M. Billinghurst, S. Weghorst, T Furness III

Human Interface Technology Laboratory

University of Washington

Box 352-142

Seattle, WA 98195, USA

{grof, weghorst, tfurness}@hitl.washington.edu

Virtual Reality (VR) appears a natural medium for three-dimensional computer supported collaborative work (CSCW). However the current trend in CSCW is toward the Open Shared Workspace, in which the computer is adapted to work with the user’s traditional tools, rather than separating the user from them as does immersive VR. One solution is through Augmented Reality, the overlaying of virtual objects on the real world. In this paper we describe the Shared Space concept - the application of Augmented Reality for CSCW. This combines the advantages of Virtual Reality and the Open Shared Workspace paradigms. We have developed a collaborative game based on this concept and present preliminary results which show that this approach may be better for some applications.

Keywords: Augmented Reality, Virtual Reality, Computer Supported Collaborative Work

Advances in computing technology have recently led to a paradigm shift in how computers are used and the interfaces for using them. Computers increasingly facilitate cooperative activity between users, leading to the development of unique interfaces to allow Computer Supported Collaborative Work (CSCW). In this paper we describe Shared Space, a new approach for three-dimensional CSCW currently under development at the Human Interface Technology Laboratory. Through the use of Augmented Reality, Shared Space maintains continuity with the users existing workspace while enhancing it by facilitating three-dimensional CSCW.

The Shared Space concept arises from the merging of two different research directions in CSCW; Virtual Reality and the collaborative "Open Shared Workspace". While most CSCW interfaces have been designed to support two dimensional interaction several recent papers have described interfaces for three dimensional CSCW. Wexelblat (1993) points out that three dimensional CSCW is best conducted in an immersive virtual environment; in this setting computers can provide the same type of collaborative information that people have in face-to-face interactions, such as communication by object manipulation and gesture. Work on the DIVE project (Carlsson and Hagsand 1993), GreenSpace (Mandeville et. al, 1996) and other fully immersive multi-participant virtual environments has shown that collaborative work is indeed intuitive in such surroundings. Gesture, voice and graphical information can all be communicated seamlessly between the participants. However most current multi-user VR systems are fully immersive, separating the user from the real world.

Another research direction in CSCW is the concept of the collaborative "Open Shared Workspace" (Ishii and Miyake 1991). The basis of the open shared workspace is that CSCW interfaces must maintain a continuity between existing individual user tools and collaborative tools, and should not block the potential use of current tools and methods. To achieve this, video and computers are integrated seamlessly into the workplace allowing remote participants to collaborate and use familiar tools such as pen and paper to communicate in naturalistic ways. Users are not just restricted to computer-based tools and the computer is integrated into the existing workspace rather than isolating the user from it. Clearly the open workspace paradigm is at odds with immersive virtual environments.

Augmented Reality (AR) allows the overlay of virtual images onto the real world and so combines the advantages of both the Open Shared Workspace and immersive Virtual Reality approaches. Single user Augmented Reality interfaces have been developed for computer aided instruction (Feiner et. al. 1993), manufacturing (Caudell and Mizell 1992) and medical visualization (Bajura et. al. 1992). Shared Space is a unique Augmented Reality approach in which see-through head mounted displays allow multiple users to work in both the real and virtual world simultaneously, facilitating CSCW in a seamless manner. Figure 1.0 shows two users using a Shared Space interface, in this case a collaborative web browser. The users can see each other and virtual web pages floating around them in space, shown in figure 2.0. As can be seen the users can quite easily interact with the real and virtual world simultaneously.

Figure 1.0 The Shared Space Configuration Figure 2.0 The Users View

The ability to use computer supported collaborative tools seamlessly with traditional tools has been shown to be a key factor in determining the success of CSCW interfaces. Grudin (1988) found that if tools force users to change the way they work, then they are generally rejected. The Shared Space approach attempts to avoid this by augmenting the way the user already works, enhancing rather than replacing their current tools and work practices.

Schmalsteig et. al. (1996) identify five features of collaborative AR environments:

• Virtuality: Objects that don’t exist in the real world can be viewed and examined.
• Augmentation: Real objects can be augmented by virtual annotations.
• Cooperation: Multiple users can see each other and cooperate in a natural way.
• Independence: Each user controls their own independent viewpoint.
• Individuality: Displayed data can be different for each viewer.

