Smart Devices in Meetings

Smart Devices in Meetings
Florian Weiß

Abstract— Meetings are a valued practice to exchange knowledge, create solutions and work towards common goals. However, problems such as imbalanced participation or undesired behavior lead to unsatisfying results. Current techniques and electronic tools strive to support collaborative work, but introduce new challenges. This work concentrates on novel approaches for meeting support, and divides the topic into two areas: smart devices providing feedback and interactive applications for brainstorming support. First the problem and existing solutions are examined before novel applications are presented. For each category designs, implementations and results of user studies are investigated. Providing ambient feedback about behavior during meetings helped to balance the participation level without causing distraction. Digital multitouch tables facilitate the creation and categorization of ideas as well as face-to-face collaboration. To give insight into the similarities and differences between the systems, a comparison is presented. While all implementations provide valuable benefits, combined approaches fusing the different aspects could be promising, resulting in even more powerful tools.

Index Terms—Meeting Support, Smart Devices, Tabletop Interface, Social Visualization, Collaboration

1 INTRODUCTION

Meetings are an important part of our every day life. They are frequent activities at work, school, university or in the scientific field. Sharing information, discussing options and making decisions are vital for problem solving. Although meetings build a platform to represent individual opinions, create new ideas or achieve common goals, they are often seen as ineffective and inefficient by the attendees. Lack of collaboration or missing coordination could result in decreased motivation and rejection. Shyness due to the fear of judgment and aggressive dominant behavior are undesired and interfere with the process [5]. Nevertheless, teamwork and exchanging knowledge are crucial factors to obtain solutions to complex challenges. Hence the question is: How can this situation be improved?

This topic has been addressed over the past years. Brainstorming techniques [16] have been developed over time to aid idea generation. Behavioral studies evaluated the influence of a facilitator or mediator attending discussions or decision making scenarios. Even electronic systems were designed and developed to test their influence on meet- ings, since these days nearly everyone is familiar with smart phones and personal computers and since the capabilities of these devices are constantly improving. The additional possibilities offer new opportu- nities to build on previous results and enhance existing methods with technology.

In the context of Computer Supported Cooperative Work (CSCW) conferences are held1 to investigate how computer systems can support collaboration and affect groups. The purpose is to offer new designs or ways of interaction or visualization. Particularly smart devices have the potential to affect the user’s behavior in positive ways. They are electronic devices with embedded sensors and transmitters which monitor the environment and communicate with each other in order to share information [22]. This technology can adjust according to a new situation and react in a desired way or give feedback about the current state. Due to the ability of devices to communicate between each other and due to the diversity of input methods (like multitouch and other tangible interfaces), various implementations are possible. Due to the potential combi- nations, different aspects of problems can be addressed and optimized.

In this paper I present a survey and review of smart devices, their influence on meetings and how they support and affect them. I start with an overview of related work and existing problems in the general field of collaboration. Secondly, existing novel approaches with smart devices are presented in two parts. On the one hand smart devices with interactive interfaces and on the other hand smart devices pro- viding feedback about the behavior of the participants are presented. For each I give a general overview and examples of implementations and studies. I conclude with a comparison, discussion and outlook for future work.


2 GROUP DECISION MAKING

Brainstorming has proven to be an effective technique for meetings and is frequently used for collaborative problem solving [16]. It relies on communication and information sharing to optimize the result. Brainstorming consists of two phases: a storming and a norming phase. The first phase is solely used to generate ideas as a base for the second phase. The focus is on quantity and even unusual ideas are appreciated to create a bigger picture. During the norming phase categories and structures are formed in order to combine similar concepts or ideas, thus motivating discussion and reflection. However, new challenges arise reducing the effectiveness of collaborative work.

To be able to design systems which support idea generation and decision making meetings, the potential problems have to be investi- gated. Three key problems were identified by Diehl and Stroebe [5].

• freeriding(socialloafing)
• productionblocking
• evaluationapprehensions

Social loafing, where people tend to put less effort into their work when they are part of a group, can occur from free riding, due to the believe that other team member can additionally cover their part. This can result in the sucker effect, where individuals reduce their level of participation in order to balance the contributions. ”When everyone gets the same grade for the group project, why should I do all the work on my own?” As a consequence, the group is less productive and outcome is more likely to be disappointing. Production blocking occurs when participants cannot release their ideas or comments. It is possible that statements are forgotten or suppressed during the process, since they are no longer relevant for the current state of the discussion. The fear of judgment or negative feedback is evaluation apprehension. Members withhold their ideas, restrain themselves and lower their participation.

