Constructing a streaming video-based learning forum for collaborative learning.

As web-based courses using videos have become popular in recent years, the issue of managing audio-visual aids has become pertinent. Generally, the contents of audio-visual aids may include a lecture, an interview, a report, or an experiment, which may be transformed into a streaming format capable of making the quality of Internet-based videos acceptable to learners using a limited bandwidth. Although the streaming technique enables learners' accessibility to audio-visual aids, such usage only supports the instructors' perspectives of instructional videos. In fact, distance learners can contribute ideas in both text and audio-visual formats. However, previous research has not considered the feasibility of distance learners' contributing audio-visual aids without a video camcorder. The objective of this study is to demonstrate that audio-visual aids from distance learners should be considered in designing a web-based course. To demonstrate this concept, this report first introduces a screen camcorder tool that enables learners to record activity in a computer desktop as videos in standard format or streaming format. Then a collaborative learning strategy, Jigsaw II, is applied to encourage expertise groups to contribute streaming videos for training other learners. Moreover, a summary of learners' use of videos contributed by others supports instructors' pedagogical requirements. Hence, a multidimensional analysis of streaming video-based collaborative learning is applied to demonstrate the impact of videos contributed by distance learners. Finally, a preliminary survey of technology acceptance is administered to 37 learners. The results show the feasibility of audio-visual aids contributed by distance learners without a video camcorder.

**********

As web-based courses using videos have become popular in recent years, the issue of managing audio-visual aids has become noteworthy. Using videos for distance learning is not a new idea (Kozma, 1986; Levne, 1992; Zigerell, 1991). For instance, a distance learning course may transmit videos conventionally by VCR, cable TV, broadcast TV, VCD, or DVD. Although research indicates that a face-to-face (synchronous or asynchronous) situation is not essential for learning, some reports show that distance learners prefer videos to other media. Hence, a distance-learning instructor may often provide videos for distance learners by way of the Internet.

Audio-visual aids may include a lecture, an interview, a report, the operation of a machine, or an experiment. In fact, the contents of such aids are discipline-dependent. For instance, the major portion of a video for a physics course may demonstrate an experiment. The instructor is typically the only provider of audio-visual aids in most web-based courses, but few instructors are involved in planning or designing instructional videos. Although some studies involve instructors as contributors of videos, most of them focus on teacher professional development. For instance, the Teach-Scape system was developed to support teachers' reflections on their classroom instruction videos or video-based case studies that illustrate and analyze exemplary teaching in real classrooms (Pea, 2001). Generally, audio-visual aids are widely considered as replacements for learning activities that cannot be implemented in a classroom situation. As Benney (2001) stated, "Most certainly, video is a successful medium for taking students into the field to observe facts, processes, and emotional events which would other-wise be impossible to see and experience."

To provide instructional videos over the Internet, the video signals must be converted into a streaming format to make the quality acceptable to learners on a limited bandwidth. A streaming format means that the requested video signals are gradually transmitted from server to client. The best-known commercial products for providing video-streaming functionalities are the Microsoft[TM] Windows Media Server and the RealNetworks[TM] Real-Media Server. Consequently, a web-based course involves preparing the contents of videos, converting them into a streaming format, and setting up a media server to provide instructional videos through the Internet.

Existing learning strategies for video-based instruction can be categorized into three types: (a) passive watching, (b) learning/practicing while watching, and (c) learning/practicing after watching (DeMartino, 2001). With the passive-watching strategy, students are not engaged in a discussion activity or a learn-by-doing activity. This strategy generally results in poor long-term learning outcomes (Schluger, Hayes, Turino, Fishman, & Fox, 1987). With the learning/practicing-while-watching strategy, a student or a learning group can pause, stop, forward, or rewind a video clip to self-pace learning by joining a discussion or learn-by-doing activity with learning partners. The learning/practicing-after-watching strategy involves a discussion activity or learn-by-doing activity following a teaching video, providing a reflective opportunity to deepen long-term outcomes.

