & Rosen, 1995). The students did not readily use the system. Students either did not use CaMILE or if they used it, it was not the in way that the course instructors had hoped.
Our first exploration of collaboration in support of project-based learning was in a mechanical engineering class. ME3110: Creative Decisions and Design at Georgia Tech is a required introduction to design for juniors. Students work in teams of four to six (seven to ten teams per class), where all teams are working to design and construct a complex mechanical artifact. For example, one class had to create a vehicle to climb a vertical length of pipe and then pop balloons attached to the top of the pipe. The problem and the design are too complex for an individual to solve in a single quarter (ten weeks), but simple enough for a team to complete.
The instructors of ME 3110 approached us to help solve some problems that they perceived in the class. The instructors felt that students needed support in maintaining context on their problem outside of class. In addition, because students had busy schedules, they complained about the difficulties of meeting and of working together. We hypothesized that technological support for the student groups might help students maintain continuity in the problem context, and provide a means of working together outside of class and face-to-face meetings.
The first version of CaMILE (used in Fall '95 in ME 3110) was a Macintosh-based program (Figure 1). Students created notes in threaded discussions. When students created a note, they were asked to identify the kind of note that they are creating (e.g., a rebuttal, an alternative idea, a new idea, a question) and then they were offered suggestions of phrases that are useful in a note of that type (via the procedural facilitations). As noted earlier, students using CaMILE could easily attach media of any sort, which appeared as icons in the margin of their note. We hypothesized that students would use CaMILE to communicate with one another outside of class and would attach the results of their work with the various design tools to share with the other members of their group. The threaded discussion would help to provide context as well as helping the students integrate the results of different phases of the design activity as they shared their documentation.
Figure 1. First Version of CaMILE: Macintosh-based
We used the Macintosh-based version of CaMILE for two quarters in ME 3110, but our recognition that something unexpected was happening came along very early. Students were hardly using CaMILE, except for when they were required to use it for assignments that we designed to encourage use. The notes that they were posting often had little content, e.g., "I agree" as a complete note. Students were not using it in place of face-to-face collaboration - in fact, they continued to complain about the difficulty of scheduling and having meetings. There were very few documents being posted in the notes, and students were not reading those that were posted.
We conducted interviews with a sample of the students, and used surveys (based on these interviews) for getting broader response to our questions. We decided that there were several factors causing the lack of use of CaMILE, but the most significant factor was that students in the same group saw one another much more often than we anticipated.. As one student put it, "I'm in a group with my best friends. We take three or four classes together a quarter. Why do we need to talk over the computer?" The problem with scheduling and having meetings was caused not just because the students because the students had scheduling conflicts but simply because they were too busy. In addition, there were problems with the reliability and availability of the software. Many of the students owned Windows computers and CaMILE was only available for MacIntosh, necessitating that the students use the software on campus. Occasionally, there were problems with the server that would cause the notes to be lost.
A second early attempt at using CaMILE was in a problem-based course in Sustainable Development and Technology. This was an interdisciplinary engineering course in which students learned to design solutions to problems in ways that were friendly to the environment (Hmelo et al., 1995). The students were asked to use CaMILE to collaborate as they worked on their problems and projects. We expected the students to use the software to exchange documents and discuss issues they were grappling with. Throughout the course, students were encouraged to use CaMILE to assist them in their collaborative efforts by encouraging them to post their questions and to engage in dialogue around these questions and their collective problem-solving efforts as they worked on group projects. In fact, this rarely occurred. An occasional note was posted by students in order to share resources (e.g., articles and books) with other groups but threads never exceeded two notes (and the second was usually written by one of the course instructors in the hope of engendering dialogue). Our goals for using the software were not achieved for a variety of reasons: (1) access problems, (2) hardware platform incompatibilities, (3) failure to adequately integrate the collaboration technology into the course. To determine the reasons that students did not use CaMILE, we analyzed the logs that the students were asked to maintain. At this time, we were still using the Macintosh-based version of CaMILE. This caused two types of problems.
The lack of explicit integration of CaMILE into the course was probably the major reason for our lack of success in using CaMILE. Students were encouraged rather than required to use CaMILE. Even if they had been required to use it, the benefits to the students were not made clear. In the students' logs, they noted how difficult it was to collaborate and to meet outside class. They also noted that it was hard for them to do more than attach their documents together the day that a paper or case study was due. The students were focused on getting their particular piece of the project completed rather than learning from each other. The reports that the students handed in were consistent with the lack of integration: often the same information was repeated in sections written by different students or worse, they may have contradicted each other. Despite the collaboration difficulties, the students had no idea how CaMILE might be of benefit.
