This study examined online group discussion from students' perspectives to determine what characteristics students identify as meaningful to their online learning. Participants were 142 graduate students in the Southeast. The multiple regression results revealed that quality of online discussion was positively associated with communication media, computer experience, discussion interest, and strategy for preparation. Males reported statistically higher scores in quality of discussion. In addition, quality of discussion was negatively associated with group partner.
Introduction
Interactions between students and their instructor and among the students themselves are significant to the process of online learning (Palloff & Pratt, 1999). Research indicates that small groups facilitate learning as compared to individual learning (e.g., Bruffee, 1999; Du, Zhang, Olinzock, & Adams, 2008; Johnson, Johnson, & Stanne, 1986), and that peer group work has significant impacts on varied learning outcomes in both face-to-face and online learning environments (e.g., Bruffee, 1999; Harasim, 1990; Scardamalia & Bereiter, 1996; Uribe, Klein, & Sullivan, 2003). As a new learning method, however, a caution should be added that using technology can result in students becoming bored, inattentive, or even frustrated with the online discussion experience (Berge, 1999), and many instructors have indicated a lack of student participation in online discussions (Jin, 2005). Being very different from traditional learning through face-to-face communication, misunderstanding and miscommunication are more likely to happen and are also detectable in an online environment. In addition, online communication technology is relatively new as an educational tool, and learners may experience a learning curve with the technology as well as with the learning method.
Research and instructional efforts have been made to deal with challenges faced by students and instructors in online learning environments and to facilitate successful online collaboration (e.g., Bonk & King, 1998; Clark & Mayer, 2003; Harasim, 1990; Kaye, 1991; Mason, 1991). The purpose of this study is to examine online group discussion from a student's perspective to determine what characteristics students identify as meaningful to their online learning and develop a deeper understanding of the dynamics of online discussion.
Related Literature
Collaborative learning is not a new idea in education and the benefits of online collaborative learning have been widely researched (Roberts, 2004). Using group work as an instructional strategy has been a specific focus within the area of collaborative learning (Bonk & King, 1998; Koschmann, Hall, & Naomi, 2002). However, few studies have examined the details of group discussion (Thompson & Ku, 2006). Faculty often uses group projects and discussions to engage students in a cooperative and/or collaborative learning environment. In examining group dynamics in an online environment, Fisher, Thompson, and Silverberg (2005) indicate one of the strengths of group work is that it helps a student explore his or her thinking, providing opportunities for knowledge construction with their peers. However, distance learners have indicated experiencing a sense of social isolation (Lally & Barrett, 1999). This sense of isolation can be addressed by having group members work together in unique ways, providing opportunities for students to attend to the academic and social components of the online class (Gabelnick, MacGregor, Matthews, & Smith, 1990). Groups are complex systems that are dynamic and adaptive (McGrath, Arrow, & Berdahl, 2000). As complex systems, groups can also be investigated from a systems perspective. A systems perspective recognizes and studies every component in terms of how that component affects the system and how the system affects each component (Carabajal, LaPointe, & Gunawardena, 2002). With online groups there is the additional component of the technology tools, which can not be ignored when examining online groups (Fisher et al., 2005; McGrath et al., 2000).
Another important component to groups and online discussions deals with the preparation of the group members. Preparation by group members is an important component for successful group discussion (Curtis & Lawson, 2001; Petress, 2004). Jonnasen (1996) refers to computer conferencing as a “mindtool” that prompts a larger amount of reflection and analytical thinking while still connecting learners. Students have found group projects more rewarding when they were actively involved in the pre-planning, reading, and implementation (Fisher et al., 2005).
Du et al. (2008) found using writing as a way to prepare for small group discussions provides opportunities for rich learning experiences. Cashman, Gunter, Gunter, & Shelly (2004) used “assigned conversation,” which was a focused study of reading assignments, and found that this method increased students' level of preparation, active participation, and the amount learned. Du, Durrington, and Mathews (2007) add a word of caution that preparation that is suitable for interaction in more routine learning tasks may have an opposite effect and actually constrain the discussion when the task is less structured and the learning objective is more conceptual.