With each type of learning strategy, the usage of audio-visual aids in learning typically is an instructor's decision. Distance learners are not expected to provide videos to depict an idea or communicate with others. Hence, a distance-learning instructor tends to provide a Bulletin Board System (BBS), an e-mail list, or a Web Bulletin Board (WBB) to support asynchronous, collaborative learning after watching an instructional video (Repenning, Ioannidou, & Phillips, 1999). In fact, distance learners can contribute ideas not only in text format but also in audio-visual format. For instance, the teacher education program at Tufts University has students create a videopaper using the VideoPaper Builder, a multimedia authoring tool developed jointly by TERC and the Concord Consortium (DiMattia, 2002). However, distance instructors usually do not consider the feasibility of audio-visual aids contributed by distance learners. Although using a WBB to enable distance learners to communicate with each other is popular, WBBs are not designed for (networked) video-based instruction. Distance learners can help one another learn through an asynchronous, text-formatted Web forum. Alternatively, distance learners can asynchronously communicate with each other by using video clips. These issues are referred to as the contributor and usage problem.

The required computer hardware for videoconferencing--a video camera, a video capture card, and network bandwidth--is expensive for a distance learner. Consequently, most web-based courses do not completely implement a video-based communication environment because of the consideration of distance learners' financial burden. Moreover, video-based instruction is generally regarded as a synchronous communication tool or a passive medium, as is the case with recorded lectures. However, a noteworthy idea is that distance learners can contribute audio-visual aids--as opposed to videoconferencing--without additional computer hardware. In other words, distance learners do not need a video camcorder to record computer screen activity as video and oral explanation as audio in a video-based communication environment. The major difficulty is how to organize and guide learners to contribute their videos for each part of a learning topic. However, distance learners can join the process of asynchronous, video-based instruction, helping to eliminate the problem that instructors tend to spend too much time lecturing. The three components of a collaborative learning strategy--sharing, interdependency, and involvement--allow learners to contribute videos for teaching each other (Salomon, 1991). Thus, refining a selected collaborative learning strategy is essential to the effective use of videos for peer learning. These issues are referred to as strategy refinement of video-based instruction.

Because web-based courses use a media server to provide streaming videos for instruction or communication, logs of media server usage are generated when a learner plays, pauses, stops, forwards, or rewinds a clip of a teaching video. One record of media server logs can indicate, for example, the filename of the streaming video that was accessed, the IP address of the computer that has accessed that streaming video, the time and date that the streaming video was accessed, the duration of an action, action commands, and the using protocol. In other words, analyzing media server logs can partially answer the question "Can the metadata applied to streaming videos offer learning strategies and information that was previously unavailable?" (Doerksen, Mattson, & Morin, 2000). Thus, an instructor can figure out the impact of the innovative methodology for video-based instruction/learning. This issue is similar to web usage mining, by which web log records are used to track user access patterns of Web pages (Han, & Kamber, 2001). However, some features of learners' behavior records--play, forward, rewind, and stop--are found only in media server logs. Hence, it is necessary for an instructor to have an effective tool to extract pedagogical meanings from logs of a streaming video-based learning system. Some commercial products generate various statistical results from the media server logs to help the server administrator improve server efficiency. However, such statistics may not satisfy the requirements of a distance learning instructor. Instead, the online instructor needs to observe a learner's use of Internet-based videos for learning from various aspects of the learner's behaviors, but never from the aspect of server efficiency. These issues are referred to as the statistical and analysis problems.

To demonstrate the feasibility of a video-based learning forum, a screen camcorder tool that enables learners to record screen activity as videos in streaming format is introduced. The screen camcorder enables distance learners to develop instructional videos without a video camcorder. Then a collaborative learning strategy (Jigsaw II), proposed by Elliot Aronson (Aronson, & Patnoe, 1997; Slavin, 1995), is applied to encourage expertise groups to contribute streaming videos for training other learners. Next, this article proposes a data cube framework as a repository for learners' use of streaming videos. The data cube repository enables an instructor to acquire summary information about a learner's behaviors, such as video clips used more than twice, average duration before stopping a video clip, the rewinding interval of a video clip, and so forth. Finally, a preliminary survey of technology acceptance is presented, assessing the usefulness and ease of use of this novel system as perceived by students. Table 1 indicates the requirements for overcoming difficulties, and the relationships between the requirements and the proposed solutions.