Several of these features are not unique to AR environments, but Augmented Reality has several advantages over immersive virtual environments, including:

• Local participants can see each other's facial expressions, gestures and body language, increasing the bandwidth of communication between the participants by providing a channel for non-virtual, i.e. natural, communication.
• Participants can use familiar real world tools to manipulate the virtual images, increasing the intuitiveness of the interface.
• Users are able to refer to notes, diagrams, books and other real objects while viewing virtual objects.
• The entire environment doesn’t need to be modeled, considerably reducing the graphics rendering requirements. Augmented Reality can be used to enhance the complexity already in the real world, rather than replace it entirely.

There are several different approaches for facilitating three-dimensional collaborative work. The most obvious is adding collaborative capability to existing screen-based three-dimensional packages. However a two-dimensional interface for three-dimensional collaboration can have severe limitations. For example, Li-Shu (1994) developed a workstation based collaborative CAD package but users found it difficult to visualize the different viewpoints of the collaborators making communication difficult. Communication was also restricted to voice and pointing with a graphical icon, further compounding the problem.

Alternative techniques include using large parabolic stereo projection screens or holographic optical systems to project a three-dimensional virtual image into space. CAVE like systems (Cruz-Neira et. al. 1992) and the responsive workbench (Krug et. al. 1995) allow a number of users to view stereoscopic 3D images by wearing LCD-shutter glasses. These images are projected on multiple large screen projection walls in the case of the CAVE, or a large opaque table top display for the responsive workbench. Unfortunately in both cases the images can be rendered from only a single user’s viewpoint, so only one person will see true stereo. This makes it impossible for users to surround the Responsive Workbench table, or to spread themselves throughout the CAVE and see the correct stereoscopic image. The devices are also difficult to integrate into the users workplace requiring bulky hardware such as a projection screen or large beam splitter, are not portable and require expensive optics.

Mechanical devices can also be used to create volumetric displays. These include scanning lasers onto a rotating helix to created a three-dimensional volumetric display (Soltan, et. al. 1995), or using a rotating phosphor coated plate activated with electron guns (Blundell, et. al. 1994). These devices are also not portable, do not permit remote collaboration, and more seriously do not allow direct interaction with the images because of the rotating display surface.

In introduction we described some immersive multi-user virtual environments that demonstrated how intuitive 3D CSCW was in such settings. There are far fewer examples of multi-user augmented reality systems. Amselen (1995) and Rekimoto (1996) have explored the use of tracked hand held LCD displays in a multi-user environment. Amselen uses LCD panels as portable windows into a shared multi-user immersive environment, while Rekimoto attaches small cameras to LCD panels to allow virtual objects to be composited on video images of the real world. These displays have the advantage that they are small, light weight, portable and higher resolution than head mounted displays. Unfortunately they do not support a true stereoscopic view, and are not hands free. In addition, users must hold the LCD panel in front of their face - obscuring their facial expressions from other participants.

Klaus et. al. (1995) also use video compositing techniques to superimpose virtual image over a real world view. Their system is also multi-user, but is monitor and workstation based so users get the impression that the virtual objects are superimposed on a remote real environment rather than their local environment. In addition it is difficult for users to change the real camera position. Their architecture is designed to support distributed users viewing the same real environment remotely rather than local users interacting in the same real environment.

Our work is most closely related to that of Schmalsteig et. al. (1996). They also use see-through head mounted displays to allow users to view 3D models collaboratively. Their principle application is scientific visualization and they describe a 3D slide show under control of one user. All the participants wear see-through head mounted displays to view the 3D virtual models superimposed on the real world. They report users finding the interface very intuitive and conducive to real world collaboration because the groupware support can be kept simple and mostly left to social protocols.

There are many unresolved issues in developing Shared Space interfaces. Clearly this approach is advantageous when the user needs to use real world objects or information in conjunction with 3D virtual objects. However, one of the most interesting questions is whether the improved communications bandwidth between users affects performance for the same task in a virtual or augmented reality setting. If this is the case then the ease of communication may make the Shared Space approach valuable for 3D CSCW interfaces that don’t rely on other real world objects. In the remainder of this paper we describe a pilot study conducted to address this issue. We begin by describing our experimental design and then present the results, and finally some conclusions and future work.

In this research we were most interested in exploring the differences in task performance between the same collaborative task performed in an immersive virtual environment and augmented reality Shared Space configuration. A performance improvement in the Shared Space configuration could imply that the increase in communication bandwidth does indeed affect task performance.