From these early failures, we learned many lessons related to both the technology and the way that the technology is integrated with the curriculum. These lessons are simple but critical for successfully using computer-supported collaborative learning technology. First, we learned that students need to have something to talk about and that this something needs to be made explicit. We have had subsequent success by embedding CaMILE discussions in or attaching them to interesting contexts (Figure 2). We refer to this as anchored collaboration (Guzdial & Turns, 1996). In this mode, CaMILE discussions are reached via one-click links from Web pages, that are relevant and thought provoking. In the example discussed below, a CaMILE discussion may be attached to images of the students' designs for commentary. Second, students need to have ubiquitous access to the hardware. By redesigning CaMILE to run on the World Wide Web, we dealt with the technology issues.
We redesigned our effort by changing our goals, the technology, and how we integrated the technology with the curriculum:
Goals: Sustained discussion where the topic of the discussion frequently addresses classroom learning issues and curricular objectives.
Software: CaMILE was re-designed as a Web-based application (Figure 2). As a Web-based version, the new CaMILE was available cross-platform, so students could use it from their dorm rooms on their own computers. Further, we added a new capability that was not available in the Macintosh-based CaMILE. Each note in Web-based CaMILE was accessible at a distinct address (URL) which could be referenced from another Web page as a link. We were now able to create anchored collaboration (Guzdial & Turns, 1996), where a discussion on a given topic (anchor) was accessible through a simple hypertext link.
Curriculum: We combined this new capability with a new emphasis in our description to the students of the role for CaMILE. Rather than using CaMILE for within-group discussion (which was still acceptable and encouraged), we used CaMILE for between-group and within-class discussions. For example, we used CaMILE as a forum for allowing ME 3110 student groups to review and critique other groups' interim design reports midway through the quarter. Each report was placed on the Web, with a link to a note heading a thread focused on discussing just that report.
Facilitation: Although students initiated some of the discussions, course instructors needed to provide some structure as well. This may be as simple as setting up a Web-page such that there are places for different discussion topics or more complex such as scanning in and posting images that might serve as contexts (Figure 2 and 3)
The results for use of Web-CaMILE in Mechanical Engineering have been quite different. The goals of using CaMILE were now focused on student projects (Figure 2). This was accomplished using our new framework. Because the projects were an integral part of the curriculum, and the basis for the students grades, our goals changed to having students engage in a sustained discussion centered around the projects. In these discussions, the students wrote much more substantive notes. While they were not excited about using CaMILE per se, they were interested and excited in the activities that we were providing for them. Competing design groups were eager to see one another's designs, as well as to critique (and defend) the designs. We feel that students are getting a better appreciation for the cohesiveness and integration of the components of the design process by contrasting work across multiple groups.
Figure 2. An Anchored Collaboration using WebCaMILE in
Mechanical Engineering: Student report is on the left, thread
of notes on right.
We have also overcome the issues of software access and reliability that plagued the earlier CaMILe trials through the Web-based implementation. The course faculty facilitated the discussions by setting up Web pages with images scanned from the students' designs and provided links into the CaMILE discussions. This helped to provide concrete contexts for discussion. This new approach has had success in other courses as well.
A later test of Web-based version of CaMILE was conducted in a graduate class in Cognition and Instruction. The students represented a number of disciplines: industrial engineering, civil engineering, architecture, computer science, and psychology. They shared a common interest in developing innovative learning environments but had little knowledge of the theoretical and research base outside their own specific discipline. In order to encourage the students to apply what they were learning, they worked in interdisciplinary project teams to write a proposal for the development and evaluation of a learning environment.
Figure 3. An example Web page with a scanned image (right)
used as an anchoring context to a WebCaMILE discussion (left):
Students critiquing each others projects in Cognition and Instruction.
The intent was to have CaMILE integrated into the students' activities throughout the course. By creating a virtual place for the students to communicate outside the classroom, the instructor's intent was to have the students continue their discussions of the issues from class. The instructor had several goals for using the system:.
Promote reflection on topics relevant to the curriculum.
Encourage exploration of issues from different perspectives. Because of the interdisciplinary nature of the group, it seemed likely that the students would have different views on the issues and projects
Discussions of project-related work. The students would use CaMILE to post their initial ideas that would eventually coalesce into a small number of project teams. CaMILE would serve as a forum for the students to post and get feedback on their design ideas.
The instructor facilitated the discussion by setting milestones for the students to post their introductions, formation of project teams, and initial and final proposals. In addition, the notes from the blackboard were posted after each class. Links to various discussions were provided from the class home page.