The characteristics of group members are another important component of online discussion. Teachers use various methods in forming online groups. Some will mix students into groups attempting to balance technology skills, leadership ability, content knowledge, and diversity based on their personal knowledge of the students. Other teachers randomly assign students to groups. Carabajal et al. (2002) indicate the importance of balance in online discussions. Online discussions foster equal participation among the participants, but it doesn't lend itself to patterns of leadership in which one person dominates what is designed to be a shared space. This poses a conflict because, as Pavitt and Johnson (1999) state, online groups need an effective moderator or the group loses coherence and becomes a group of individuals formulating their thoughts online. In addition, if one member is particularly adept at the skills required by the group task, that individual's skills overshadow the group's ability to succeed. Winograd (2003) addresses the moderator's role as the leader of the discussion. In this role the moderator serves as the motivator for participants by encouraging interaction while providing a trusting discussion environment. The online discussions need to allow each group member to bring their knowledge, abilities, backgrounds, and experiences to the group process as they construct new knowledge.
Other studies have examined the nature manner that may influence online discussion, including writing skills and fast response. Tuckerman's (1965) four-stage team development theory is well received in organizational research and practice, and is proven to be effective in team facilitation and management (Du, Durrington, & Olinzock 2006; Guzzo & Dickson, 1996; Guzzo & Shea, 1992; Schwarz, 1994). According to Tuckerman, team growth is a sequential and developmental process, including four stages in the following order: forming, storming, norming, and performing. Each stage is characterized by two major dimensions: the task-related issues and peer relationship growth (Tuckerman, 1965).
Through the four varied stages, a team faces different issues related to the team task and member relationships. In general, at the forming stage, members tend to be polite and reserved, but not trusting with one another. Typical team behaviors at the forming stage are testing other members and getting oriented with the team task. At this stage, members are uncertain about their roles in the team.
At the storming stage, team members start to communicate more about their feelings and thoughts. They get to know each other better and typically obtain more information about the team task as well. As more communication and collaboration happen, however, arguments, conflicts, and disagreements typically emerge. Leadership is often challenged at the storming stage, and the team needs more clarifications on individuals' roles and responsibilities. Excessive storming leads to anxiety and tension, while suppressed storming leads to resentment and repeated and unhealthy conflicts.
After the storming stage, the team will have gained shared understanding, and thus will achieve norms regarding team procedure, relationships, and team performance as well. In the norming stage, the team shows more cooperation and cohesion as a united whole, with joint efforts and shared understanding.
Finally, when reaching the performing stage the team will work with openness, trust, and flexibility, and become able to perform as a unit. At the performing stage, members assume their varied roles, encourage cooperation and collaboration, and develop more interdependence. Teams at this stage are able to focus their energy on the tasks and function efficiently as a whole.
Online groups face many challenges due to the lack of face-to-face communications or shared social context. The fading or blurry physical, temporal, and psychological boundaries make it difficult for online teams to establish a team identity or sense, which is critical for effective team performance. Appropriate selection and utilization of communication media may help learners better overcome some of the difficulties. With a variety of information and communication technologies, it is very important yet challenging to select and utilize appropriate media for different tasks and at different team development stages. Thus, media research provides another lens to look into the dynamics of online collaboration.
Media richness and social presence theories are the most widely accepted rational theories that explain media choices and media behaviors. Media richness theory (Daft & Lengel, 1984; Havard, Du, & Olinzock, 2005) measures the richness of media in terms of the capacity for immediate feedback, multiple cues, natural language, and personal focus on voice tone and inflection. Media have varied capacities to reduce ambiguity and thus facilitate mutual understanding (Daft & Lengel, 1984; Havard, Du, & Xu, 2008). Richer media facilitate more accurate and meaningful transmission and exchange of ideas. However, as discussed earlier, tasks of different types and complexity have different requirements for information richness in order to achieve maximal group performance. Some tasks require more information and richer media than others for the best team performance.
