GEMSTEAM

Principal Investigators:
Charles Steinfield, Department of Telecommunication
Erik Goodman, CASE Center, College of Engineering
Thomas Muth, Department of Telecommunication
Research Associates:
Carleen Maitland, Department of Telecommunication
Tim Hinds, CASE Center, College of Engineering
Hsun-I Hsu, Department of Telecommunication

Introduction

During the summer of 1996, the Michigan State University Department of Telecommunication and the Case Center for Computer Aided Engineering and Manufacturing undertook a preliminary study of the effect of communication technology augmentations on engineering task performance. This preliminary study is part of a larger program of research investigating methods for improving the efficiency and quality of design teams in geographically distributed contexts. The summer project involved four EDS interns working at MSU in a simulated corporate environment. We specifically investigated how augmented communication tools enhanced or hindered engineering teams' abilities to collaborate using computer aided design software. Engineering teams used the Unigraphics computer aided design software which, although requiring some time for our interns to learn essentially “on their own,” does provide some of the support needed for collaborative engineering design efforts. Communication modalities to facilitate collaboration over a distance included in-person meetings, email, copies of printed documents (simulating what one might obtain via fax), videoconferencing, electronic whiteboard, and file sharing. Our primary goal for this phase was to develop a set of methods for assessing the effectiveness of augmented communication tools complementing CAD software, and identifying work processes that can help engineering teams gain the greatest benefit from them. These methods will then be used in the subsequent phases of the research program.

The project began with Unigraphics training for the interns. Once accomplished, this was followed by two two-week design phases in which the engineers used the communication and engineering tools to solve a problem. Employing instruments such as pre- and posttest questionnaires, a daily communication log, and a review of the quality of the interns’ engineering designs, data was collected to assess the effectiveness of the various tools. Outcomes of interest were objective measures of the speed, accuracy, and quality of solutions, as well as members' perceptions of the solution quality, degree of consensus, and satisfaction with the collaborative process.

Project Design

Subjects & Environment

The four subjects were undergraduate and graduate students of Mechanical Engineering working as interns for EDS during the summer of 1996. The workspace for the interns, located in the Engineering Building at Michigan State, was established to simulate an office environment. Partitions were set up so that each intern had a work space similar to cubicles found in engineering offices. The partitions also enhanced the simulation of distance between the team members. The internships began in mid-May and the students were given 6 weeks to train on the Unigraphics computer aided design software package which they were expected to use during the engineering design phase of the experiment. During this training period the interns worked together. Once training was complete, one week was used to choose a project and gather technical requirements for that project. Again, interns remained in one group to ensure they were equally exposed to all information concerning the project. Immediately before the two-week design phase began, the interns were broken into two groups and randomly assigned to the two experimental conditions. After the two-week period (7/15-7/26), a presentation was made on the groups’ accomplishments. During this time period, a second project was chosen, again with the interns working together to ensure equal exposure. For the second project the group members were changed so that no two interns would work together again during the second phase. The groups were again randomly assigned to the experimental conditions. This second design phase occurred from 8/5-8/16, with the interns reporting their results during the subsequent week.

Design Projects

The project chosen for the first two-week period was Robot Wars, a competitive mechanical sporting event featuring radio-controlled robots in combat. The project was chosen because it offered a design problem that could be subdivided into individual tasks. Additionally, extensive project specifications were included with the project description, which was downloaded from a Web page. To ensure that the project could be completed in the two-week time frame and that the design teams were working on similar, although not exactly the same, design approaches, it was necessary to establish additional design specifications for the project. This allowed a more meaningful comparison on the dependent variable of quality of the teams’ designs.

The project chosen for the second two week period was the 1996/1997 ASME Student Design Contest project which involved moving two ping pong balls and a golf ball from a platform to a box. This project was obtained from the ASME Web site and, similar to the first project, included detailed design specifications. No additional design specifications were needed to ensure completion within the two-week time frame.