COMPUTER DESKTOP CAMCORDER

Many commercial products can record computer desktop activity and sound from a microphone as videos in standard AVI formats, such as Lotus[TM] ScreemCam (1), TechSmith[TM] Camtasia (2), and Hyperionics[TM] HyperCam (3). Recording computer desktop activity means that any learner's action on a computer desktop--such as opening a file, executing a program, showing slides for a lecture, using a browser, and even moving the mouse-pointer--are videotaped. Furthermore, explanations to learners while they are at the computer can be recorded through the microphone of a computer as the audio track of a teaching video. Since a computer desktop or screen is like the viewfinder of a video recorder, the software is called a desktop camcorder or a screen camera.

Originally, desktop camcorder tools were used to train users of new software with video-based instruction. Since learners may watch the produced teaching video without a live instructor, desktop camcorder tools should provide functions to support an instructor emphasizing the focus on computer screen. Hence, most desktop camcorder tools include the following features:

1. Learners can determine the video capturing area by selecting a screen, window, or fixed region as video input.

2. Learners can zoom in or zoom out during a capture.

3. Learners can annotate a caption on the captured video.

4. Learners can use a microphone to input audio during a capture; hence, explanations of a topic can be recorded while the video is producing.

The TechSmith[TM] Camtasia computer desktop camcorder is used for demonstration. Figure 1 illustrates the use of the TechSmith[TM] Camtasia to capture slides, mouse movements, and audio from a microphone input as teaching videos. Learners can press the capture hotkey or click on the red record button on the toolbar to start a recording. During the recording, the Camtasia recorder can be minimized to avoid obstructing the learners' lines of vision. Learners can stop capturing by activating a hotkey, and a dialog box will be activated to save the captured video to a file.

[FIGURE 1 OMITTED]

In general, the captured videos are produced in standard video formats. However, the captured videos can be easily transformed into proper streaming formats that depend on Microsoft[TM] Windows Media Server or RealNetworks[TM] Realmedia Server for streaming media services. Both companies provide freeware to transform videos in standard AVI formats into proper streaming formats. Fortunately, TechSmith[TM] Camtasia tools support both transformations. Consequently, users can easily produce streaming videos regardless of which technology they use.

This study proposes a mechanism for a distance learner to contribute teaching videos without a video camcorder. Most important of all, such teaching videos for peer learning come from the perspective of learners rather than that of an instructor. On the other hand, the features of a conventional web-based distance learning systems do not encourage learners to contribute teaching videos for peer learning. Traditionally, an instructor provides teaching videos, and learners discuss those teaching videos in a learning forum. Thus, most distance learners expect teaching videos contributed only by an instructor rather than by their learning partners or even by themselves. To resolve this problem, the current study applies a collaborative learning strategy with the aim of intensifying learners' interdependence.

REFINED COLLABORATIVE LEARNING STRATEGY

The elements of a collaborative learning strategy include heterogeneous grouping, interdependence, individual accountability, and group processing (Warschauer, 1997). These elements ensure that learners try to help each other. Previous research has developed a learning environment for implementing a collaborative learning strategy (Persico & Manca, 2000). However, no asynchronous, collaborative learning environment has been based on streaming video-based techniques instead of web-based or videoconferencing techniques. Furthermore, collaborative learning strategies are required to promote learning motivation and learning performance when a novel technique is applied to distance learners.

Before developing of a prototype system, it is necessary to consider which collaborative learning strategy is most suitable for a computer desk-top camcorder technique. It is difficult to prove that one collaborative learning strategy is more suitable than others, because no literature reports the use of computer desktop camcorder technique in collaborative learning activities. However, we have used research to estimate the efforts of technical requirements to implement collaborative learning strategies in an internet-based environment (Chang, 2001). Furthermore, the selected collaborative learning strategy should encourage distance learners to contribute instructional videos for the training of fellow learners from their own individual perspectives. Considering these issues, the Jigsaw II collaborative learning strategy was chosen because its expertise-group mechanism supports distance learners' mastery of a topic. Then members of an expertise group are encouraged to share their knowledge in a streaming video style.