Based on these facts a number of customized brainstorming tech- niques have been developed. Nominal brainstorming, where each par- ticipant is isolated during the storming phase to reduce evaluation ap- prehension, or a group passing technique, allowing members to build on the others’ ideas to decrease production blocking, are examples for such techniques. To aid and support the categorization process during the norming phase, Post-its and whiteboards can be used.

With the help of technology existing approaches can further be improved. Electronic Brainstorming Systems (EBS) are designed to enhance the meeting process and outcome. Using computer networks enables distributed conferences [8] or tracking and storing the process for later review. Participants can work in parallel to prevent production blocking. Individual contributions can be visualized on a shared interface to reduce free riding and support collaboration. A certain kind of anonymity prevents evaluation apprehension.

Although EBS address the previously stated problems, working on a personal computer reduces face-to-face interaction, which has shown to be an essential factor [7]. Facial expressions and body language are important parts of social interaction and have a big influence on our behavior. The way the participants act during meetings can improve collaboration or shut it down. Therefore facilitators are present during discussions to mediate the sessions. Providing feedback on behavior and group dynamics have been proven to lead to higher satisfaction and consequently a better overall performance [19]. As a result the participants reflect about their current actions and can adjust accord- ing to the situation. Since personal computer or smart phones are con- sidered disruptive, new ways to support face-to-face collaboration are investigated.

Smart devices offer great potential to combine the advantages of EBS and to enable face-to-face interaction for meetings. Embedded sensors, networking and communication between devices allow reac- tion according to measured information, whereas multitouch displays provide access to a wide range of interaction possibilities. Since there are multiple factors that influence the outcome, different approaches are followed to overcome the problems and improve the general re- sult. In the next part I introduce some of the novel approaches which support brainstorming and decision making meetings with smart de- vices.


3 INTERACTIVE INTERFACES TO SUPPORT BRAINSTORMING

To enhance existing brainstorming techniques interactive smart de- vices can be utilized. The goal of the later described systems was to combine the beneficial aspects of conventional brainstorming with the support that smart devices could provide, to overcome the gen- eral problems. Taking the already proven methods into consideration, similar approaches have been derived, adapted and implemented.

The research of Buisine et. al [2] showed that the collaborative behavior can be influenced with multitouch displays, but that other new difficulties arise. If the user interface is too complicated and overloaded with information the participants lose focus of the creative process. To avoid negative impact on the creative process of brain- storming, the research of special requirements is an important topic, as described in several publications [9, 18].

Depending on the chosen devices and implementation, different focus areas are addressed and investigated. Some studies concentrated on the enhancement of idea generation with smart devices and its features [2, 3, 11, 17], others examined user satisfaction or the impact of external influences like time pressure on meeting behavior [17].

In this chapter three different approaches, all supporting brain- storming meetings in a multitouch table environment, are presented. The systems use different input devices to generate and store the ideas. Each system implements a set of gestures to manipulate the individual representation of the participant’s thoughts and organize them on the table. The systems are actively used by the participants to improve the brainstorming sessions.

3.1 Brainstorming on Collaborative Multitouch Table

With Firestorm [3] an interactive multitouch tabletop system to support brainstorming was designed and implemented. A set of design rules optimizing the user interface were established and followed after a preliminary study. During the main study the authors compared a conventional brainstorming technique with the refined version of Firestorm.

For the best support of brainstorming the storming phase and the norming phase were optimized separately. To improve the number of ideas generated during the storming phase, a wireless keyboard was used by each participant. Every user could write down their ideas, which were visualized on a one line note, similar to Post-its or cards. The notes were displayed after pressing the enter button and arranged spiral-shaped in the middle of the tabletop, readable for every participant. Colors of the cards referred to the author of the idea. After the idea creation during the storming phase, the norming phase to categorize and structure ideas started. Firestorm offered several options to support this brainstorming phase. Notes could be rearranged or discarded using gestures on the multitouch table. Through flipping an idea card it was transformed into a container with the ability to store other ideas (see figure 1). This way hierarchical structures could be created. Dragging single notes or drawing a lasso around multiple cards could aggregate them into a container. Due to the implemented archive function, the complete idea generation and categorization process could be stored and later analyzed in detail.

For the main study Firestorm was compared to a whiteboard brainstorming session. Groups of four participants processed similar problems under each condition. During the whiteboard setup the outcome of the storming phase was printed and handed out to every user. Thereafter the board was cleaned and used for the norming phase, to reach a fair comparison between the two conditions. For the evaluation the number of ideas generated during the first phase and the categorization during the second phase was observed.