Simulated Distance

To accomplish the project’s goals it was necessary to simulate an environment in which design teams worked from geographically distributed sites. During both of the two-week design periods the interns worked in such a way as to simulate communication patterns of teams that had one member in Western Europe (France) and the other team member in the U.S. in the Central Standard Time zone. It was assumed that this would create work days that overlapped by approximately one hour at the hypothetical end-of-the-work-day. It was also assumed that on a global engineering project the team members would dedicate this one hour of overlapping time to communicating with their project partner. The “French” team members would therefore begin their day by receiving documents from their American counterparts. The “French” team member would spend the day working on the project and then at the end of the day would have the opportunity to consult with their partner. The American team members would start the day with the consultation and end the day by sending off the altered files.

To make the simulation as realistic as possible, it was necessary to establish rules for communicating. First the interns were instructed that they were not allowed to talk to one another except for the designated one-hour time interval and then only through the allowed communication modalities. Sending email or leaving voice mail was allowed to occur throughout the day. “French” team members were instructed to check their email in the morning to get the files left by the Americans the night before. Since email received during the day would ruin the time simulation (as their American counterparts should be sleeping), the American interns were restricted to receiving it only during the one-hour overlap.

Communication Modalities

Although the goal of the project was to create communication differences between the groups, we believed it was necessary to create the differences while maintaining a realistic work environment. Team A was given augmented communication modalities that may or may not be available to an engineering team in industry. Team B, the group without augmentation, was provided communication tools usually available in an engineering office. The following chart shows the communication modalities that were assigned to the two groups.

The differences between the groups were in their methods of sharing computer files, access to the PictureTel video conferencing system, and access to the electronic white board. Here, the differences in file sharing and FTP were that FTP was accomplished by a method that increased the probability for protocol conversion problems and decreased the transmission speed. File sharing simulated a process that would occur at a company with a server accessible from both work locations. The directory used for file sharing in this case resided on a server in the Engineering Building. This reduced the difficulties in transmission and gave the interns access to support personnel. Sharing files required both users to have some familiarity with UNIX as it was necessary to change the permissions on the document. FTPing did not require users to change permissions on files. The FTPing was accomplished via the Teamgems server which is located at an EDS site remote from the MSU campus. Using a remote server increased the

transmission distance and hindered access to support personnel.

Video conferencing, a communication modality available only to Team A, was provided using PictureTel software running on a PC with an ISDN line. The electronic whiteboard was provided using Live Share Plus, an add-on to the PictureTel software. The electronic whiteboard is a free-hand sketching and on-screen text tool. It allows two locations to view a sketch that has been created by both parties, although only one site can edit a document at any given time. The sketching is performed by moving a mouse to direct a pen on the screen. This is a task not easily mastered to produce sketches comparable to those that could be generated by hand with a pencil and paper. The whiteboard also allows the user to place fonts on the screen in locations directed by the cursor. Delays occurred in providing some of the communication modalities. During the first design phase even the Team A interns only had access to phone, voice mail, sharing of printed documents through use of office mailboxes (simulating fax), email, and electronic file sharing. Video conferencing became available only during the last three days of phase 1 and the electronic white board was not functional. During the second design phase the interns maintained access to the above mentioned communication modalities, with video conferencing and the electronic white board available to Team A for the entire duration.

Instruments

To gather data on the interns’ daily communication patterns and perceptions of the usefulness of their tools and activities, a diary form was developed. The form asked them to record each communication event, which communication modality was used, and the duration and usefulness of the communication. This form was made available to the interns as a Web page linked to the Teamgems page on an EDS file server. The form was password protected and the data the interns entered into the form was sent via email to the graduate research assistant in charge of collecting data when the intern pressed the “send” button located on the form. In order to create the form in html format and have the data emailed, it was necessary to have access to a cgi_bin directory on a file server. HTML forms refer to an executable program or PERL/TCL script file which parses and formats the data and then mails it. The interns had considerable input to the physical design of the form to ensure their maximum satisfaction with the tool they would use every day. The form was tested during the week in which design specifications were being gathered. After they used the form for several days, a training meeting was held in which interns were able to ask questions about using the form and were given explanations of variables on the form. The final design of the form presented the options for most variables as text with an adjacent button to make filling out the form as easy as possible.