Briefly, there are 10 steps to implement the original Jigsaw collaborative learning strategy in the face-to-face classroom situation, including (Aronson, & Patnoe, 1997):

 1. Divide students into 5- or 6-person jigsaw groups.
 2. Appoint one student from each group as the leader.
 3. Divide the day's lesson into 5-6 segments.
 4. Assign each student to learn one segment.
 5. Give students time to read over their segment at least twice and
 become familiar with it.
 6. Form temporary "expert groups" by having one student from each
 jigsaw group join other students assigned to the same segment.
 7. Bring the students back into their jigsaw groups.
 8. Ask each student to present her or his segment to the group.
 9. Float from group to group, observing the process.
10. At the end of the session, give a quiz on the material.

Learners, who only read one part of instructional material rather than the entire one, are inconvenient to design their teaching videos because Aronson's Jigsaw collaborative learning strategy, Jigsaw I, requires that instructional material be divided into appropriate topics (the third step in aforementioned ten-steps-list). Hence, Jigsaw II, refined from Jigsaw I by Slavin, is used in the video-based learning forum (Slavin, 1995). Although there are many differences between Jigsaw I and II, the latter was selected to enable learners to get an overview of the day's lesson before designing their own instructional videos. Furthermore, there is no competition between groups at the end. Hence, it is called as a refined collaborative learning strategy from Jigsaw II. When an instructor wishes to construct a video-based learning forum on the Web, some steps are important to encourage learners' contribution of video clips in the Jigsaw II collaborative learning strategy. First, distance learners should be assigned to heterogeneous groups or jigsaw groups according such characteristics as culture, mother tongue, and grade-point average. Then, each distance learner in a heterogeneous group is respectively assigned to a study group for mastering a specific learning topic. The learners collaboratively study a learning topic to support mutual progress becoming an expertise group while they are designing the teaching video on the topic on which they have become experts. Finally, each learner of an expertise group returns to his/her heterogeneous group to present a well-organized, video-based report for teaching the topic on which he/she has become an expert.

In Figure 2, the required processes of Jigsaw II are illustrated, including heterogeneous groups and the use of expertise groups. Figure 2 depicts a situation of dividing nine learners into three groups for collaborative learning using Jigsaw II. From the perspective of a distance learner, a computer desktop camcorder supports his/her tutoring of learning partners in a heterogeneous group by video-based instruction. A learner, especially one who is not learning well in a traditional setting, is more convincing to others with the use of an instructional video contributed by the expertise group. To construct a video-based collaborative learning activity on the Internet, a distance-learning instructor first sets up a media server as a streaming video provider and a channel for learners to upload video files from their own PC at home or a remote learning center. Then the members of an expertise group not only study a learning topic, but also plan a teaching video. Finally, these members produce a teaching video on a computer desktop camcorder and upload the video to a media server. Consequently, video-based instruction in a heterogeneous group becomes feasible. The system integrating Jigsaw II collaborative learning with video-based instruction is depicted as Figure 2.

[FIGURE 2 OMITTED]

"Should there be various media? Does text suffice for the presentation, or can we expect a better student reaction from different media? Will the media enhance the student motivation, his/her comprehension of the contents, other learning related matters?" (Okamoto, Cristea, & Kayama, 2001). It is important to determine whether various media is necessary or if the text suffices for the presentation and whether the media enhances learners' motivation and comprehension. Learners' behavior records of various media, statistical description of records, and pedagogical understanding of the records can help determine the answers to such questions. However, because logs of a media server are huge and unstructured, it is difficult for a distance learning instructor to extract this information. Consequently, a framework is proposed to assist an instructor in acquiring the required summary information.

ANALYZING VIDEO-RELATED BEHAVIORS

Previous researchers have investigated the effects of teaching strategies on learners' behaviors in a video-based distance learning system (McMahill, 1993). Their studies provide guidelines for a distance learning instructor to promote learning outcomes in a video-based distance learning environment (Milheim, & Lavix, 1992). However, a distance learning instructor requires more than guidelines for using videos. For instance, the instructor may need to know who has watched a specific video, what a learner did while he/she watched a specific video, and how many times a specific learner watched a specific video. Furthermore, the instructor needs to examine the access logs of teaching videos to determine learners' behaviors in order to evaluate individual accountability and group processing during collaborative learning. Streaming techniques enable a distance learning instructor to observe how learners utilize videos for learning.

