Patterns of scaffolding in computer-mediated collaborative inquiry

There is wide agreement on the importance of scaffolding for student learning. Yet, models of individual and face-to-face scaffolding are not necessarily applicable to educational settings in which a group of learners is pursuing a process of inquiry mediated by technology. The scaffolding needed for such a process may be examined from three perspectives: the organisation of activities, the affordances of tools, and process-level guidance. The purpose of the present study was to assess three tutors’ contributions to university students’ computermediated discourse organised as a question-driven inquiry process. In all, 17 students participated in the discourse as part of their studies in a cognitive psychology course. Quantitative and qualitative analyses were conducted to investigate critical patterns and possible effects of the tutors’ process-level scaffolding. During the process, the more experienced tutors acted as metalevel commentators, whereas the less-experienced tutor participated in the inquiry as a co-inquirer who also produced inquiry questions. More elaborate scaffolding to foster students’ metacognitive awareness of the inquiry strategies was generally lacking in all three tutors’ data. Implications for pedagogical improvement in collaborative, computer-supported educational settings are discussed. Patterns of scaffolding 2 Introduction The possibilities and challenges of computer-supported collaborative learning (CSCL) have been intensively investigated in recent educational research (Dillenbourg, 1999; Koschmann, Hall & Miyake, 2001; Lehtinen, 2003). The CSCL approach is built on social constructivist and sociocultural perspectives of learning, and it can be defined as a pedagogical approach in which participants collaborate through the computer in a learning situation, supported by several network-based collaboration tools (Lehtinen, 2003). According to many studies, advanced collaborative technology can be beneficial for students’ learning, but it does not by itself guarantee successful results; appropriate pedagogical arrangements and guidance during the collaborative learning endeavour are also required (Dysthe, 2000; Hara, Bonk & Angeli, 2000). Most studies on CSCL have, however, concentrated on students’ activity; the instructor or tutor’s role has been investigated less often. Puntambekar and Kolodner (1998) argued that, in complex learning environments, support for learners should be distributed across the various agents that play a role in a learning process. Probably the most effective approach is an integration of various modes of support— both those embedded in the pedagogical approach, tasks, and tools, as well as those manifested in the human tutors’ situation-specific guidance. With reinforcement from the literature, we suggest three complementary pedagogical levels to be addressed when discussing support for students in CSCL settings: 1. The task organisation level, which means the initial “framing” or organising of the learning group’s activities, or structuring the task according to the chosen pedagogical approach (Bielaczyc, 2001; Wells, 2000); 2. The tool level, which means built-in structures or software tools in the educational technology environments for structuring and directing students’ work (Reiser, 2004); 3. The process level, which involves human coaching, situation-specific guidance, and expert participation during the collaborative activity (Rourke, Garrison & Archer, 2001). The present study concentrates on this third level while recognising that the levels are intertwined. According to a number of researchers (Edelson, Gordin & Pea, 1999; Hakkarainen & Sintonen, 2002; Scardamalia & Bereiter, 1994), organising the collaborative process as question-driven inquiry provides methods for facilitating students’ learning and knowledge building efforts in CSCL. In practice, institutional and curricular constraints often diminish the possibilities of arranging university courses mainly according to high-level inquiry principles. For instance, in the Finnish higher education system, university courses are usually arranged as parallel, semester-long courses run by one instructor, with two lectures each week, and a final Patterns of scaffolding 3 exam in the end; thus students’ questioning has hardly any influence on course content or direction. Therefore, it would also be desirable to find ways to enrich existing educational practices by bringing in elements of higher-order inquiry into traditional lecture courses. Background and context In an earlier study (Muukkonen, Hakkarainen & Lakkala, 1999), we investigated undergraduate students’ engagement in a cognitive psychology course in which lecturing was supplemented, between the lectures, by computer-mediated collaboration. Collaboration was organised as question-driven inquiry discourse, and 13 undergraduate students from a Finnish university were provided with web-based asynchronous software that included built-in scaffolds to support collaborative inquiry. The study results provided evidence that the pedagogical arrangements and the tools used stimulated the students to produce questions and elaborate explanations about the topic. Yet the discourse was not very deep; elaboration of students’ own explanations into more theoretical explanations was only weakly evident in their work. In a subsequent study (Muukkonen, Lakkala & Hakkarainen, 2001), the present researchers analysed students’ engagement in a similar lecture course in which there were also tutors participating in the students’ computer-mediated inquiry discourse between the lectures. In this course, the general pattern of all the students’ knowledge productions was very similar to the previous course: The students produced many problems and, especially, their own explanations for the problems, but not as many elaborated scientific explanations as were expected. There was, nevertheless, some variation in separate groups’ knowledge productions, which provoked our interest in analysing the tutors’ activity in the