There are some tasks that can only be performed in a Shared Space configuration, but we were interested in finding a performance difference in a task that could be performed in either augmented reality or an immersive virtual environment. We also wanted a task that forced collaboration between the participants, so we developed a two player game. The game involved moving different colored cubes or balls around a virtual space and placing them in a target configuration. The objects are randomly distributed in space around the users, and a target configuration of six random objects is shown in wireframe in front of them. All objects are space stabilized with respect to the real world.

Each of the players has a different role. One is the "spotter" and can see all the virtual objects. His role is to search the space and find the objects needed to complete the target configuration. Once an object has been found, it can be made visible to the other player by selecting it and saying the command "spot". Object selection is by head pointing. In the absence of 100% accurate voice recognition an experimenter enters the spoken commands on a keyboard. Once an object has been spotted it remains highlighted to show the spotter which objects are visible to the other player. The spotter also has a row of miniature object icons attached to his field of view showing him which objects remain to be spotted. Figure 3.0 shows the view of the spotter. The six objects in the upper right hand corner are icons representing the objects to be spotted, while the wire frame objects at the bottom are four of the six target objects.

The second player is the "picker". Her role is to find the objects that have been made visible by the spotter, select them by looking at them, and pick them up by saying "pick". The object then becomes attached to her viewpoint and she can move it over the target object and drop in place by saying "drop". The picker cannot spot objects and only those objects that have been spotted are visible to the picker.

The role division between the players forces them to collaborate and the spotter needs to tell the picker where the objects are in order for them to be quickly picked. Subjects are free to use any means to communicate. The time to place all six target objects is measured and is used to determine task performance.

While playing the game both players wear Virtual i-O i-glasses! and have their head positions tracked by Polhemus Fastrak magnetic trackers. Both are connected to the same SGI Onyx RE2 computer to minimize network delays, while a second SGI Indy computer provides audio feedback. The players stand side by side and are free to move around the real world. The Virtual i-O glasses can be either used in see-through mode, or covered with a plastic shield for immersive viewing.

An important aspect of Augmented Reality interfaces is accurate registration of virtual objects with the real world. Azuma (1995) describes a very accurate calibration technique for reducing static and dynamic registration errors. In our experiments we used a simplified form of this technique to rapidly calculate initial head orientation. At the beginning of each game users see red crosshairs overlaid on their field of view. They align these crosshairs as accurately as possible with a horizontal and vertical surface twelve feet away and their head position is measured several times. The average of this measurement is used to set the initial tracker orientation offset. Stereopsis also introduces errors into AR interfaces. To get correct close-in stereo the eye offset from the tracker must be measured in a time consuming process. We avoid the need to do this by placing the virtual objects eleven feet away from the user - well outside the distance where stereo perception relies on vergence cues.

There were five experimental conditions tested:

• Real World - Real Body: In this case subjects used the glasses in see-through mode and could see the real world and each other. This corresponds to the ideal Shared Space configuration.
• Real World - No Body: The glasses were used in see-through mode, but a sheet was dropped between the subjects; they could see the real world, but not each other.
• Virtual World - No Body: The glasses are used in immersive mode, but the subjects did not have a virtual representation of each other.
• Virtual World - Virtual Body: The glasses were used in immersive mode and the subjects were each represented by virtual avatars. These block like avatars are similar to those used in other multi-user virtual environment and have a virtual head and tracked virtual hand, as shown in figure 4.0.
• Virtual World - Virtual Body - No Walls: The glasses are used in immersive mode, and the players have a virtual avatar, but the virtual world has only a ground plane.

When a real or virtual body is present participants can use both voice and pointing gestures to show where objects are located. Without bodies they can use voice only. In other pilot studies it was found that in the real world case users were able to use real world objects (computers, shelves etc..) to locate the virtual objects. This is because the virtual objects are space stabilized with respect to the real world. To minimize this effect virtual reference objects were also added to the immersive virtual world, including a ground plane, multicolored wall panels and columns. Players are able to use these virtual reference objects to specify the location of the target objects. In the last condition these walls and columns are removed to test for environmental effects.

We used a counterbalanced experimental design. Each pair of subjects were assigned one of the configurations to start and then cycled through all conditions. There were four target configurations used and subjects experienced each target configuration in each condition. The location of the target objects varied from trial to trial, but subjects all experienced the same set of target locations. Before experimental trials began subjects were give a short explanation of the task and several training games.

A total of 36 subjects (18 pairs) were run through the experiment. There were 20 women and 16 men ranging in age from 19-45. Most had no prior experience of virtual environments and none had played this type of game before. Each played a total of twenty games (four in each of the five conditions) and their times for each game were measured. Several subjects were not able to complete twenty games because of simulator sickness or other factors, however all completed at least two games in each condition. Subjects were also given the short post-experience questionnaire found in appendix A.