While streaming videos are gradually transmitted to learners, the clip of a video can be played, paused, stopped, rewound, and forwarded. As a learner triggers any one of those actions, the media server generates a log record, formally called operational data. Table 2 depicts field names contained in original log records of a media server. Some fields that are relevant to this article are briefly described. For instance, c-starttime denotes a time stamp for a record generated in logs; x-duration indicates the time interval between two forward, rewind, pause, or stop events; c-rate denotes the status of streaming before a command. One can find detailing descriptions about the log file format of a media server in the URL: http://msdn.microsoft.com/

As Table 2 shows, the usage of a media server is completely monitored and recorded in log files. From a different viewpoint, the abundant operational data is a reflection of learners' learning processes in a stream video-based distance learning environment. Hence, a distance learning instructor can understand learners' status by analyzing logs of a media server. For instance, the analysis results can reveal whether or not a learner repeats a clip, if a learner skipped a clip, and if a learner paused a clip. However, it is extremely difficult for an instructor to extract the required information because logs of a media server are huge and unstructured.

Some commercial products provide functionality for analyzing logs of a media server (4). Those products summarize the usage of a whole media server. For instance, they easily provide summarizations of the most downloaded clips, the average percentage played by clip, the top users, clip requests by day of the week, and user interactions by clip. However, the results generally do not show a learner's behaviors when using videostreaming. In a video-based distance learning situation, the instructor is concerned with each learner's status rather than the usage of a media server. For instance, the behavioral records of a distance learner who always plays a clip less than 10% of the full length may indicate that he/she is not motivated by opportunities to collaborate with peers. Hence, the instructor may want such information as how many clips a learner watched, how many clips a learner contributed, and a learner's average percentage played by clip. Consequently, novel architecture for interpreting media server logs is necessary for distance learning purposes.

Data cube technology enables an instructor to analyze learners' behaviors from various perspectives (Chang, Chen, & Ou, 1998). A data cube is a "subject-oriented, integrated, time-variant, and nonvolatile collection of data in support of management's decision making process" (Inmon, 1996). The data cube model consists of a fact table and some dimension tables. The attribute in a fact can be explained by various detail levels in the related dimension. For instance, by cube repository the instructor can find the number of contributing clips of a learner, a group, or all learners. Acquiring this information does not necessitate learning a query language or complex spreadsheet techniques. For instance, the product of WebTrends[TM] (www.WebTrends.com) can provide a report, called user interactions by clip, of identifying clips with the most replays, rewinds, and forwards. It first finds out the number of replays, rewinds, and forwards with every clip; then it sums the requests of each clip. Finally, sums of clip requests are divided by the number of interactions to calculate the average number of user interactions by clip. The report generally shows clips in descending order by average number of interactions. Table 3 illustrates a sample report of user interactions by clip. The clips are sorted by the third column, indicating the average number of interactions. Although the clip-request number of forloop.asf is less than that of selection.asf, the rank of forloop.asf is higher than that of selection.asf.

Table 3, a summarized report of all learners' behaviors, provides meaningful information to a media server manager. However, this table would not be helpful to an instructor who wants to observe the relation between a learner's behaviors and interactions of a clip because it does not allow him/her to trace the details of every learner's behaviors. For instance, an instructor cannot know whether or not there is a learner who never requests the clip one (i.e., mms://media.dyu.edu.tw/recursive.asf). The "drill-down" operation of data cube technology does enable an instructor to understand a report from an additional dimension. For instance, the number of clip requests in Table 3 can be drilled down through the learner dimension. Table 4 illustrates that such information can be generated by cube repository on the assumption that there are three groups in the streaming video-based learning environment.