course. Three separate groups, including the tutors, took part in the course, which provided an opportunity to examine the influences of the tutors’ process level scaffolding, since the organisation of activity and tools were the same for all groups. We will begin by providing a background for the research through an examination of earlier studies of scaffolding and tutoring. From individual scaffolding to tutoring virtual communities of inquiry As Brown, Ash, Rutherford, Nakagawa, Gordon and Campione (1993) stated in the context of discovery learning, tutors should take the difficult role of being in the middle ground of guided inquiry learning. They should not rely too much on students’ spontaneous inquiry and should intervene if students are not able to make progress on their own. On the other hand, too much control prevents students’ self-directed learning activity and restricts their cognitive efforts. These challenges can be related to choices that must be made during process guidance. Patterns of scaffolding 4 Questions to be asked include, for instance, whether the tutor should take an active role as an expert model in contributing to the process, or allow the students to work as much as possible for themselves. Also, how do tutors give guidance that helps the learners assume responsibility for cognitively more challenging tasks in the inquiry? The concept of scaffolding was developed in order to define what kind of help should be given to a learner and was first introduced by Wood, Bruner, and Ross (1976) in their investigation of an adult helping a learner in a task that is too difficult. According to them, scaffolding included such elements as arousing the learner’s interest in the task; reducing the complexity of problem to suit the learner’s level of expertise; directing the learner’s activity towards the task goals; highlighting the critical features of the task; and modelling the solution to a task. Most subsequent studies of scaffolding or successful expert tutoring (e.g., Lepper, Drake & O ́Donnell-Johnson, 1997; Wood & Wood, 1996) were based on individual tutoring situations. Yet, Kolodner (2001) stated, models of individual scaffolding are not necessarily applicable to educational settings in which a group of learners is pursuing a common goal. Another common feature of the above studies of scaffolding and tutoring is that the tasks that the learners faced were rather formal, structured problems that had known, correct solutions. In instructional settings following question-driven inquiry, by contrast, the problems are often open-ended; students are treated as experts who have knowledge to share, and they are encouraged to create (what is for them) new explanations (Hakkarainen, 2003; Scardamalia & Bereiter, 1994). In such an approach, the purpose is to promote students’ cognitive advancement and metacognitive awareness, with support given only when necessary, in order to coach the students gradually to take up responsibility for higher-order aspects of learning (Bereiter & Scardamalia, 1987; Hogan & Pressley, 1997; Ligorio, Talamo & Simons, 2002). In addition, if the communication is mediated by technology, it further changes what is appropriate in scaffolding and tutoring. Tutors have to learn new ways to support students’ work ‘virtually’, through asynchronous or synchronous communication channels, compared to more familiar ways in face-to-face contact (Lakkala, Ilomäki, Lallimo & Hakkarainen, 2002; Ligorio et al., 2002; Russell & Perris; 2003). Previous studies and models of teachers’ or tutors’ contributions to group learning or collaborative inquiry were based on face-to-face group instruction or classroom settings (Levin, 1999; Mercer & Fisher, 1993). The teacher is very often at the centre of discussions in traditional classrooms, and collaborative technology does not necessarily, by itself, change the situation for the better. Guzdial (1997) reported findings about the teachers’ and the tutors’ dominant roles in college courses in which a CSCL forum was used. In eight courses out of ten, it was the teacher or Patterns of scaffolding 5 teaching assistant who produced the highest number of notes; also, the second most productive writer was the teaching assistant in three of the ten courses. Lahti, Seitamaa-Hakkarainen, and Hakkarainen (2003) found that the teacher mediated a great deal of interaction between the students in a university-level design course following progressive inquiry. This finding addressed the teacher’s cognitive centrality, which refers to a person’s crucial role in sharing knowledge with other members of a group, and thus relates to a person’s social status (Stasser, 1999); Lahti and her colleagues raised questions about the inquiry-hindering effects of such centrality. Purpose and scope of the study The focus of the present study is on the tutors’ contribution to students’ computermediated collaborative inquiry. The setting was a 4-month, undergraduate university course in cognitive psychology, in which students pursued collaborative inquiry in three separate groups, using a web-based groupware system between weekly lectures, with the support of a tutor participating in each group. The organisation level scaffolding was provided by structuring the computer-mediated activity following the ideas of progressive inquiry, a pedagogical framework developed by Hakkarainen and his colleagues (Hakkarainen, 2003; Muukkonen, Hakkarainen & Lakkala, 2004) for facilitating knowledge-building practices in educational settings. The progressive inquiry model posits that students’ genuine questions and previous knowledge of phenomena are a starting point of a deepening process, in which students explain phenomena, share their expertise and build new knowledge collaboratively with the support of technology and knowledge sources. The progressive inquiry framework is based on theories of knowledge building (Scardamalia & Bereiter, 1994), an interrogative model of scientific inquiry (Hintikka, 1982), and concepts of distributed expertise (Brown et al., 1993). The role of the instructor in the investigated