The time to complete each target configuration was measured. Performance times were averaged for each subject pair and then across the entire subject pool to give a measure of performance under the different conditions, shown in figure 5.0.

Figure 5.0 Average Subject Performance Times

(RW+RB = Real World and Real Body, RW = Real World,

VE = Virtual Environment, VE + VB = Virtual Environment with Virtual Body, VE+VB+NW = Virtual Environment with ground plane only and Virtual Body)

It appears that users performed better in the real world condition where they could see each other, or in the virtual world case where they had a virtual body and there were no surrounding virtual walls. However these measured performance times had a large variance and using a one factor repeated measures ANOVA we found no significant difference between conditions [F(4,81) = 1.355, p = 0.25].

The lack of a significant difference between performance times may be due to two confounding factors; the use of body cues, and learning effects. Various strategies were used by the players. Approximately half (10 pairs) used techniques that didn’t rely at all on real or virtual body cues, such as designating object location by clock or compass direction. Many of these subjects would use body cues for the real world conditions and other non-body cues for the virtual world conditions. Figure 6.0 shows the performance for those who used any non-body cues at all and those who didn’t. A two factor ANOVA (use of body cues, experiment condition) found a highly significant difference in results between those that used non-body cues and those that didn’t [F(1,4)=6.53, p<0.05]. Subjects often found using non-body cues faster because they didn’t need to look at their partner before searching for the target objects.

There was also a learning effect over time. Figure 7.0 shows the average performance time for the first two and last two trials for each condition for all users. As can be seen users performed better with practice in all conditions. A two factor ANOVA (time, experiment condition) also found a highly significant difference in results between the performance times for the first two trials and the last two [F(1,4)=21.1, p<0.0001].

Figure 7.0 Performance Times for the First and Last Two Trials

When we consider only those subject pairs which didn’t use any non-body cues to aid performance we can use a paired one-tailed t-test to test for difference in performance between the RW+RB case and VE+VB case. Doing this we find a significant difference over the first two trials and over all trials, as shown in the table below.

Table 1.0 Paired T-test results comparing RB+RW to VE+VB performance.

T-values with ostriches are significant at p<0.01, df = 9, t critical = 1.833.

This implies users performed significantly faster in the real world/real body condition than in the virtual world/virtual body case. However there isn’t a significant difference between real world/real body performance compared to virtual world/virtual body performance when there are no virtual walls present, as shown in the table below. Many users found it easier to find the objects when there were no virtual walls distracting them.

Table 2.0 Paired T-test results comparing RB+RW to VE+VB performance.

T-values with asterices are significant at p<0.01, df = 5, t critical = 2.01.

In the post game survey subjects where asked to answer questions on a scale of one to seven for each of the conditions before they knew what their actual performance was. Appendix A lists the questions and anchors for the answer scales.

Subjects thought that the virtual world conditions let them explore new things more than the real world cases. Figure 8.0 shows the average response to the first question: "How much did the search game have surprises and let you explore new things?". Using a one factor repeated measures ANOVA we find a significant difference between conditions [F(4,115)=5.2, p<0.001].

There was also a significant difference in response to the second question "How good was your team at playing the search game?". A one factor repeated measures ANOVA gave the following results, [F(4,114)=7.65, p<0.0001]. On average users felt they we best at playing the game in either the real world/real body case or virtual world/virtual body case as shown in figure 9.0. The perception that they were playing better in the real world real/body case is even more apparent when the results from those subjects that only used body cues are considered.

Figure 8.0 Users Rating of Exploring New Things (1 = not much, 7 = very much).

Figure 10.0 Subject Rankings for Performance in each Condition.

In question seven of the post game survey subjects were asked to rank each condition according to how much they liked it. The most enjoyable condition was ranked first and the least enjoyable last. Figure 11.0 shows that the Virtual World/Virtual Body case was ranked as most enjoyable condition. Using a Friedman two-way ANOVA we again find a significant difference between these rankings [Chi-Square = 24.95, df = (4,22), p< 0.0001]. Subjects enjoyed the novelty of being able to interact which each others virtual avatar, even though many didn’t use the virtual body cues to aid them in their game performance. Subjects also reported that the virtual avatars increased their sense of presence in the virtual environment.

Figure 11.0 Average Ranking of How much Users Liked Each Condition.