Although a distance learning instructor can use cube repository to know how a summary report is calculated, it is still difficult for an instructor to discover learners' behavior patterns by tabular reports. For instance, all items in Table 2 can be drilled down along learner dimension, which is comp5osed of three calibration types (all, groups, and learners). If there are 20 students in the streaming video-based learning environment, the total number of detailing items--not including the time dimension, interaction dimension, or date dimension--is three hundred, coming from 20(students) * 3(items) * 5(clips). If the date dimension is considered, four calibrations--including day, month, quarter, and year--are added. The total number of detailing items will explode and diagnosing learners' behavior patterns becomes too complex. Fortunately, visualizing the drilled down report, shown as the bottom left picture of Figure 3, supports an instructor in identifying potential behavior patterns. For instance, an instructor can realize that the number of clip requests of Group I is not in descending order with a difference greater than three (16-19). Hence, an instructor can further explore the potential unusual behavior in Group I by analyzing how every group paused, forwarded, or rewound a video clip. From the bottom right picture of Figure 3, an instructor can tell that the clip interaction of Group I was as usual. Hence, clip interaction is not a key point to explain why the clip requests of Group I are special between clip 2 and clip 3.

[FIGURE 3 OMITTED]

PRELIMINARY EVALUATION

The evaluation presented in this section attempts to analyze distance learners' acceptance of a desktop camcorder for collaborative learning. This evaluation does not attempt to propose a validated means to improve learning performance or collaborative learning efficiency when using of the desktop camcorder tool. However, a key objective is to demonstrate that asynchronous video-based communication can be used not only for instruction, but also for learning. This experiment was conducted at Da-Yeh University, where a sample of 37 students (18 females and 19 males) were studying programming design in the Information Management Department during a semester from February 2003 to June 2003. Those students had not previously used a desktop camcorder or similar software. They took a training course on using a desktop camcorder before beginning the next collaborative learning activity. Furthermore, the Jigsaw II collaborative learning strategy was explained in detail to the learners. Then the instructor assigned a project to every Jigsaw group as their task. During the period of collaborative learning, students could communicate with each other only through the Internet. The objective was to determine how useful and easy to use the distance learners thought a desktop camcorder was.

A questionnaire used in Davis's Technology Acceptance Model (TAM) was chosen as the evaluation tool (Davis, 1989). TAM supported the managers' understanding of the process of adopting new technology: "(TAMs) ... are widely accepted among the MIS research community as tools for evaluating information system applications and predicting usage" (Doll, Hendrickson, & Deng, 1998). In other words, the results from a TAM can provide evidence to be weighed for or against adopting a desktop camcorder in collaborative learning. Originally, there were 12 items classified into two categories (perceived usefulness and perceived ease of use) in Davis's TAM questionnaire. In the present evaluation, one overall ranking question was added to the original questionnaire. The contents of the questionnaire used in this evaluation are as follows (Table 5):

Each question in Table 5 was scored on a 7-point scale, with strongly disagreed (1) and strongly agreed (7) as the two endpoints. Previous research studies reported good reliability of TAM. Similarly, an alpha reliability estimate of .82 was obtained for the 12 items. The reliability estimates for the individual scales were .86 (usefulness) and .73 (ease of use). All 12 items received agreement ratings of over 62% slightly agreed to strongly agreed on the survey. The six ease of use questions received agreement ratings of over 78% slightly agreed to strongly agreed. The results show that on the average the respondents' (learners') perceived usefulness (Q1~Q6) was relatively lower than their perceived ease of use (Q7~Q12). Through informal interviews, the students indicated that video-based instruction was not critical because they were not always in a distance-learning situation. One student said, "It (teaching video) will be very useful in a heterogeneous group if we cannot meet each other (learning partners in a heterogeneous group) after finishing the project. Hence, the video would become very useful to teaching my learning partners." While this preliminary study was not without limitations, it proved the usefulness and ease of use of the desktop camcorder tools for collaborative learning.

CONCLUSIONS

Although streaming techniques have enabled learners' accessibility to educational multimedia on the Internet in recent years, such usage firmly supports the instructors' perspectives. In other words, most teaching videos and audio-visual aids that are available on instructional web sites are contributed by instructors or instructional teams. Although students could contribute video clips in some studies, a video camcorder was needed to record the required footages. However, it is not realistic to presume that every learner can acquire a video camcorder. Hence, this report has introduced the use of a software application, a computer desktop camcorder, to allow distance learners to contribute teaching videos for collaborative learning without a video camcorder. It is proposed that a computer desktop camcorder is integrated into a well-known collaborative learning strategy, Jigsaw II, as additional support for heterogeneous group learning. Furthermore, a data cube model is proposed to support a distance learning instructor extracting pedagogical meanings contained in the large amount of teaching video logs. After distance learners were trained in the use of a desktop camcorder, the results of a preliminary evaluation of technology acceptance demonstrated that most learners agreed that a computer desktop camcorder was useful and easy to use for collaborative learning. Most important of all, this study has demonstrated the feasibility of using audio-visual aids contributed by distance learners.