course was a pedagogical organiser and content knowledge expert; he gave the weekly lectures, as in a typical undergraduate university course, but did not participate in the computer-mediated activity. The process-level scaffolding for the students was arranged by providing each student group with a tutor. The tutors were post-graduate students in educational psychology with good skills in using information and communication technologies (ICT) but varying expertise in teaching. Their role was to keep the discourse active, deepening, and focused on the generated research questions. They were also expected to encourage the students to build on each other’s ideas and to plan, monitor, and evaluate the inquiry process themselves. Patterns of scaffolding 6 The questions addressed in the present study are the following: 1) How do the tutors participate in university students’ computer-mediated progressive inquiry discourse; 2) How does the nature of the tutors’ scaffolding influence the students’ engagement in inquiry; and 3) What recommendations might be appropriate, based on the results, to develop models and practices for scaffolding computer-mediated collaborative inquiry? Research methodology Course setting The investigated course was a 15-week, undergraduate university course, Psychology of Learning and Thinking II. It yielded two-credits to complete a ten-credit minor unit in psychology offered to degree students at the University of Helsinki. The course consisted of weekly 2-hour lectures and tasks between the lectures. Eighty students took part in a course, but they were divided in two conditions; 17 of them chose to use the Future Learning Environment (FLE) between lectures for collaborative inquiry. This study is based on an analysis of these 17 students’ and the three tutors’ written postings to FLE’s common database. The rest of the students were in the other condition; they did not use any groupware, but participated in the course in a more traditional format, following lectures and writing learning logs. Comparison of all students’ knowledge productions in the two conditions has been reported in another study (Muukkonen et al., 2001). The tool used in the investigated course, Future Learning Environment (FLE), is an asynchronous groupware system designed for supporting knowledge building and progressive inquiry in educational settings (see http://fle.uiah.fi). FLE is an Open Source and Free Software, and its development is continuing (see http://fle3.uiah.fi). The Knowledge Building module provides threaded discourse forums for working together to solve problems and develop common ideas. In the investigated course, a new forum was founded for each problem formulated in the first lectures, and the students were guided to continue the discourse by writing more specific problems, explanations and scientific knowledge under each problem. The notes were visible to all members in the same study group. In the discourse forums, each note had to be categorised by choosing a scaffold label corresponding to the progressive inquiry elements (Problem, Working theory, Deepening knowledge, Comment, Metacomment, or Summary); similar categories had been employed in Computer-Supported Intentional Learning Environment (CSILE; see Scardamalia and Bereiter, 1994). The system also provided each participant with a Virtual WebTop for storing and sharing documents. During the investigated Patterns of scaffolding 7 course, the WebTops were used to share study materials, articles and editable documents; the contents of these documents are not analysed in the present study. As already mentioned, the course was enriched by the elements of progressive inquiry. During the first two lectures, all the students were guided to formulate research problems. Initially, they produced the research problems individually, continued by discussing their problems with a peer, and then engaged with a small group. Each group presented the most interesting questions to all the participants in the lecture, e.g., “What is intelligence and how can it be measured?” Finally, the teacher grouped the questions together in three lists, consisting of 3-4 questions, to be further studied in FLE course in three groups. The students themselves chose the questions to investigate (and hence the group in FLE). The groups consisted of four to seven students. The requirement for course credit was to contribute actively to the progressive inquiry process in FLE by writing one’s own questions and explanations, and by commenting on the productions of the other group members. Information sources used were the weekly lectures and scientific articles distributed by the teacher. At the end of the course, the students were also expected to write a summary of their own contributions and learning process. We provided each group with a tutor participating in their FLE-work, as the chosen means of scaffolding the inquiry process. The tutors were acquainted with the principles and goals of progressive inquiry, but they did not receive any explicit or specific guidance about the ways that they might contribute to the FLE discourse. The tutors did not meet the students in person during the course. The teacher in the course, who gave the lectures, sent one starting note to each FLE group (including general instructions for working), but did not otherwise participate in the FLE-work. The teacher’s contribution to sharing theoretical and scientific information to the students was crucial. The students had lively conversations with the teacher during the lectures, but these were not recorded or documented in the study. Participating students and tutors The students in the course were undergraduate students from many fields of study at the University of Helsinki, including forensics, mathematics, history, languages, and education. They took part in this two-credit course to complete a ten-credit minor unit in psychology. They had a previous course, Psychology of Learning and Thinking I, before enrolling for this course. The mean grade (on a three point scale, with 3 being the highest) from the prior