Group formation for web-based collaborative learning with personality information.

This article addresses the question of how to build human relations in web-based collaborative learning. Although the field of web-based education/training (WBE/T) has grown rapidly in the last few years, there has been little systematic research on the issue of group formation in such environments.

It is the difficult to build human relations in such situations, due to the associated problems in computer-mediated communication. In physical environments, a person often guesses other's personality through nonverbal communication, which is very limited in Computer-Mediated Communication (CMC).

The aims of this article are three-fold. First, the concept of personality information for awareness (PIA) is proposed, followed by the design of a a PIA-based system called "PIAGETS", which supports human relations building for collaborative learning in WBE. Next, the details of self-analysis subsystems and user aide subsystems, consisting of YG personality test module and Group Formation module respectively are given. Finally, the details of a basic investigation of human relation in face-to-face class are provided. The aim of this experiment is to clarify the human relation building with focus on group formation process in face-to-face classes which hopefully could be applied or extended to online class, i.e. web-based collaborative learning.

There are many researches in collaborative learning environments through Internet, where Computer Mediated Communication (CMC) plays an important role (Ogata, Matsuura & Yano, 1996). However, some issues regarding methods and technologies still remain.

CMC refers to the use of networked computers for communication, interaction, and exchange of information. Examples of CMC technologies include e-mail, bulletin boards, newsgroups, and computer conferencing (Miltiadou, 2001). There are many researches, which compare the face-to-face versus online instructional situations (Johnson, Aragon, Shaik, & Palma-Rivas, 2000; Tiene, 2000; Sumner & Hostetler, 2002), to study the effectiveness of CMC in learning. Tiene summarized the key differences between online and face-to-face discussions based on the following four factors: access, timing, mode of expression and visual cues. One potentially limited aspect of the online discussion is its lack of visual cues. Nonverbal cues can be important in any effort to communicate.

Other researches have studied the effects of CMC and social presence in online education (Jiang & Ting, 2000; Hewitt, 2001; Miltiadou, 2001; Leh, 2001; Tu, 2002). It is not enough to just introduce CMC technologies into web-based education/training (WBE/T) to have effective learning where, how to increase learner's motivation is a critical issue.

It is widely believed that e-Learning has lower satisfaction ratings and higher dropout rates than the traditional classroom (Horton, 2000). The improvement of social presence in online instruction may help. Social presence is the degree to which a person feels "socially present" in a mediated situation. It is an important factor in distance learning and one of the theoretical constructs in distance education. People feel intimate and share more if they feel socially present. However, when social presence is lacking, people recognize the environment as impersonal and share less (Leh, 2001). Gunawardena (1995) suggest that instructors and moderators need to learn to adapt to telecommunications media by developing interaction skills that create a sense of social presence. Horton suggested several techniques that designers and instructors can use to keep learners interested, energized, and enthusiastic, and one of them is to build a learning community. In a learning community, learners take responsibility for their own education and training and for that of other members of the community. The concept of a community goes beyond an individual course to empower learners to own the learning process.

Thus, it is necessary to build learner-learner or teacher-learner human relations. Our claim is that to motivate learners in WBE/T, the formation of communities, i.e. the building of human relations plays a central role. To build human relations in such environments, attention is focused on the learner and the exchange of personality information.

THE PURPOSE OF THE STUDY

To build human relations in traditional face-to-face classrooms, a learner performs both verbal and nonverbal communication activities with a partner or classmate. However, in web-based collaborative learning such as WBE/T, it is more difficult to build human relations, because of the problems in CMC. One problem in CMC is the limited exchange of information--especially nonverbal communication--when compared with face-to-face communication (Matsushita & Okada, 1995; Tiene, 2000).

The aim of this research is to provide support for building human relations in web-based collaborative learning, by overcoming the limitation in the exchange of information, especially about the learner's personality (Hishina, Okada, Ogasawara, Sakamoto, & Suzuki, 2002).

OUR APPROACH

It is important to take into account human relations in web-based collaborative learning, where information about the learner's personality plays a central role. This personality information can be classified into the following four types: (a) self-analysis, (b) self-disclosure, (c) person analysis, and (d) person perception (Buss, 1986; Sugino, Kameshima, Andou, Komaki, & Kawabata, 1999). In face-to-face situations, a person often grasps his own personality and guesses a partner's personality.