Although there were several threads on the class discussions, a post-course questionnaire indicated that the students found the discussions that were anchored in their projects the most useful. The students initially posted their ideas and tried to encourage other students to join them to form teams. The discussions were generally productive but there were also some unintended consequences. As the groups formed, the discussions were used for project work (although this was largely by one group). The students took the role of critiquing each others proposals very seriously as was evident in their contributions to the discussions (Figures 3 and 4). The projects themselves provided a concrete anchor for the discussion of issues in cognition and instruction. These were not often articulated in the abstract so students rarely followed through on the issues discussed in class. For example, students did not have protracted discussions about transfer but issues related to transfer came up in the discussions of the projects as in this comment that one of the students made: "In the evaluation section, you said that you would try to determine whether the transfer was mindful or un-mindful in nature. Has one proved to be more effective than the other?
Is unmindful transfer not considered successful transfer ? I'm
a little confused ..."
Figure 4. Critique and Review of Proposals using WebCaMILE in Cognition and Instruction
In general, we have found that anchored collaborations are useful in many classroom contexts. In most classrooms (especially project-based classrooms), there are artifacts to discuss or critique: from the latest open-ended design assignment, to peer review of student work. Because CaMILE facilitates anchored collaboration, it is now being used in several classes around Georgia Tech's campus. The importance of anchoring the discussion in concrete instances and setting up meaningful goals for the use of the system is extremely important.
In terms of lessons learned about computer-supported collaborative learning, we gained an appreciation for the complexity of designing for a classroom context. Students, at least in higher-education, need a good reason for using a computer-based collaboration tool. By designing and re-designing both the tool and the contexts for use, we have come to an understanding of the affordances needed to make computer-supported collaborative learning acceptable and useful to students in higher education.
At a meta-level, an important lesson from our experience is that a good mistake can be a much more powerful learning experience than an immediate success. Our first trials with CaMILE helped us develop the notion of anchored collaboration, which has been a powerful concept in our later work. Experimental evidence gathered from more still more contexts of use suggests that anchored collaboration is more effective than traditional collaboration supports (i.e., newsgroups) in encouraging on-line collaboration. In a recent comparison of two sophomore computer science classes, one using news groups and the other using CaMILE, the CaMILE-using class wrote many more notes in longer threads. This suggests that students who use CaMILE engage in more sustained discussions, one of our goals for the system. It is not surpassing that creating a space for students to talk about and something to talk about should lead to extended discussions. However, we have found that collaboration supports must be designed to encourage this happy coincidence to occur.
Blumenfeld, P. C., Soloway, E., Marx, R. W., Krajcik, J. S., Guzdial, M., & Palincsar, A. (1991). Motivating project-based learning: Sustaining the doing, supporting the learning. Educational Psychologist, 26, 369-398.
Collins, A., Brown, J. S., & Newman, S. E. (1989). Cognitive apprenticeship: Teaching the craft of reading, writing, and mathematics. In L. B. Resnick (Ed.), Knowing, Learning, and Instruction: Essays in Honor of Robert Glaser (pp. 453-494). Hillsdale, NJ: Erlbaum.
Guzdial, M., Turns, J., Rappin, N., & Carlson, D. (1995). Collaborative support for learning in complex domains. In J. L. Schnase & E. L. Cunnius (Eds.), Computer Support for Collaborative Learning '95 (pp. 157-160). Hillsdale NJ: Erlbaum.
Hmelo, C. , Vanegas, J., Realff, M. Bras, B., Mulholland, J., Shikano, T., & Guzdial, M. (1995). Technology support for collaborative learning in a problem-based curriculum for sustainable technology. In J. L. Schnase (Ed.), In J. L. Schnase & E. L. Cunnius (Eds.), Computer Support for Collaborative Learning '95, (pp. 169-172) . Hillsdale NJ: Erlbaum.
Newstetter, W. C., & Hmelo, C. E. (1996). Distributing cognition or how they don't: An investigation of student collaborative learning. In D. C. Edelson & E. A. Domeshek (Eds.), Proceedings of ICLS 96 (pp. 462-467). Charlottesville VA: AACE.
Salomon, G., & Perkins, D. N. (1989). Rocky roads to transfer: Rethinking mechanisms of a neglected phenomenon. Educational Psychologist, 24, 113-142.
Scardamalia, M., Bereiter, C., McLean, R., Swallow, J., & Woodruff, E. (1989). Computer-supported intentional learning environments. Journal of Educational Computing Research, 5, 51-68.
Turns, J., Guzdial, M., Mistree, F., Allen, J. K., & Rosen,
D. (1995). I wish I had understood this at the beginning: Dilemmas
in research, teaching, and the introduction of technology in engineering
design courses, Proceedings of the Frontiers in Education Conference
. Atlanta, GA.