course for the 17 members of the FLE groups was 1.8 (SD = .72). There was no statistical difference in the mean grades of the other students in the same course (M = 1.9, SD = .68); therefore, we Patterns of scaffolding 8 concluded that the students in the FLE-groups were not exceptional students in that sense. The age of the 17 students varied from 21 to 30, with a mean of 24 years. Year in studies varied from 0 to 6 years, the average being 3 years. The majority of students taking part in the course were females, which was also reflected in the gender distribution of the investigated FLE groups: 5 males and 12 females. The three tutors were post-graduate psychology students, all females, who were members of the research group conducting the present research. The tutors were of ages 30 (Tutor 1), 39 (Tutor 2) and 26 (Tutor 3) years old. Tutors 1 and 2 were competent in using ICT and had substantial teaching experience as lecturers in adult education. They also knew the theoretical basis of progressive inquiry very well, and Tutor 1 had participated in the design work of FLE, but they had not previously participated in tutoring a progressive inquiry process. The youngest tutor was familiar with computers and the basics of cognitive psychology and progressive inquiry, but she did not have earlier teaching or tutoring experience; she was a newcomer in the research group. Data collection and analysis The material produced by the three study groups to FLE’s discourse forums constituted the data analysed in the study. The material was retrieved for analysis so that all the posted notes were arranged in hierarchical order based on the reply structure; the first notes of each thread were listed in chronological order. The course was in Finnish; therefore, all the text examples presented in the present article have been translated into English. Table I is a summary of the data used. Table I. Summary of the database material used in the study Group 1 Group 2 Group 3 Total Number of students 4 7 6 17 Discourse forums in FLE 5 4 5 14 Produced notes 56 87 95 238 Text pages in the data file 24 42 34 100 a The knowledge produced by the tutor in each group is also counted in the numbers of the table. The number of text pages was counted from a file, which included all the database material accumulated to the database of each group’s discourse (writing times of notes, note headings, authors’ names, inquiry scaffolds chosen, and the text of each note). Descriptive statistics of the groups’ discourse engagement. To get an overview of the variation in the study groups’ discourse engagement, several measures were computed from Patterns of scaffolding 9 each group’s data: the number of the postings, the number of starting notes and reply notes, and the usage of inquiry scaffolds. An examination of the content of the notes indicated that they usually consisted of several ideas. In order to analyse the richness of the produced knowledge, the notes were segmented into propositions, each of which was considered to represent a separate idea. To analyse the reliability of the segmentation, an independent coder classified, similarly, approximately 5% of all notes into segments. The inter-coder reliability (single measure intraclass correlation; see McGraw & Wong, 1996) was .88, indicating that the reliability of segmentation was satisfactory. The teacher’s three starting notes (including general instructions for working) were not included in the analyses. Analysis of the content of the tutors’ scaffolding. In order to get an overview of the tutors’ scaffolding, the content of the tutors’ notes was analysed, in detail, by methods of qualitative content analysis (see Chi, 1997). The same segmentation of note texts into ideas was used as in the earlier analysis, and each segment was categorised according to the object of scaffolding it represented. The categories were derived from several preliminary analyses of the tutors’ notes; the following four categories were used in the final classification: 1. Ask explanation-seeking question: Ideas assigned to this category were research questions into the inquiry process, questions about refocusing the discussion or questions directed explicitly to students to elaborate their explanations. 2. Produce expert’s explanation: Ideas in this category represented the tutors’ own explanations of the problems discussed, or references to scientific explanations or theories. 3. Review and evaluate the discourse: Ideas assigned to this category included summaries or reviews of the discourse, references to students’ contribution, positive and supportive evaluation, or unmodified quotations from students’ texts. 4. Recommend study practices: Ideas assigned to this category represented instructions for working in the course or in the inquiry process, and technical advice about FLE. In the content analysis, each segment or idea in the tutors’ notes was classified to one of these four scaffolding categories. All of the ideas fitted into these four categories of ideas, which were set up as mutually exclusive. To analyse the inter-coder agreement of the categorisation, an independent coder classified approximately 40% of ideas in the tutors’ notes. The coefficient for coder agreement (Cohen’s Kappa) was .89, which is indicative of a clearly stronger congruity than chance alone would produce. Analysis of the discourse progression. In the third phase of data analysis, more exploratory methods were used to evaluate the possible influence of the tutors’ scaffolding practices on the students’ engagement in inquiry. For this purpose, the unit of analysis was Patterns of scaffolding 10 extended to cover the entire material for each group to enable a process perspective on the discourse. Two researchers read the data several times and formed their individual opinions of it. After that, characteristic differences were jointly identified through discussions between the researchers. Consequently, the process analysis presents a synthesis of what were considered as the most influential features in the process-level scaffolding in the three groups.