To assess the interns’ perceptions of their abilities to use the various communication modalities and to assess their attitudes towards group work in general, a pretest questionnaire was administered. The instrument contained 35 questions and was given to the interns after they were assigned to their groups but before the design phases of the project began. A posttest questionnaire was administered to assess the interns’ perceptions of how well they had worked together and their perceptions of the quality of their work. The posttest questionnaires were administered immediately following each two-week design period.

Results

Pretest Questionnaire

The pretest questionnaire for the first design project revealed that the interns were a fairly homogenous group. They showed interest toward the project, felt knowledgeable about the topic, and felt the project would be difficult. Differences did occur in their opinions on how long the project would take to complete. One member of Team B felt the project would take more than 3 weeks although only 2 weeks were scheduled. The other interns felt the project would take between 2 and 3 weeks. In terms of their teammates, the interns indicated they did not know one another before the summer and they felt they could work together well. They also expressed the opinion that one or both team members were knowledgeable about the project. In response to questions about working as a team, one intern disclosed that he preferred to work alone. The remainder of this intern’s responses, as well as the other group members’ responses, demonstrated agreement on the importance of keeping members informed, reaching consensus, and allowing all members to participate in the assigned task. All four of the interns also indicated that they usually speak up at meetings and that they only occasionally see each other outside of work, if at all.

The interns’ experience with different technologies varied. Each of them indicated they were daily users of email and had computer-aided design experience, although none had had experience on Unigraphics. The variance in their responses occurred in their use of video conferencing and exchanging files using FTP (file transfer protocol). Experience with these technologies varied from never having used them before to frequent use. The final section of the pretest questionnaire asked the interns to rank five characteristics of a team-designed engineering project. In their rankings all of the interns agreed that customer satisfaction was their first priority. From there three out of the four interns ranked working as a team as number 2 and consensus about the solution as number 3. The 4th and 5th rankings were split between goals already mentioned and the remaining two, which were speed and technical quality.

Unlike the first design project, in the second design phase the interns felt they had some or no experience with the project topic. Despite this lack of experience they did feel they were somewhat knowledgeable about the topic and indicated they were somewhat interested. The interns were split on their perceptions of the difficulty of the project, with Team A seeing it as not at all difficult and Team B anticipating it to be difficult. All of the interns felt the project would take between 1 and 2 weeks. The interns also were split on their opinions of their team's knowledge about the project. Team A and one member from Team B felt two or more people were knowledgeable. One member from Team B felt no one was knowledgeable. All interns anticipated they would work quite well with their team members and would enjoy the encounter.

After the experience of the first project, changes can be seen in the interns’ attitudes towards keeping other team members updated. While two of the three still felt it was important one intern strongly disagreed with a statement that said updates were important. The other member of Team B felt that groups often waste time keeping each other informed while the other two disagreed with this statement. When asked about working alone, the one intern in Team A indicated he preferred working alone. Both members of Team B preferred working in groups. Team A considered achieving consensus a waste of time while Team B disagreed with this statement. Everyone agreed that it is important that all participants have input to the problem solution. The interns continued to keep their relationship restricted to work, never seeing one another outside the project. The interns experience with technologies had changed and they were now all able to indicate that they had video conferencing experience. There was still, however, variance in their responses to having had FTP experience. The member of Team A was more experienced than the members of Team B.

Changes also occurred in the interns’ perceptions of desired project outcomes. The interns still agreed that customer satisfaction was the highest priority while technical quality was the next highest. Continuing to work as a team was third, speed was fourth with consensus ranking fifth.

Daily Log

The data available from the interns’ daily log sheets shows that they used the various media with the following frequencies. The relative numbers of events are important here as the chart represents accumulated data from each intern. The result is that in some cases communication events were double counted, as when the event was reported by both interns.