Social presence theory (Short, Williams, & Christie, 1976) studies media in terms of the degree to which they are perceived to convey the presence of a communication party. The quantity of social presence is how much one believes another party is present. In communication, the psychological distance among communicating parties is referred to as immediacy (Havard, Du, & Xu, 2008). Thus, there are two forms of immediacy: technological immediacy and social immediacy. Technological immediacy is inherent, while social immediacy can be changed (Heilbronn & Libby, 1973). Heilbronn and Libby (1973) state that the maximum amount of exchanged information ensures technological immediacy, and social immediacy is conveyed through communications with verbal or nonverbal cues. Walther (1995, 1996, 1997) suggests that information and communication technology (ICT) is also able to convey social information, just as face-to-face communications, but with lower transfer rate. Walther (1995, 1996, 1997) has also found that ICT-mediated groups have greater social discussion, depth, and intimacy than in face-to-face groups.
In a review of social presence theory and studies on information and communications technology-mediated communication, Gunawardena (1995) concludes that immediacy enhances social presence, which in turn enhances interactions. As related to online collaborative learning, it indicates that the online teams, with assistance from the instructor or an external moderator, should promote the use of media that better convey the notion of social presence in order to increase interaction among the members.
The literature has identified some variables related to students' satisfaction with the collaborative group process. These variables need to be further studied by investigating student perceptions related to these group system variables. The research question, what are students' perceptions of online discussion quality and contributing factors for those perceptions, guided this study.
Method
Setting Background
“Multimedia and hypermedia design” is an online graduate course delivered by a large public university via the WebCT course management system. The course is offered entirely online as part of a graduate program in instructional technology. The course has been offered six times in the past two years and has a typical enrollment of approximately 25 students. Most instructional and learning activities in this course involve intensive online interactions and collaborations using both synchronous (chat room) and asynchronous (bulletin board) conferencing tools.
Students in the course were required to post their reading syntheses to the message board for each of four sections, read all of their peers' postings for that section, and then select at least three messages to respond to for each of those four sections. Each response was indicated by the writer as “challenge,” “elaboration,” or “how I am influenced by the message.” After students had written their responses, they could edit the text of their original posting to include peers' feedback. Around 20% of the final grade depended on the frequency and quality of students' postings to peers, and how they had included peers' feedback in their final reading syntheses.
Group work was also emphasized in a team project. Students worked in teams of 3 or 4. Each team created three mindtool projects focused on the implementation of a certain computer tool in a classroom or training environment. Teammates codeveloped every mindtool project by each making a preliminary post to their team message board with a first draft of a project idea, interacting with teammates with critiques and recommendations, jointly finalizing the project document, and then posting revised documents to the class message board. Around 30% of the final grade depended on students' collaboration efforts in teamwork—“team agreed on a development schedule” and “respond to all teammate postings with thoughtful feedback.”
Additionally, the course was not lecturebased. Instead, the instructor let the students self-explore the subject topics through collaborative discussion and application projects.
Survey Instrument
The quantitative data source included a student survey containing 37 questions related to computer expertise, computer experience, group partner preferences, strategy for preparation, communication media use, manner of discussion, and discussion topic interests. The survey was administered at the end of the semester and quantitative data were collected and analyzed to seek patterns of students' perceptions of online discussion quality in the context of higher education.
The participants in this study were graduate students at a major university in the southern United States. In order to gain a broader spectrum of participant samples, we recruited participants and gathered data from the same, graduate-level course on multimedia and hypermedia design in two different class sections with the same instructor: a completely web-based class delivered through WebCT.
The participants were 142 graduate students from a public university in the Southeast. Specifically, of the participants in this sample, 65.5% were female and 34.5% were male. The sample was 66.9% African American, 28.2% Caucasian, 2.8% Latino, and 2.1% Asian American. Overall, the participants in age groups 21-30, 31-40, 41-50, and 51-60 were 54.2%, 16.2%, 23.2%, and 6.3%, respectively. Among these participants, 46.5% reported that they were full-time students.
The participants were asked about (a) their self-reported computer expertise, ranging from novice (scored 1) to expert (scored 3), and (b) their previous experience with online course, ranging from none (scored 1) to more than two (scored 4). In addition, several multi-item scales were used for the present study:
Group Partner
Informed by related literature (Petress, 2004), four items were used to assess students' preference for group partners (e.g., a good knowledge of the subject matter and a strong technology background). Possible responses ranged from strongly disagree (scored 1) to strongly agree (scored 5). Internal consistency (Cronbach's alpha) was .71.