However, in web-based collaborative learning, we think that it is difficult to grasp and guess individually the four types of personality information. This is because a person does not pay much attention to the personality information in CMC. To overcome such drawbacks, in this research, awareness is regarded as a key factor, which is provided by exchange of personality information. In web-based collaborative learning, awareness is important to create efficient collaboration.

Proposal of Personality Information for Awareness (PIA)

Taking into account all the aspects previously mentioned, a new concept of PIA is proposed. PIA represents personality information for (a) self-analysis, (b) self-disclosure, (c) person analysis and (d) person perception. To build human relationships, each learner performs the communication activities shown in Figure 1. The explanation of each type follows.

1. self-analysis: the information in which student A perform self-analysis in order to form the self-concept. By this, student A discovers him/herself.

2. self-disclosure: the information in which--from the results of self-analysis--student A selects, sometime elaborates, and shows to the public (student B in case of Figure 1) only the selected/elaborated information.

3. person analysis: the information in which, student B notices and analyzes the personality information presented by others (student A in case of Figure 1).

4. person perception: the information which--from the results of personal analysis--student B acknowledges whether or not to build an interpersonal relationship with student A.

Note that although Figure 1 shows student A as the focus of attention, the same activities are valid for student B.

Designing a PIA-based system: PIAGETS

To support the building of human relation for collaborative learning in WBE, a PIA based system named PIAGETS is proposed. PIAGETS is the abbreviation for "Personality Information for Awareness applied to Group learning in web-based Education/Training Support system".

PIAGETS is composed of the following five functions (1)-(5), as shown in Figure 2:

1. Subsystems for self-analysis,

2. Subsystems for self-disclosure

3. Subsystems for personal analysis

4. Subsystems for personal perception

5. Subsystems for user aide

[FIGURE 1 OMITTED]

Regarding the first four functions (1) to (4), it is up to the user to make active use of them, and they are in a one-to-one correspondence with the four types of PIA information.

On the other hand, regarding the subsystem for user aide (5), the system itself provides active support to the user, independent of whether or not the user wants to make active use of this function. One way to realize this support is through intelligent agents, which learn the user's preferences (Okada, 2000).

Development of PIAGETS

Currently, the development concentrates on the subsystem for self-analysis and subsystems for user aide. The design philosophy and overview for both subsystems follows.

1. Subsystem for self-analysis: Personality Test Module

This YG test module can be used as a psychological testing to analyze self-personality objectively. It is presumed that the user would make active use of the subsystem for self-analysis. Moreover, the aim is to implement the functions required for the system itself to provide active support to the user, based on the self-analysis data.

In the YG test module, a modified version of the YG personality test developed by Yatabe, was adopted. The original YG personality test is a traditional method to investigate personality in Japan.

[FIGURE 2 OMITTED]

The original YG personality test is derived from the following processes.

First, to identify the individual's character and their characteristics, a questionnaire must be answered. The original YG personality test has 120 questions, which would require a lot of time and effort to answer. To reduce the time required to answer the questionnaire, a reduced trial YG personality test developed by Uehara (1997), consisting of 36 questions, was used.

Based on the responses, the strength of 12 predefined personality traits are calculated. These 12 personality traits belong to one of two classes, where six belong to the "General emotionalism" class (D: Depression, C: Cyclic tendency, I: Inferiority, N: Nervousness, O: Objectivity, Co: Cooperativeness) and the remaining six to the "Extroversion/Introversion" class (Ag: Agreeableness/Aggressiveness, G: General activity, R: Rhythmic, T: Thinking extroversion, A: Ascendance, S: Social extroversion).

According to the outcomes of this calculation, the individuals are classified in five personality group types A, B, C, D, or E. However, there are cases in which the individuals are judged to be of two or more of the A, B, C, D, or E types, for example, belonging to both type A and type C.

In Uehara's trial test, the five personality types are named as follows: A as "average type," B as "black list type," C as "sedative type," D as "adaptive type," and E as "eccentric type."