Differences occurred in the communication patterns of Team A and Team B, with email being the most common mode of communication for Team A. This team decided to update one another on the project’s status through daily email messages. This system was developed by the team as part of their approach to the project design. Team B used the telephone as its most popular mode of communication and rarely emailed one another.

In addition to frequencies of medium use, the daily diaries also allowed the interns to judge the usefulness of the communication event. The usefulness was judged by using a 5 point Lickert type scale with 1 indicating low usefulness and 5 indicating high usefulness of the communication event. Team A rated telephone and fax as more useful than Team B, although Team A used the telephone and fax less than Team B.

Due to the loss of one intern during the second design phase, it is not possible to report figures in the same manner as the first design phase. It was possible, however, to gain important information from the interns’ daily logs. The daily logs showed that Team A, now with video conferencing available, employed this communication option and found it to be very useful. However, for both teams, face to face was the most frequently used form of communication. For Team A the second most frequent form of communication was both telephoning and video conferencing. For Team B the second most frequent form of communication was faxing, with telephoning being the third most common form. Team A rated email with the highest utility, with video conferencing second, while Team B rated fax and telephoning with the highest utility scores.

Posttest Questionnaire

The Posttest Questionnaire measured the interns’ perceptions of their group experience. The items were divided into five sections: How Well the Team Worked Together, Views on the Group’s Solution, Satisfaction with the Group, Views on the Team’s Communication and Team Leadership/Organization. From the first design phase, the overall questionnaire produced little variance in the interns’ responses.

In the first section on working together, the interns responded that they worked efficiently and functioned well as a group. From the second section it was revealed that the interns felt that their engineering designs, although not perfect, were perfect for what was needed. Satisfaction with their groups was high with all of the subjects agreeing or strongly agreeing that they were satisfied with their group work experience and indicating they would like to work with their team member again. They also indicated that consensus about the project solution was achieved.

In terms of their team’s communication, the fourth part of the instrument showed that all group members agreed their ideas were given attention by the other group member and that everyone contributed equally to the project. There was also agreement that coordinating with other group members was not difficult. The interns’ perceptions differed, however, on how easy it was to communicate. One member from each team felt it was easy to communicate while the others felt it was difficult. Variance was also present in the interns' responses to a question concerning working alone and communicating only at the end of the project. Three of the four felt that they worked together with continuous communication, while one intern from Team B felt that they worked alone and only communicated towards the end. Team A also disagreed on whether one person dominated the conversations. One member strongly disagreed while the other neither agreed or disagreed with this statement.

The final section of the questionnaire was designed to assess the interns’ perceptions about team leadership. The results from this section demonstrate a difference in interns’ perceptions of leadership among the group members. Team A had one intern who indicated that it was clear from the beginning who was the leader while the other member disagreed. Team B had one member neither agreeing or disagreeing and the other member strongly agreeing with the idea that a leader was evident from the beginning. On a subsequent leadership question, one which asked if the group had one person who functioned as a leader, the interns were consistent with their previous responses. However, to an item that stated “no one was really leading our group,” both members of Team B were in agreement with the statement. Team A, being consistent with their previous responses, were split on their agreement with this item.

Team A felt the team did not work efficiently while Team B was split with one member agreeing and the other disagreeing.. All disagreed that they were wasting time. They all agreed that they were able to quickly figure out what everyone needed to do and that there time was spent on necessary activities. Team B felt they functioned well as a group. Both members of Team B felt they had the right mix of skills to do the project. And interns either agreed or neither agreed or disagreed that their solution was exactly what was needed. Team A felt there were no flaws in their solution while Team B agreed that were some flaws. The members of Team B indicated they would like to work with their team again while Team A neither agreed or disagreed. Team B agreed they were satisfied with the experience of working in a group and that there was consensus on their project design. The members of Team B felt the group listened to their ideas. Although they disagreed about the ease of communication, with one member believing it was easy and the other, not. For Team B, coordination was not a problem. Both members of Team B disagreed that they worked alone, only communicating at the end of the project. Both members of Team B agreed that everyone contributed equally and that no one tended to dominate the group's discussion. Both members of Team B felt that there was no leader in their group.