Strategy for Preparation
Informed by related literature (Cashman et al., 2004), three items were used to measure strategies that students may use for online preparation (α = .60) (e.g., study in advance and contact instructor in the case of anticipated issues). Possible responses ranged from strongly disagree (scored 1) to strongly agree (scored 5).
Communication Media
Informed by related literature (Havard et al., 2008), two items were used to assess communications media that students may use for online discussion (α = .73) (e.g., use of charts and bulletins). Possible responses ranged from strongly disagree (scored 1) to strongly agree (scored 5).
Manner of Discussion
Informed by related literature (Thompson & Ku 2006), three items were used to measure how students approach online discussion (α = .73) (e.g., take the lead and not wait for the others to initiate the discussion). Possible responses ranged from strongly disagree (scored 1) to strongly agree (scored 5).
Discussion Interest
Influenced by related literature (Jin, 2005; Du et al., 2008), three items were used to measure students' interest in online discussion (α = .62) (e.g., their interest and comfortable level on theoretical framework and technical components). Possible responses ranged from strongly disagree (scored 1) to strongly agree (scored 5).
Quality of Discussion
Influenced by related literature (Du et al., 2008), four items were used to measure quality of online discussion (α = .89), including deep learning, competence, and productivity. Possible responses ranged from extremely poor (scored 1) to excellent (scored 5).
Data Analysis
Descriptive statistics were presented first. Following that, zero-order correlations were computed among the independent variables and quality of online discussion. Finally, a simultaneous multiple regression analysis was conducted using gender, computer expertise, computer experience, group partner, strategy for preparation, communication media, manner of discussion, and discussion interest to predict quality of online discussion.
Results
Descriptive Statistics
Table 1 presents the descriptive statistics relating to the study variables. The mean score for computer expertise was 2.21 (SD = .58), indicating that the participants, on average, considered their level of computer expertise between “intermediate” (scored 2) and “expert” (scored 3). The mean score for computer experience was 3.23 (SD = 1.18), indicating that the participants, on average, took “two” (scored 3) or “more than two” (scored 4) online courses previously.
Descriptive Statistics and Pearson Correlations
| Variables | M | SD | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
|---|---|---|---|---|---|---|---|---|---|---|
| 1. Gender (female = 0, male = 1) | .35 | .48 | — | |||||||
| 2. Computer expertise | 2.21 | .58 | −.12** | — | ||||||
| 3. Computer experience | 3.23 | 1.18 | −.09** | −.18* | — | |||||
| 4. Group partner preferences | 4.03 | .80 | −.29** | −.22** | −.18* | — | ||||
| 5. Strategy for preparation | 3.90 | .63 | −.25** | −.22** | −.18* | −.41** | — | |||
| 6. Communication media use | 3.84 | .63 | −.20* | −.31** | −.27** | −.28** | −.30** | — | ||
| 7. Manner of discussion | 3.07 | .86 | −.10** | −.23** | −.26** | −.02** | −.07** | −.24** | — | |
| 8. Discussion interest | 4.15 | .48 | −.14** | −.07** | −.33** | −.39** | −.18* | −.28** | .16** | — |
| 9. Quality of discussion | 3.76 | .69 | −.13** | −.04** | −.04** | −.08** | −.09** | −.30** | .27** | .29** |
| Variables | M | SD | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
|---|---|---|---|---|---|---|---|---|---|---|
| 1. Gender (female = 0, male = 1) | .35 | .48 | — | |||||||
| 2. Computer expertise | 2.21 | .58 | −.12 | — | ||||||
| 3. Computer experience | 3.23 | 1.18 | −.09 | −.18 | — | |||||
| 4. Group partner preferences | 4.03 | .80 | −.29 | −.22 | −.18 | — | ||||
| 5. Strategy for preparation | 3.90 | .63 | −.25 | −.22 | −.18 | −.41 | — | |||
| 6. Communication media use | 3.84 | .63 | −.20 | −.31 | −.27 | −.28 | −.30 | — | ||
| 7. Manner of discussion | 3.07 | .86 | −.10 | −.23 | −.26 | −.02 | −.07 | −.24 | — | |
| 8. Discussion interest | 4.15 | .48 | −.14 | −.07 | −.33 | −.39 | −.18 | −.28 | .16 | — |
| 9. Quality of discussion | 3.76 | .69 | −.13 | −.04 | −.04 | −.08 | −.09 | −.30 | .27 | .29 |
Note: N = 142. *p < .05. **p < .01.