Table 1 Relationships Between the Proposed Solutions and Requirements

PROBLEMS               REQUIREMENTS                      SOLUTIONS

Contributor and usage  * Learners as video contributors  Computer
                       * Without a video camcorder       desktop
                                                         camcorder
Strategy refinement    * Encourage contributing videos   Refined Jigsaw
                       * Collaborative learning          II
Statistical and        * Pedagogical records of the      Data cube
  analysis               usage of the instructional      repository
                         videos
                       * Observable behavioral patterns

Table 2 A List of Field Names in Logs of a Media Server

Audiocodec               avgbandwidth        c-buffercount
c-bytes                  c-cpu               c-dns
c-hostexe                c-ip                c-os
c-osversion              c-pkts-lost-client  c-pkts-lost-cont-net
c-pkts-lost-net          c-pkts-received     c-pkts-recovered-ECC
c-pkts-recovered-resent  c-playerid          c-playerlanguage
c-playerversion          c-quality           c-rate
c-resendreqs             c-starttime         c-status
c-totalbuffertime        channelurl          cs(referer)
Cs-uri-stem              cs(user-agent)      Date
Filelenghth              filesize            Protocol
s-cup-util               s-dns               s-ip
s-pkts-sent              s-totalclients      sc-bytes
Time                     transport           Videocodec
x-duration

Table 3 An Example of "User Interactions by Clip" Report

Clip                                    Average       Clip      Plays,
                                        Number of     Requests  Rewinds,
                                        Interactions            or
                                                                Forwards

1 mms://media.dyu.edu.tw/recursive.asf        3.42       90      308
2 mms://media.dyu.edu.tw/pointer.asf          3.16       45      142
3 mms://media.dyu.edu.tw/array.asf            2.39       41       98
4 mms://media.dyu.edu.tw/forloop.asf          2.32       31       72
5 mms://media.dyu.edu.tw/selection.asf        1.53       32       49

Table 4 Drill-Down Clip Requests Along Learner Dimension

Clip  Group I  Group II  Group III  Clip Requests

1      45        26         19           90
2      16        15         14           45
3      19        13          9           41
4      12         8         11           31
5      15         9          8           32

Table 5 Contents of Questionnaire for Evaluation

USEFULNESS   Q1. Using a computer desktop camcorder in collaborative
             learning would enable me to accomplish tasks more quickly.
             Q2. Using a computer desktop camcorder would improve
             collaborative learning performance.
             Q3. Using a computer desktop camcorder in collaborative
             learning would increase my productivity.
             Q4. Using a computer desktop camcorder would enhance my
             effective ness on collaborative learning.
             Q5. Using a computer desktop camcorder would make it easier
             to do collaborative learning.
             Q6. I would find a computer desktop camcorder useful in
             collaborative learning.

EASE OF USE  Q7. Learning to operate a computer desktop camcorder would
             be easy for me.
             Q8. I would find it easy to get a computer desktop
             camcorder to do what I want it to do.
             Q9. My interaction with a computer desktop camcorder would
             be clear and understandable.
             Q10. I would find a computer desktop camcorder to be
             flexible to interact with.
             Q11. It would be easy for me to become skillful at using a
             computer desktop camcorder.
             Q12. I would find a computer desktop camcorder easy to use.

Acknowledgements

This research was supported by the National Science Council of Taiwan under contract number NSC 90-2520-S-212-001 and NSC 89-2745-P-212-001. The author wishes to express appreciation to Dr. Cheryl Rutledge for her editorial assistance.

Notes

1. Lotus ScreenCam. URL: http://www.lotus.com/home.nsf/rightframe/1screencam

2. TechSmith Camtasia. URL: http://www.techsmith.com/

3. Hyperionics HyperCam. URL: http://www.hyperionics.com/hc/index.html

4. Hit List Commerce. URL: http://www.marketwave.com/ Webtrends Log Analyzer. http://www.webtrends.com/

References

Aronson, E., & Patnoe, S. (1997). The jigsaw classroom: Building cooperation in the classroom (2nd ed.). New York: Addison Wesley Longman.