Characteristics of Engineering Designs

Outcomes of interest for the engineering designs were the speed, accuracy, and quality of the solutions. The judgment of the interns’ designs was provided by the Associate Director of the Case Center for Computer-Aided Engineering and Manufacturing, a person familiar with the technical specifications of the projects and having extensive engineering experience in industry.

Design Project 1 - Robot Wars

Team A - High Tech - Lawnmower Design

Overall Design/Presentation

First and foremost, this design was a very unusual solution to the problem. The idea of a spinning blade to immobilize the opponent vehicles is a concept not yet seen in actual competition. Also, the conical shield refined a design concept that had not been previously exploited. However, the design appeared to be vulnerable to maneuverability difficulties when positioned near a corner of the contest arena. The level of detail in the final design was very complete. The project team utilized sound engineering principles in developing a simple design for their vehicle. The final design was presented in a clear and organized manner utilizing good presentation materials

Team Organization

It was clear from the presentation of the design materials that one team member dominated in development of the design concept and in detailing the final design. This is not uncommon in most engineering teams and was expected in this case given the level of technical expertise possessed by one of the team members.

Engineering Research/Analysis

The team did adequately research various components of their design, i.e. motors, batteries, servo controllers, battery life, etc. Although, they did not present any engineering analyses. The team should have calculated and presented results for critical items such as maximum vehicle speed, blade speed, blade inertia, optimal cone thickness, optimal blade thickness, energy required to damage/destroy a tire, etc. Given that the total robot weight was below the 9.1 kg target, the design, although very good, was far from being optimized.

Team B - Low Tech - Pincher Design

Overall Design/Presentation

This design utilized a gripping mechanism as an offensive weapon, which is a concept which has been seen in previous competition. However, the pincher was coupled with a highly effective defensive feature, an asymmetric shell. This design would fare moderately well in actual competition, but would not be in the winner’s class. The level of detail in the final design was not as complete as was expected. Only assembly drawings of the final design were available, with no details of individual components. The project team utilized sound engineering principles in developing a somewhat complex vehicle. One problem area would be the drive train design. With its wheel base shorter than its track width, it would experience severe maneuverability problems. This team presented their final design in a clear and organized manner, utilizing good presentation materials

Team Organization

The team was evenly staffed in both talent and work load. Both members exhibited the same level of technical expertise and appeared to have shared equally in the total project work load. The members developed a good working relationship throughout the design project.

Engineering Research/Analysis

This team also did adequately research various components of their design, i.e., motors, batteries, servo controllers, battery life, etc. They also presented some engineering calculations. Specifically, calculations were presented for the pincher mechanism. But, just as their counterparts, they did not present any engineering analyses for critical items such as maximum vehicle speed, optimal shell thickness, etc. This robot’s weight was also far below the 9.1 kg target, indicating the design was not optimized.

Design Project 2 - ASME Student Design

Team A - High Tech - Lever Arm with Ramp Design

Overall Design/Presentation

This design utilized a lever arm with a trough device to transport the balls from the platform to a ramp extending to the delivery box. Of the two designs presented, this proposal showed the greatest potential for winning in actual competition. The critical features of the ball transportation mechanism are simplicity of design and the speed of a delivery mechanism. This design was quite simple in its approach to the problem. It incorporated a fast delivery device which created a sizable initial velocity for the balls. It also utilized gravity as a major component. The one area where this design lacked was in not containing a mechanism or device to stop the balls once they were delivered to the final crate. The level of detail in the final design was very complete. The project team utilized sound engineering principles in developing a simple design for their project. The final design was presented in a clear and organized manner utilizing good presentation materials

Team Organization

Both team members shared equally in the concept design stage for their project. However, due to the exit of one of the team members early in the detailed design phase, most of that work was completed by only one team member.