The mean score for manner of discussion was 3.07 (SD = .86), indicating that the participants were “not sure” (scored 3) about their role in online discussion (i.e., whether to take the lead or to wait for others to initiate the discussion). Meanwhile, the mean scores for group partner preferences, strategy for preparation, communication media use, and discussion interest were 4.03 (SD = .80), 3.90 (SD = .63), 3.84 (SD = .63), and 4.15 (SD = .48), indicating that the participants, on average, tended to “agree” (scored 4) with the items in these scales. Finally, the mean score for quality of discussion were 3.76 (SD = .69), indicating that the participants considered the quality of online discussion between “average” (scored 3) and “above average” (scored 4).
Zero-Order Correlations Among Study Variables
Table 1 also includes zero-order correlations among independent variables and quality of online discussion. Quality of online discussion was found to correlate significantly with communication media, manner of discussion, and discussion interest. The only pair of independent variables with a correlation greater than r = 40 were group partner preferences and strategy for preparation (r = 41).
Regression Analysis
The hierarchical regression procedure was used to explain variances of quality of online discussion. In the multiple regression analysis, eight independent variables (i.e., gender, computer expertise, computer experience, group partner, strategy for preparation, communication media, manner of discussion, and discussion interest) were introduced. Regression results indicated that the model significantly predicted quality of online discussion, R2 =.372, R2adj = .334, F (8, 133) = 9.833, p < .001. Together, this model explained 37.2% of the variance in quality of online discussion. A summary of the results is presented in Table 2.
Predicting Quality of Discussion: Results from Multiple Regression (N = 142)
| Independent Variable | B | SE | β | t | F | Total R2 | Total R2adj |
|---|---|---|---|---|---|---|---|
| Intercept | −2.032 | .791 | −2.568 | 9.833 (8, 133)*** | .372 | .334 | |
| Gender (female = 0, male = 1) | −2.670 | .140 | −.465*** | −4.782 | |||
| Computer expertise | −2.174 | .101 | −.147*** | −1.719 | |||
| Computer experience | −2.274 | .061 | −.469*** | −4.491 | |||
| Group partner preferences | −.537 | .115 | −.621*** | −4.689 | |||
| Strategy for preparation | −2.388 | .112 | −.354** | −3.450 | |||
| Communication media use | −2.623 | .101 | −.566*** | −6.192 | |||
| Manner of discussion | −2.127 | .065 | −.160*** | −1.952 | |||
| Discussion interest | −2.521 | .133 | −.361*** | −3.914 |
| Independent Variable | B | SE | β | t | F | Total R2 | Total R2adj |
|---|---|---|---|---|---|---|---|
| Intercept | −2.032 | .791 | −2.568 | 9.833 (8, 133) | .372 | .334 | |
| Gender (female = 0, male = 1) | −2.670 | .140 | −.465 | −4.782 | |||
| Computer expertise | −2.174 | .101 | −.147 | −1.719 | |||
| Computer experience | −2.274 | .061 | −.469 | −4.491 | |||
| Group partner preferences | −.537 | .115 | −.621 | −4.689 | |||
| Strategy for preparation | −2.388 | .112 | −.354 | −3.450 | |||
| Communication media use | −2.623 | .101 | −.566 | −6.192 | |||
| Manner of discussion | −2.127 | .065 | −.160 | −1.952 | |||
| Discussion interest | −2.521 | .133 | −.361 | −3.914 |
Note:*p < .05. **p < .01. ***p < .001.
As documented in Table 2, six variables were found to have a statistically significant effect on quality of discussion. Quality of discussion was positively associated with communication media (β = .566, p < .001), computer experience (β = .469, p < .001), discussion interest (β = .361, p < .001), and strategy for preparation (β = .354, p < .01). Males, compared with females, reported statistically higher scores in quality of discussion (β = .465, p < .001). Finally, quality of discussion was negatively associated with group partner (β = −.621, p < .001).