A web-based YG personality test, which will be referred as YG test module, was implemented, as a Perl-based CGI and a set of HTML web pages. An example of answer input screen and the result display screen for the YG test module are shown in Figures 3 and 4, respectively. The user answers the questions given in the input screen by selecting one of the radio-buttons (Figure 3). When the user finishes with the answers, the module calculates, which of the personality group he belongs to and displays the results on the screen (Figure 4). The outcomes of this YG test module are used to match users, to realize the Group Formation, which is a function corresponding to the user aide subsystem.

2. Subsystem for user aide: Group Formation module

In a face-to-face class, the group formation process can be described as follows. In many cases, the learner often chooses a friend as a partner for collaborative learning. When for some reason if the group cannot be formed with friends, then the teacher might build the groups based on other mechanisms such as grouping by names or by the seat distance between students. If a partner for collaborative learning is decided, a person often grasps his own personality and guesses a partner's personality.

On the other hand, in an environment such as WBE, it is difficult to select a partner for collaborative learning, because there are not enough criteria (e.g., information about seating distance) for group formation.

To overcome such problems, a system to do the grouping for collaborative learning, characterized by using the personality information from the YG test module as the matching criteria is proposed.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

METHODS AND RESULTS

For the development of our Group Formation module, we have investigated the group formation process in face-to-face, taking into account the individual's personality.

To measure the influence of the student's friends, pal or "buddy" on his personality structure, the group formation in the face-to-face class was observed (Investigation 1). The next stage is to compare the groups identified from observations, against the groups formed based on the YG personality test information (Investigation 2). The results of this comparison will be used in computer-assisted group formation for collaborative learning.

Investigation 1: Groups Identified From Observations

To identify the student's buddy, the student's seat selection in the classroom was recorded. First, based on the observations of the students seating position in the classrooms, the student's friendship relationship was guessed, as shown in Figure 5. For instance, the learners surrounded by lines in Figure 5 were assumed to have friendly relationships.

[FIGURE 5 OMITTED]

Investigation 2: Groups Formed Based On The YG Personality Test Information

In the previous section, the personality was investigated using the YG personality test module. A sample of the personality data that was obtained with the YG module is shown in Table 1. Note that data from student 20 is not included as she was absent on the day of the YG test.

Then, the seating places in the classroom of the students during four classes were investigated. The students' seating positions were plotted according to the axis shown in Figure 6, and the physical distance between students calculated using the Formula 1.

From the position of the seats in the classroom, it is assumed that students with average physical distance of two or less (for example, pair) show a friendly relationship.

[FIGURE 6 OMITTED]

Table 2 summarizes the result of the mean distance and the YG test module for each pair.

DISCUSSIONS

From the results of the Figure 5 (Investigation 1) and Table 2 (Investigation 2), it is possible to see the coincidence between all the groups identified through observations and the groups formed based on resulting personality information from the YG test module. Through this experiment, it was confirmed that buddies tend to sit close to each other.

Basically, the data of the YG test module is judged to belong to one of the five personality types A to E. However, taking into account the cases in which data can belong to several types simultaneously, there are 31 possible combinations types (calculated based on the combinatorial Formula 2).

It is possible to see from the results shown in Table 2 that the pairs are composed of learners belonging to different personality types. For instance, regarding pair 2, the result of the YG test module shows that student 6 is only type C, while student 12 is both type C and D.

From our studies of group formation on face-to-face classes, we can see that for collaborative learning environments, the groups are composed by individuals with different personality types. Based on this finding, we would like to see whether this assumption, which was tested in a face-to-face scenario, is still valid for collaborative learning in CMC environments.

Algorithm for Grouping Formation

Taking into account the experimental results explained before, an algorithm for group formation based on personality information--which is implemented in an online Group Formation support--is proposed. The Grouping Formation module uses the judgment result data from the YG test module. Based on the results of the judgment result data, the system proposes the learner's grouping (pair) so that each group is composed from members with different personality types.

The group-formation process (pair) is done with a seven-step algorithm explained bellow:

Step 1: the learner's name is stored in the arrays, one array for each of the 31 type combinations.

Step 2: for each of the 31 arrays, the learners are reordered randomly.

Step 3: the number of learners is counted at each array, and the 31 arrays are ordered in descending order, based on the total number of learners for each type, for example, the array type with most learners first.

Step 4: to form the pairs, one learner is picked up sequentially, from the array with most learners and the second with most learners. Output the resulting pair.