Engineering Research/Analysis

The team did research the standard components of their project -- i.e,. the motor. They also performed a fair amount of engineering calculation in determining those factors which would increase their chances in winning a competition. This was achieved by calculating various system parameters that would minimize the time balls would be in motion. Again, a stopping device should have been designed and analyzed to minimize the amount of time the balls would be allowed to bounce once in the final delivery box. There should have also been some supporting calculations showing how far the balls could be thrown down the ramp and the associated sliding vs. rolling estimates.

Team B - Low Tech - Ball Scoop with Ramp Design

Overall Design/Presentation

This design utilized a scoop mechanism that cradled the balls while transporting them up an incline to be delivered to a ramp extending to the final delivery box. As in the Team A design, Team B pushed the bounds of interpretation of the rules to their best advantage. The overall design of this transfer mechanism incorporated simple techniques such as utilizing gravity for a large portion of the energy required to deliver the balls. However, it also used a fairly complex mechanism to transport the balls to the ramp device. At first glance, this transport device appears simple enough. But, after a more thorough review, it contains features which make it fairly complex in nature. First, the device is string driven. Strings are not nearly as reliable in alignment as rigid links. The device also uses two strings, creating an unstable situation for fine adjustment. The scoop design had an additional area of concern, as it would attempt to force the balls to roll in opposite directions at the same time. This would create a situation where friction would consume a fair amount of energy in transporting the balls. With some time for adjustment and fine tuning, this design would perform well in the competition, but it would not be among the winners. The team presented very complete detail for their final design. They presented their final design in a clear and well organized manner utilizing good presentation materials.

Team Organization

This team was evenly staffed in both talent and work load. Both members exhibited the same level of technical expertise and appeared to have shared equally in the total project work load. The members developed a good working relationship throughout the design project.

Engineering Research/Analysis

This team also researched the major component of the design constraints; the motor. They also presented a considerable amount of engineering calculations related to minimizing the time required to deliver the balls. Some additional work would be necessary to optimize the design -- specifically, refinement of the scoop mechanism and development of a stopping system.

Analysis

Interns’ Perceptions

Effectiveness of the Tools for Engineering Design and Communication

As in any technology-supported work environment, equipment downtime can seriously curtail the efficiency of workers. In a global, concurrent engineering environment, equipment and system unreliability can be particularly crippling. Yet, downtime does occur, and our project was not immune to this problem. Summer maintenance work on campus computer networks caused some of the downtime, sometimes preventing interns from using the Unigraphics software, email, ftp, filesharing, videoconferencing and the whiteboard, and hindered some of our computer-supported data collection instruments. Nevertheless, this replicates the real world environment, and provides a preview of some of the difficulties we might experience when conducting research on truly distributed work teams, rather than simulated geographical dispersion.

The engineering tools available to the students included the Unigraphics software package. The interns found the system was difficult to learn compared to other design packages. However, they also found the system offered greater capabilities. During their design phases, the interns found that the layers and/or part file functions of the software were extremely helpful for collaborative work. The interns were able to share the file that contained their design, allowing each member to view the progress of the other member. Expertise with this tool varied among the interns and may have been responsible for some of the interns’ perceptions such as those related to leadership. The intern with the highest skill level perceived that during the first design phase his team had a clear leader while his teammate did not.

The impact of using the layers function of Unigraphics was highly dependent on the mode of communications being used. The team using file sharing found this tool very useful as sharing the file was quite easy. In the first design phase, where Team B had trouble using FTP, the layers function was not as useful.

The teams used the Unigraphics software only during the detailed design phase of the project. During the conceptual design phase, they preferred to draw sketches of their ideas by hand rather than putting in the effort of making a computer drawing. During this conceptual phase, the Team A interns found that holding the drawing up to the camera during a video conference was an effective way of conveying their ideas. During the design phase, however, video conferencing was less useful because they were working from their Unigraphics files, which could not be communicated over the video conferencing system. The interns felt that to redraw the document using the electronic white board was too difficult and not worth the time.