Discussion
The present study linked a number of factors to quality of online discussion. Regression results revealed that quality of online discussion was positively associated with communication media selection and use, and computer experience. According to Guzzo and Dickson (1996), teams at different stages of development face different issues and problems, which generate different needs for social presence and richness of media. For example, at the forming stage, during which the major task is information seeking, the richer the medium is, the more information input members can detect and process. In addition, the more social presence is perceived, the more likely and easier it is to establish trust. Similarly, at the storming stage, when conflicts occur frequently, there is a strong need for social presence to address these conflicts, especially affective conflicts. Face-to-face is identified as the richest medium (Daft & Lengel, 1984) and also the best medium to convey social presence (Du & Sun, 2007).
When students think about computer experience and using communication media, they immediately associate it with a means to provide immediate feedback, flexibility within the course, and help with diverse cultural issues that may arise. Using online posting as a form of communication, the student is able to obtain ideas from other participants with different ethnicity and cultural differences, and to share similar experiences of their own. They can also enter into a chat room to share information from each other. By using charts and bulletins, the tasks of all students participating in the online discussion will be able to achieve successful outcomes in an efficient and timely manner. With continuous upgrades in new technology software, such as Blackboard, these changes allow students to communicate more effectively. The tools within the content learning system allow the student to share discussions, e-mails, and chats at their convenience.
Students are more confident and knowledgeable after taking the online course, and shared a positive experience with accessing course information. In their review, Guzzo and Shea (1992) found that almost all studies on group collaboration have reflected an underlying input-process-output model. All the qualities that each member brings to the team are inputs, such as expertise, personality, and strength. The inputs by using computer technology are then processed through group interactions and activities, such as information exchange, cooperation, collaboration, and taking leadership role, which are transformed into output of the teamwork later. For online collaborative teams, all the input identification and processing are expected to happen with assistance of ICT. However, it can be very difficult for online collaborative teams to detect or benefit much from some of these inputs, such as personality. Media richness theory and social presence theory enlighten the understanding of rational choice of media. Social constructivist theories help to understand media behaviors, which are not always consistent with the rational theories. These theories enable instructors and instructional designers to understand why one medium works better than the others at a certain stage, and such knowledge enables the instructors to better provide facilitation and moderation as needed.
Results revealed that quality of online discussion was positively associated with discussion of topic interest and strategies for preparation. Students answer questions through discussion on topics of interest because the students are familiar with the information or topic being discussed. Due to the online learning environment, it is not thought of as a social network, but it allows students to become focused on the academic issues, such as course content, final projects, and individual assignments to be discussed (Winograd, 2003). Students also indicated they preferred advanced preparation when participating in online discussions, which may explain why they don't have a sense of inadequacy when participating in online discussions that are prepared for in advance. Students want to be successful in the course. This allows them to avoid feeling unknowledgeable about the course content (Roberts, 2004). Therefore, students prepare by researching and discussing the information beforehand, and then posting significant information that can benefit other students.
An interesting result indicated quality of discussion was negatively associated with students' strong group partner preference. One reason students often avoid partners when completing projects is that one partner may not be as motivated as the other in completing tasks or assignments, which leaves one person to complete the projects. This situation leaves that person resentful of the other participants in the group. Another reason students avoid partners in group projects is that they become too confident, and believe they know more than the others in the group. This tends to offend the others in the group, decreasing their motivation to participate and are less willing to listen to their partners' ideas, and feel they have become a member of a group that is run by a dictatorship. Yet another reason students avoid group partners is that they may be shy and become distant and less willing to vocalize their thoughts and ideas because they think their partners will see their ideas as insignificant and not worth discussing (Jin, 2005). Finally, they may feel that their expertise in certain areas that are the topic of discussion is the only pertinent information that is relevant or necessary. This generally lowers the confidence and self-esteem of other students who lack the experience that their partner may have. They will feel that they lack the knowledge required for the course or assignment (Thompson, & Ku, 2006).