In this pilot study we have explored whether or not seeing the real world and collaborators affects task performance compared to an immersive virtual reality setting. We compared the Shared Space configuration against four other configurations. There was no significant performance differences between conditions across all users. However users could be divided into two groups; those that used non-body cues to aid communication and those that didn’t.

There was a significant performance difference among subjects that didn’t use non-body cues; they took less time to complete the task in the real world/real body case than in the equivalent virtual world/virtual body case. This implies that the increased communications bandwidth facilitated by seeing the real world and a real collaborator does indeed aid task performance when body cues are used. However there was no significant difference when users could no longer see the virtual walls in the environment, showing that the environment also had a heavy influence on the task. Several users commented that the task was easier without the virtual walls because they could find the objects quickly against the black background with were fewer environmental distractions.

Regardless of whether they used body cues or not, subjects thought they performed better in the real world/real body case. They ranked this condition highest in terms of how well their team performed and also rated it highly in how good their team was at playing the game. The virtual world/virtual body condition was also highly ranked which is interesting because although users though they performed well, performance on the virtual world/virtual body/no walls case was better. Part of this discrepancy may be due to the fact that users enjoyed this condition so much.

In summary, subjects perceived that they performed better when they can see the real world and their real collaborators and appeared to actually perform better under this condition when using body cues. However subject performance is complicated by the different strategies used, learning effects, and the influence of the virtual environment, so further studies will need to confirm this effect.

It is interesting that seeing the real world and a real collaborator would have a noticeable effect in such a simple task. The usefulness of having users think they are performing well may be enough to justify applying the Shared Space approach in many types of collaborative virtual environments. We may hypothesize that the performance effect will be even more noticeable in tasks that require more complex interactions, for example collaborative design or visualization. For these applications users may rely more on the nuances of conversation that are almost impossible to capture with virtual avatars. In the future we will extend the Shared Space testbed to support usability studies for these types of applications. In particular there are several areas of research that need to be addressed:

• The problem of close-in stereo: Our current configuration only allows users to view objects that appear further than arms length away. To collaboratively view objects closer than that requires more accurate calibration techniques. The eye position relative to the head tracker must be measured exactly and user’s positions must be calibrated with respect to each other. In addition the head mounted displays we are using have a fixed focal length making it difficult to design environments containing objects close to and far away from the users.

• Application Identification: The Shared Space approach will only be appropriate for certain applications and research needs to be conducted to identify the most promising. Our current work is focused on interfaces for wearable computing because these typically use see-through headmounted displays and are used in collaborative settings. The Shared Space approach may not be applicable to wide area distributed VR, but another challenge is to bring non-local users into a Shared Space application with local users.

• Rigorous Usability Studies: Although these results look promising they have originated from a small pilot study and very application specific task. We need to extend this work to test a wider variety of users over a larger application domain. Even in these results we see a difference in performance response over time. We need to determine if Shared Space configurations enhance performance of only novice users, or if they are beneficial for all levels of expertise. We also need to determine the hardware, software and interface requirements for developing effective Shared Space interfaces.

This work is supported by a grant from the Washington Technology Center in collaboration with Virtual i-O Ltd. Hunter Hoffman developed the post experiment survey used, and Edward Miller and Sisinio Baldis provided useful criticism.

Amselen, D. (1995) A Window on Shared Virtual Environments. Presence, Vol. 4(2), pp. 130-145.

Azuma, R., and Bishop, G. (1994) Improving Static and Dynamic Registration in an Optical See-Through HMD. Proceedings of SIGGRAPH ‘94 (Orlando, Fl, July 24-29, 1994), pp. 194-206.

Bajura, M., Fuchs, H., and Ohbuchi, R. (1992) Merging Virtual Objects with the Real World: Seeing Unltrasound Imagery Within the Patient. Proceedings of SIGGRAPH ‘92 (Chicago, IL, July 26-31, 1992), pp. 203-210.

Blundell, B.G., and Schwarz, A.J. (1994) A Graphics Hierarchy for the Visualization of 3D Images by Means of a Volumetric Display System. Proceedings of the IEEE Region 10’s Ninth Annual International Conference, (Singapore, Aug. 22-26, 1994), pp. 1-5. Vol. 1. IEEE New York, NY.

Carlson, C., and Hagsand, O. (1993) DIVE - A Platform for Multi-User Virtual Environments. Computers and Graphics. Nov/Dec 1993, Vol. 17(6), pp. 663-669.

Caudell, T.P., and Mizell, D.W. (1992) Augmented Reality: an application of heads-up display technology to manual manufacturing processes. Proceedings of the Twenty-Fifth Hawaii International Conference on System Science, Kauai, Hawaii, 7-10 Jan, 1992,

Vol. 2, pp. 659-669.