Step 5: for step 4, if there are no more learners left on the array to form the pairs, then the next array with most number of learners is selected.

Step 6: repeat the step 4 (and step 5 when necessary), to continue the pair formation process.

Step 7: when there is not two arrays of different types, then form the pairs randomly with learners from the same array.

This Web-based Grouping Formation module was implemented, based on CGI with Perl and a set of HTML Web pages.

LIMITATIONS

Currently, to our knowledge, there are not many research studies for CMC-based collaborative learning, which deals with group formation by using personality information. Thus, we have targeted our focus on this issue. Unfortunately, when we started this research, the availability of fields for experimenting with CMC-based instructions was scarce.

To overcome this barrier, we have decided to first, investigate on the group formation mechanism for collaborative learning in traditional face-to-face class. The next step planned is to apply this mechanism to CMC-based instructions and check the validity of the approach. In other words, our approach is based on face-to-face group observations, and "hopefully," we will be able to apply such observations to virtual communications. One of the current limitations is that we have to test whether the group formation observations on face-to-face class can be really applied to online class, i.e. web-based collaborative learning.

As our approach for group formation is based on the learner's personality information, another question that we need to answer is "Does the personality information remain the same for CMC and face-to-face?" Researches on personality can be classified into two opposite groups, where one group argues that personality is uniform and goes beyond the particular situation, while the other one claims that is not uniform and dependent on the particular situation (Krahe, 1992). We have adopted the hypothesis of the first group, where the personality is independent of the particular situation. Under this hypothesis, the CMC and face-to-face constitute the particular situations. Our second limitation on the current version comes from the fact that we need to test the validity of this assumption.

CONCLUSION AND FUTURE WORKS

First, the problems involved in collaborative learning environments, affected by the computer-mediated communication were discussed. To overcome such problems, the concept of PIA, and a PIA-based system called "PIAGETS"--which supports human relations building for collaborative learning in WBE--were proposed. Next, the development of two subsystems, namely YG personality test module and group formation module, were introduced.

From the data gathered from the YG test module, the influence of the student's buddy on the personality structure was studied. It was found from the results of this study that the pairs are composed of learners belonging to different personality types. Based on this finding, a computer-based group formation algorithm for collaborative learning environment was proposed.

Future works include the evaluation of the effectiveness for the proposed PIAGETS, the investigation of the human relationship building process and the implementation of additional support functions for human relationship building.

Table 1 The Personality Type Data by YG Test Module

Respondent                                           Personality Type
(Student No.)  D  C  I  N  O  Co  Ag  G  R  T  A  S     (by YG)

Student 1      2  4  2  1  4  6   3   0  0  4  2  4        C
Student 2      1  3  4  4  3  2   3   5  2  5  4  5        A
Student 3      4  4  5  4  4  2   2   4  0  5  1  6        A
Student 4      1  1  3  2  2  1   1   1  0  1  2  3        C
Student 5      1  6  5  4  2  1   2   3  2  2  3  3        C
Student 6      4  4  6  2  2  1   2   0  0  0  2  3        C
Student 7      2  4  2  2  3  2   4   1  1  3  6  4        C
Student 8      0  6  3  5  1  0   4   2  0  3  4  5        C
Student 9      0  6  3  2  2  2   4   1  2  0  4  5        C
Student 10     2  2  4  1  4  2   5   2  1  5  6  5        D
Student 11     2  4  2  2  2  2   3   2  5  6  5  6        D
Student 12     2  6  4  4  2  2   2   1  5  4  3  6        C and D
Student 13     4  4  3  4  4  4   4   2  3  4  0  3        A
Student 14     2  2  2  1  1  1   6   1  4  6  5  6        D
Student 15     2  6  3  3  3  1   5   2  2  5  5  2        C and D
Student 16     1  3  5  2  3  1   4   2  6  3  6  4        A and D
Student 17     2  3  2  3  1  3   1   4  3  4  2  2        A and C
Student 18     3  3  5  4  2  0   1   2  0  3  3  2        C
Student 19     3  2  3  4  5  0   1   1  2  0  6  2        C
Student 21     4  4  1  3  2  0   3   2  2  3  6  4        A
Student 22     1  3  3  3  4  2   3   0  0  2  4  4        A