Benney, A. (2001, June). Creating an active learning environment using digital video: What I did and how I did it. Proceedings of ED-MEDIA 2001 (World Conference on Educational Multimedia, Hypermedia & Telecommunications), (pp. 133-138), Tampere, Finland.

Chang, C.K. (2001). Refining collaborative learning strategies for reducing technical requirements of web-based classroom management. Innovation in Education and Training International, 38(2), 133-143.

Chang, C.K., Chen, G.D., & Ou, K.L. (1998). Student portfolio analysis by data cube technology for decision support of web-based classroom teacher. Journal of Educational Computing Research, 19(3), 307-328.

Davis, F.D. (1989). Perceived usefulness, perceived ease-of-use, and user acceptance of information technology. MIS Quarterly, 13(3), 319-338.

DeMartino, D.J. (2001, June). Use of video mediated instruction in teaching American Heart Association emergency cardiac care training courses. Proceedings of ED-MEDIA 2001 (World Conference on Educational Multimedia, Hypermedia & Telecommunications), (pp. 388-392), Tampere, Finland.

DiMattia, C. (2002). VideoPapers. Hands On!, 25(2), 12-15.

Doerksen, T., Mattson, M., & Morin, J.E. (2000). Streaming video: Pedagogy and technology. Proceedings of ED-MEDIA 2000, (pp. 1627-1629), Montreal, Canada.

Doll, W.J., Hendrickson, A., & Deng, X. (1998). Using Davis's perceived usefulness and ease-of-use instruments for decision making: A confirmatory and multigroup invariance analysis. Decision Sciences, 29(4), 839-869.

Han, J., & Kamber, M. (2001). Data mining: Concepts and techniques. San Diego, CA, Academic Press.

Inmon, W.H. (1996). Building the data warehouse. New York: John Wiley & Sons.

Kozma, R.B. (1986). Implications of instructional psychology for the design of educational television. Education Communication and Technology Journal, 34(1), 11-19.

Levne, T.K. (1992). Going the distance: A handbook for developing distance degree programs. Washington, DC: Annenberg/CPB.

McMahill, J.M. (1993). Videotape distance learning courses: Administrative implications for colleges and universities. Ed Journal, 7(6), 16-20.

Milheim, C.L., & Lavix, C. (1992). Screen design for computer-based training and interactive video: Practical suggestions and overall guidelines. Performance & Instruction, 31(5), 13-21.

Okamoto, T., Cristea, A., & Kayama, M. (2001). Future integrated learning environments with multimedia. Journal of Computer Assisted Learning, 17, 4-12.

Pea, R.D. (2001). Integrating web-based video case studies and community tools for teacher professional development. In symposium entitled: "Defining the video case: What we know and how we (use video to) know it." Proceedings of the Annual Meetings of the American Educational Research Association, Seattle, WA.

Persico, D., & Manca, S. (2000). Use of FirstClass as a collaborative learning environment. Innovations in Education and Training International, 37(1), 34-41.

Repenning, A., Ioannidou, A., & Phillips, J. (1999). Collaborative use and design of interactive simulations. Proceedings of ACM Conference on Computer Supported Cooperative Learning (CSCL 99). Stanford, CA.

Salomon, G. (1991). What does the design of effective CSCL require and how do we study its effects? SIGCUE Outlook, Special Issue on CSCL, 21(3), 62-68.

Schluger, J., Hayes, J.G., Turino, G.M., Fishman, S., & Fox, A.C. (1987). The effectiveness of film and videotape in teaching cardiopulmonary resuscitation to the lay public. New York State Journal of Medicine, 87, 382-385.

Slavin, R.E. (1995). Cooperative learning: Theory, research, and practice (2nd ed.). Needham Heights, MA: Allyn and Bacon.

Warschauer, M. (1997). Computer-mediated collaborative learning--theory and practice. Modern Language Journal, 81(4), 470-481.

Zigerell, J. (1991). The uses of television in American higher education. New York: Praeger.

CHIH-KAI CHANG

Da-Yeh University Taiwan, R.O.C.

[e-mail]