Cruz-Neira, C. et. al. (1992). The CAVE: Audio Visual Experience Automatic Virtual Environment. Communications of the ACM, Vol. 35 (6), pp. 65.

Feiner, S., MacIntyre, B., and Seligmann, D. (1993) Knowledge-Based Augmented Reality. Communications of the ACM, Vol. 36(7), pp. 53-62.

Grudin, J. (1988) Why CSCW Applications Fail: Problems in the Design and Evaluation of Organizational Interfaces. In Proceedings of CSCW ‘88 (Portland, Oregon, 1988), pp. 85-93.

Ishii, H., and Miyake, N. (1991) Toward an Open Shared Workspace: Computer and Video Fusion Approach of Team Workstation. Communications of the ACM, Dec 1991, Vol. 34(12), pp. 37-50.

Klaus, A., Kramer, A., Breen, D., Chevalier, P., Crampton, C., Rose, E., Tuceryan, M., Whitaker, R., Greer, D. (1995) Distributed Augmented Reality for Collaborative Design Applications. Proceedings Eurographics ‘95. pp. C-03-C-14, September 1995.

Kruger, W., Bohn, C., Frohlich, B., Schuth, H., Strauss, W., Wesche, G. (1995) The Responsive Workbench: A Virtual Work Environment. IEEE Computer, Vol. 28(7), pp.42-48.

Li-Shu, Flowers, W. (1994) Teledesign: Groupware User Experiments in Three-Dimensional Computer Aided Design. Collaborative Computing, Vol. 1(1), pp. 1-14.

Mandeville, J., Davidson, J., Campbell, D., Dahl, A., Schwartz, P., and Furness, T. (1996) A Shared Virtual Environment for Architectural Design Review. CVE ‘96 Workshop Proceedings, 19-20th September 1996, Nottingham.

Rekimoto, J. (1996) Transvision: A Hand-held Augmented Reality System for Collaborative Design. In Proceeding of Virtual Systems and Multimedia ‘96 (VSMM ‘96), (Gifu, Japan, 18-20 Sept. 1996).

Schmalsteig, D., Fuhrmann, A., Szalavari, Z., Gervautz, M. (1996) "Studierstube" - An Environment for Collaboration in Augmented Reality. CVE ‘96 Workshop Proceedings, 19-20th September 1996, Nottingham.

Soltan, P., Trias, J., Dahlke, W., Lasher, M., and McDonald, M. (1995) Laser-Based 3D Volumetric Display System: Second Generation. In Interactive Technology and the New Paradigm for Technology, IOP Press, 1995, pp. 349-358.

Wexelblat, A. (1993) The Reality of Cooperation: Virtual Reality and CSCW, in Virtual Reality: Applications and Explorations. Edited by A. Wexelblat. Boston, Academic Publishers, 1993.

The following questions were answered by subjects after completing the experiment.

1. How much did the search game have surprises and let you explore new things?

1 2 3 4 5 6 7

1 = not much 7 = very much

2. How good was your team at playing the search game?

1 2 3 4 5 6 7

1 = not good 7 = very good

3. Do you want to play the search game again sometime?

1 2 3 4 5 6 7

1 = no thanks 7 = very much

4. How well could you communicate with your Partner ?

1 2 3 4 5 6 7

1 = not well 7 = very well

5. How easy was it to collaborate with your Partner ?

1 2 3 4 5 6 7

1 = not easy 7 = very easy

6). For the next question you will be asked to give rankings of the conditions. Afterwards, each condition is listed separately, to give you a chance to write down any thought/comments.

There were five conditions in this experiment:

A. see-through your goggles, and COULD see your partner in the real world.

B. see-through your goggles, but could NOT see your partner in the real world because of the curtain.

C. immersive VR condition (can’t see the real world) where you could NOT see your partner in VR.

D. immersive VR condition (can’t see the real world) where you COULD see your partner in VR (the cube body with a cone hand) with walls and colored columns.

E. immersive VR condition (can’t see the real world) where you COULD see your partner in VR (the cube body with a cone hand) with no walls and no colored columns.

Please rank the conditions in terms of how well your team performed (put the letter of the condition where you did best first, and the letter of the condition you performed worst last and the others in between as appropriate).

7). Please rank the conditions in terms of how much you liked each condition (put the letter of the condition you liked most first, and the letter of the condition you liked least last and the others in between as appropriate).