Zhang and Carr-Chellman's (2001) research indicated that the team development during online collaborative problem solving centered on leadership building and buy-in. They also found that groups receiving external moderation established team norms at an earlier time, whereas peer-controlled groups either did not build norms at all, or did it at a much later time. With the absence or delay of norm-building, the groups experienced more relationship problems associated with lack of mutual respect, support, or encouragement. As a result, these teams conducted fewer interactive collaborations on the team task.
Theoretically, teams must resolve all the issues and challenges they face in order to move on to the next stage, and no stage can be skipped (Tuckerman, 1965). In practice, however, teams often try to jump from initial forming to final performing stage. It is common that teams avoid or suppress conflicts, and minimize personal relationship building with the hope to focus all the energy on the task itself. Thus, it is strongly recommended that a person with certain authority from outside of the team play the role of a facilitator or moderator. Such a role should encourage and help the team to address the issues they face at the current stage. It is also suggested that during the forming stage, assistance should be provided to make leadership clear in order to help the team to deal with both the task and relationship issues (Schwarz, 1994; Tuckerman, 1965).
In a review of literature on computer-assisted teams, Hollingshead and McGrath (1995) found that information and communication technology (ICT)-mediated teams had fewer interactions and information exchanges, and that they took more time on tasks as compared to face-to-face teams. They also found that, in terms of resolving conflicts, the ICT - mediated teams were not as good as face-to-face teams. Therefore, it is particularly important, yet challenging, to understand the relationship-related issues and be better able to resolve affective conflicts in an online collaborative environment.
It is difficult to establish trust among people who are only connected through ICT. For this reason, it is highly recommended for virtual teams to arrange one or more initial, face-to face meetings, if at all possible (Jeong, 2000; Mittleman, Briggs, & Nunamaker, 2000). In cases where face-to-face meetings are unavailable, initial contacts can be made through the use of rich media, which are capable of conveying both verbal and nonverbal communication cues as well as social presence (refer to the media section for detailed discussion). Mittleman et al. also suggested using an infonnal break for online team meetings, when all parties can share casual talks and social jokes with the assistance of lCT. Another useful strategy, as found by Zhang and Carr-Chellman (2001), is to promote some small social talks and encourage members to greet each other in a positive tone.
In this study, males reported statistically higher scores in quality of discussion. Usually, however, male students are more comfortable using computer taking online courses and participating in online discussions than are female students. Warschauer (2000) suggested that computer-based educational programs did not benefit female students as much as it benefited male students because females were likely to be disinterested in the learning settings presented by the available computer products, typically with drastic movement and even violent images. Girls tend to use a computer for interpersonal and social purposes, whereas boys tend to use a computer to gain power and control of the physical environment (Viadero, 1994).
Results revealed that there are no differences statistically in students' computer expertise, though the students come from diverse backgrounds. They use computers primarily for academic coursework. Regardless of cultural backgrounds, there is basically an even division of novice, intermediate, and expert computer users.
Conclusion
This study is an important complement to the existing literature in collaborative learning. Online collaborative discussion is relatively a new pedagogy with the emergence of online technologies, and thus it is interdisciplinary by nature. Many other fields, such as communication, psychology, organization and management, and information science, have developed rich resources of literature regarding team collaboration, team facilitation, media behaviors, groupware and social informatics. Instead of looking at online collaborative learning from any single perspective in isolation, this study attempted to further bridge literature from varied fields to detangle the complicated, dynamic relationships of online discussion. Perhaps the most important implication of this work is to inform online learning designers to be sensitive and cognizant of needs (not only expectations) of the students as they create future distance education experiences. We believe it is possible to create more flexible online learning environments that allow for learners to choose more experiences, though we recognize that this innovation may be more time intensive for instructors who count on online discussions for certain scalability in online environments. A possible outcome might be to consider differentiating spaces for various sorts of needs to help the learners in their quest to focus interactions on learning. The present study suggests that teachers can exert positive influences on online discussion effectiveness by designing more deeper and interesting assignments and then later by providing more frequent feedback. The teacher should find ways to better improve students' abilities, such as assigning more difficult assignments or projects. The teacher could also encourage further student relationship development by allowing the group members to rotate, form new groups, and then utilize their own strategies to improve the quality of online discussion.