Table 2 Summarizes the Result

Pair No. /         Respondent     Personality Type
Physical Distance  (Student No.)       (by YG)

Pair 1             Student 3             A
1.25               Student 17            A and C
Pair 2             Student 6             C
1.00               Student 12            C and D
Pair 3             Student 7             C
1.00               Student 13            A
Pair 4             Student 14            D
1.25               Student 22            A

Formula 1: the physical distance between students

Student [alpha] 's seating position (X1, Y1)

Student [beta] 's seating position (X2, Y2)

The physical distance between

Student [alpha] and Student [beta]:

[square root of ((X2-X1)[.sup.2]-(Y2-Y1)[.sup.2])]

Formula 2: Calculated based on the combinatorial formula

[5.summation over (k=5)][.sub.5][C.sub.k]

This case is [.sub.5][C.sub.1+5][C.sub.2+5][C.sub.3+5][C.sub.4+5][C.sub.5] 31 possible combinations types.

References

Buss, A.H. (1986). Social behavior and personality. Mahwah, NJ: Lawrence Erlbaum.

Gunawardena, C. (1995). Social presence theory and implications for interaction and collaborative learning in computer conferences. International Journal of Educational Telecommunications, 1(2/3), 147-166.

Hewitt, J. (2001). Beyond threaded discourse. International Journal of Educational Telecommunications, 7(3), 207-221.

Hishina, M., Okada, R., Ogasawara, T., Sakamoto, E., & Suzuki, K. (2002, October). A fundamental study for grouping formation module based on personality information. Proceedings of E-Learn 2002, (pp. 449-456).

Horton, W. (2000). Designing web-based training: How to teach anyone anything anywhere anytime. New York: John Wiley & Sons.

Jiang, M., & Ting, E. (2000). A study of factors influencing student's perceived learning in a web-based course environment. International Journal of Educational Telecommunications, 6(4), 317-338.

Johnson, S., Aragon, S., Shaik, N., & Palma-rivas, N. (2000). Comparative analysis of learner satisfaction and learning outcomes in online and face-to-face learning environments. International Journal of Interactive Learning Research, 11(1), 29-49.

Krahe, B. (1992). Personality and social psychology: Towards a synthesis. London: Sage.

Leh, A.S.C. (2001). Computer-mediated communication and social presence in a distance learning environment. International Journal of Educational Telecommunications, 7(2), 109-128.

Matsushita, Y., & Okada, K. (1995). Collaboration and communication. KYORITSU SHUPPAN, Inc. (In Japanese)

Miltiadou, M. (2001). Computer-mediated communication in the online classroom. International Journal of Educational Telecommunications, 7(4), 407-419.

Ogata, H., Matsuura, K., & Yano, Y. (1996, June). Knowledge awareness: Bridging between shared knowledge and collaboration in sharlok. Proceedings of ED-TELECOM 96, (pp. 232-237).

Okada, R. (2000). Learning user preferences for personalized and collaborative web searching. International Journal of Kansei Engineering, 1(4), 47-56.

Sugino, K., Kameshima, S., Andou, A., Komaki, K., & Kawabata, H. (1999). The psychology for learning the human relations. Fukumura Publication's Inc. (In Japanese)

Sumner, M., & Hostetler, D. (2002). A comparative study of computer conferencing and face-to-face communications in system design. Journal of Interactive Learning Research 13(3), 277-291.

Tiene, D. (2000). Online discussions: A survey of advantages and disadvantages compared to face-to-face discussion. International Journal of Educational Multimedia and Hypermedia, 9(4), 369-382.

Tu, C. (2002). The measurement of social presence in an online learning environment. International Journal on E-Learning, 1(2), 34-45.

Uehara, M. (1997). Yatabe Guiford personality test. In Your Personality able to understand and be changed! (pp. 81-89). TAKARAJIMASHA, Inc. (In Japanese)

MASATERU HISHINA

Tokyo International University , Kawagoe, Saitama Japan

[e-mail]

ROBERTO OKADA

Miyagi University, Kurokawa, Miyagi Japan

[e-mail]

KATSUAKI SUZUKI

Iwate Prefectural University, Takizawa, Iwate Japan

[e-mail]