The use of the communication tools and therefore the patterns of communications, as well as the measurement of their use, were all somewhat affected by the unreliability of the computer system.

As previously stated, during the design phases, the groups were separated into teams with separate communication tools available to them. The primary difference was that Team A had video conferencing, the electronic white board, and file sharing in addition to the normal office communication systems. Morever, Team B used FTP in place of file sharing. Computer system unreliability was particularly a problem for Team B in their use of FTP. The server supporting FTP was at a remote location, and the interns did not have direct access to the server system administrators. As a result, Team B members relied on faxing drawings to one another and then discussed the drawings over the phone. This technical difficulty constrained Team B to continue synchronous communication. Each member of Team A, however, was easily able to share files, allowing them to view the progress made by their teammate and then simply communicate asynchronously by email.

Based upon the diaries and interviews with the interns, we found that Team A used videoconferencing mainly in the early conceptualization phase of each project. The whiteboard was not used at all, and interns complained that it was a rather clumsy attempt to replace freehand drawings useful for brainstorming. They instead resorted to holding up drawings in front of the video camera, and discussing them. In the later stages of the project, once work began on the Unigraphics system, video conferencing was less useful and interns expressed a desire for the ability to annotate CAD drawings from the simulated distant location. Again, the whiteboard was not deemed useful here.

Email was also used more by Team A than Team B (looking only at the time when both members of group were still on the project.). They actually integrated a daily email update into their work routines, and found this to be quite helpful. Without any formal process, Team B did not use email very much.

Both groups maintained frequent contact with people outside of their group; in particular, computer network support personnel and the project directors. In actual workplaces, work teams also have to maintain contact with external sources of information and support, and with supervisors. We cannot tell if any differences in performance were due to these extra-group communications, but the research literature does suggest that groups who fail to interact with sources of expertise, those supplying critical team inputs, or the consumers of the group's work, are more prone to failure.

In general, the data collection instruments employed in the study were easily understood and completed by interns. However, reliance on computer-supported data gathering increased our vulnerability to downtime. When the network was down, the Web-based diary forms could not be filled out, leading to some unreliability in the data.

Implications

Despite the preliminary nature of the study, several hypotheses are suggested that merit more systematic study in the next phase of this research. These are briefly highlighted below.

• Different communication modalities appear to be more or less useful at different stages in the design process. Video conferencing was an effective means of communication during the conceptual phase of the engineering process. It easily allowed the interns to share their drawings in real time. Once past the conceptual phase the engineers who were simulating distributed work locations found that Unigraphics, especially with its layers functionality, was an effective tool facilitating distributed collaborative work.

• Real time annotation would improve the collaborative capability of CAD software. The interns felt that the Remarks function of the Unigraphics software would greatly facilitate collaborative work in a distributed setting especially during the detailed design phase of engineering. Once the interns entered their ideas into Unigraphics, it became difficult to have synchronous conversations about the drawings.

• For email to successfully support collaboration, a process for integrating its use into daily project work is needed. Without such a process, one group hardly used email and did not report it to be very useful. With such a process, the other group found email to be very useful. Although we cannot rule out other causes, due to the small size of the study, the group relying on email did have slightly better project solutions.

• Current whiteboard systems are not perceived to be useful in engineering design work. All of the interns felt that it required too much time to input drawings, and preferred to draw by hand, then use video or fax. This suggests that any collaborative communication technology systems should provide adequate support for sharing simple hand drawings.

• Engineering design teams need communication support not just with each other, but with the various external constituents of their work. Both groups reported significant communications with supervisors and system administrators. We can interpret this as attempts to secure adequate technical information and resources necessary to accomplish the project.

The study clearly prepared us for the next phase of research - one in which we hope to have a larger number of project teams, more than two members for each participating team, real geographic dispersion across time zones, and a more realistic office communications environment (e.g., actual fax machines).