A shared document-based annotation tool to support learner-centred collaborative learning

A shared document-based annotation tool was presented, and its usefulness in two different real-life web-based university-level courses (adult learners, n=27 and adolescent learners, n=23) was empirically investigated. The study design embodied three data collection phases: 1) a pre test measuring self-rated motivation, learning strategies, and social ability, 2) log file data analysis showing actual use of the system features, and 3) a post test in a form of an email survey. For both groups the results showed that the level of motivation has a positive effect on activity in the system and the final grade. The learners, who reported to have good time management strategies, were the most active users of the system. The level of social ability predicted both number of consecutive comments in documents and threads in document-related newsgroup discussions. Log file data analysis showed that user activity in the system was positively related to the final grade in both samples. Results of the post test showed that all the respondents agreed when asked a) if the system brought added value to the learning process, b) if the use of the system changed their studying habits favourably, and c) if they would like to use the system in other courses. Introduction This study was conducted in order to present a shared document-based annotation tool, EDUCOSM (Kurhila, Miettinen, Nokelainen, Floréen and Tirri, 2003). In this study we empirically evaluate its usefulness in two different real-life web-based university-level courses with both adult and adolescent learners. Furthermore, we investigate how learner’s self-rated motivation, use of learning strategies, and social abilities are related to cognitive outcomes (final grade) and completion of various tasks (on-line group formation and peer-to-peer annotation of the course material) in the system. First we take a look at existing shared annotation systems and discuss about our approach to learner-centered collaborative learning. Next we describe the EDUCOSM system and the research methods applied in this study. Finally, we present the results and concluding remarks. Theoretical framework Shared Annotation There exists a large research body studying both personal and collaborative annotations in several domains. Previous research (Marshall, 1997) has shown that annotations made in books are useful to subsequent readers. We categorize annotation systems into two main groups: Document-centered and discussion-centered. Discussion-centered systems providing tools to browse and annotate discussion messages and threads are out of the scope of this paper. Next we discuss about non-commercial document-centered systems where learners are able to browse and annotate web-based documents. The first generation of annotation tools such as ComMentor (Roscheisen, Mogensen and Winograd, 1994) and Group Annotation Transducer (Schickler, Mazer and Brooks, 1996) required installation of platform-specific client-side software. The second generation allowed annotation of any document on the Web. Systems like CoNote (Davis and Huttenlocher, 1995) showed annotations embedded in the document, at the nearby position that they were made. We see two problems with the first and second-generation annotation tools: First, annotations are separated from the context visually by placing them into different browser frame or window, and second, annotations are not embedded in the document at the exact position that they were made. Some of the third generation creditable annotation systems, such as Kukakuka (Suthers and Xu, 2002), concentrated on design for artefact-centered discourse, ignoring the first issue of annotation context. We agree that there is no problem with long and threaded newsgroup discussion-like annotations, but we argue that simple one or two word comments should be placed in the context that elicited them. In the EDUCOSM system, user made annotations are shown in small tool tip windows that pop up on top of annotated area. Background colour of the area indicates the type of annotation: straw-coloured stands for highlighting and light green stands for comment. The second problem was answered by some of the third generation of annotation tools, such as CritLink (Yee, 1998) and Annotation Engine (Seltzer, 2000). Both systems allow embedded annotations, but CritLink places markers around annotated text phrase indicating exact sphere of influence, as Annotation Engine generates just one pointer for each annotation. Both systems allow anyone to add annotations to any document on the Web. EDUCOSM continues this development process introducing numerous intuitive features such as document highlighting and commenting with right clicking the mouse and selecting desired text area. Learner-centered learning Throughout the 1990’s learner-centered learning environments (SCLE) and computer-mediated communication (CMC) systems such as problem-based, project-based, cognitive apprenticeships, constructivist learning environments, and goal-based scenarios, have rather focused on the affordances they provide learners for effecting their way of learning and thinking, than transmitting information from teachers to learners (Land and Hannafin, 2000). Learner-centered learning is supported theoretically by various overlapping pedagogical concepts such as self-directed learning (Candy, 1991), learner-centered instruction or learning (Felder and Brent, 1996), active learning (Ramsden, 1992), vicarious learning (Lee and McKendree, 1999) and cooperative learning (Felder and Brent, 2001). For example, self-directed learning involves dimensions of process and product referring to four related phenomena: personal autonomy, selfmanagement, learner-control and autodidaxy (Candy, 1991). All these dimensions are present in the process of learner-centered learning where the locus of control is shifted from teacher to the learner who has now a greater responsibility for her own learning. The research field of computer-supported collaborative learning (CSCL) has three theoretical views on collaborative technology: (1) learner-centered principles, (2) constructivism, and (3) sociocultural theory (Bonk and Cunningham, 1998). The learner-centered principles (APA, 1997) present a similar shift from the traditional teacher-centered approach to learner-centered approach of instruction providing learner with valuable real-life skills (Dillinger, 2001). Learning tasks in learner-centered learning environments include such techniques as substituting active learning experiences for lectures, holding learners responsible for material that has not been explicitly discussed in class, assigning open-ended problems and problems requiring critical or creative thinking, and using self-paced and/or cooperative learning. The research findings of educational literature prove convincingly that properly implemented learner-centered learning fosters motivation and elicits deeper understanding toward the subject being taught (Felder and Brent, 1996; 2001; Dillinger, 2001). The idea of learner-centered collaborative learning in the context of this study is that learners are expected to take responsibility for their own learning: The instructor gives an orientation to the topic through theoretical face-to-face lectures. S/he also gives few pointers to selected on-line resources. The system provides tools to process information and collaborate with peer learners. The system allows the user to add new, and read existing html-documents published in the Internet. The users annotate the documents by highlighting and commenting selected target areas. Same target area is open for multiple annotations, allowing users to communicate asynchronously via annotations.