The user study was analyzed regarding the previously considered design goal to support the collaboration of the participants. The first phase was mainly supported by wireless keyboard input. With the aid of a wireless keyboard users could create ideas individually and parallel to other participants. Since every participant was familiar with keyboards and its usage they could easily write down their ideas without having to learn the usage of a new interface or being distracted by the multitouch table. Indicating the author of each note, the color coding helped the users to get an overview of their contribution which served as a motivation. Table interaction was applied during the second phase. The possibility to flip cards to create containers and categorize ideas was very well received. Collecting related notes with the lasso function to build subcategories or topics, supported the norming phase and was frequently used. Through the hierarchical structure it was simple to organize and group the ideas generated during the first phase. The archive function offered additional opportunities to analyze the process of the brainstorming session. Furthermore the authors suggested to implement a new third phase, where the participants could review and recall the states during the meeting - potentially to create additional ideas based on them.

Summarizing the outcome of the main study, Firestorm is designed and optimized to support both phases of the brainstorming process individually for a better group collaboration.

3.2 Shared Table and Wall Interface for Collaboration

In the paper ”Designing for Collaborated Creative Problem Solv- ing” [9] the authors created a system to support brainstorming and decision making. Based on an analysis of influencing factors they de- cided to combine a table and a wall interface to improve collaboration. To reference this system, the term Digital Post-its is used in this paper.

The general idea was to give the participants a tool to create and arrange Post-its. As input device a pen was used. To create a new idea the participant drew a box on the table and an empty Post-it appeared. For easy editing the note stayed large and faced to the creator. Completed Post-its were reduced in size, while still readable, and could be rearranged on the table. Additionally the Post-it became visible on the wall display. Gestures were implemented to edit existing notes, to change their position or to delete them. One feature of the implementation in contrast to the ”real world” Post-its was that the participants were able to skid their ideas across the table to share them and get direct feedback. Another important factor was that parallel working of all users was enabled, since everybody had their own working space on the table. For a better overview idea grouping and clustering was supported by the wall display. Participants could organize and reorder the Post-its, draw circles around them to build groups and connect them with one another.

The evaluation of the design was tested during two brainstorming sessions. Groups consisted of two people and had to create ideas dur- ing the sessions with and without the Digital Post-its system. For the task without digital support, the original paper-based brainstorming technique was applied. To investigate the usability and effectiveness, the number of generated and exchanged ideas were compared between the different setups. The table solution offered the participants the opportunity to sit facing towards each other. The wall display was located next to the table (see figure 2), where they could group ideas.

In general the participants were not distracted by the Post-its system and preferred it over the paper-based technique. Since the user could rely on knowledge they already had about Post-its, the interface and usage was easy to learn. The additional possibility to skid the idea to the other participant was very popular. While supporting parallel input, due to the individual working areas, a fluid interaction without blocking was enabled. Although the amount of ideas generated did not increase significantly, other benefits of the system were valued. The table arrangement offered the opportunity for face-to-face communication with all its advantages like body language and facial expressions. The structured presentation of the created ideas on the wall display allowed the participants to rearrange the Post-its easily but also to review the already reached results. In some cases where the brainstorming process stagnated, the overview on the wall display initiated a new creative phase based on previous notations.

Altogether this paper showed that the combination of a wall dis- play and an interactive table can support group collaboration and profit from the individual benefits.

3.3 Voice Recognition and Semantic Knowledge Database

In their work ”WordPlay: A Table-Top Interface for Collaborative Brainstorming and Decision Making” [10] the authors implemented a multitouch table system (referenced as WordPlay). As primary input device they used a microphone activated by a press button, which was attached to the table.

The system offered two ways for creating ideas. With the mi- crophone spoken ideas were recorded and translated into text with a word recognition tool and displayed on the table (see figure 3). To correct false speech detection related to a small error rate of the tool but also to non-English accents, a multitouch keyboard was available on demand. Alternatively, participants could touch existing ideas on the table to get a list of associated statements from a linked semantic knowledge database. A collection of additional ideas grew out of the original and could be kept or discarded by the user. After a certain period of time unused associations faded away in order to reduce complexity. Ideas were displayed on the table as short phrases or sentences and could be manipulated on the multitouch interface. Resizing, rotation and repositioning by the users were fully supported. For individual decision making or brainstorming settings, special backgrounds were provided, for instance to sort and group pro and con arguments or to rate ideas.

The system was tested and evaluated in different application areas. Due to the input of ideas with a microphone, it was important to reduce background noise. The experiments distinguished between two types of meetings. One scenario was to decide between a set of possible choices, the other was a brainstorming session to generate ideas. Application areas were: lab demos, museum events and the groups internal usage.

The focus of the evaluation was to investigative the usability of the system and not the comparison with other brainstorming techniques. In conclusion the authors stated that WordPlay was easy to use for the participants. Voice recognition enabled comfortable recording and visualization of ideas or phrases on the table. Associations retrieved from the linked semantic knowledge database improved the idea generation and led to new directions to think of. Hiding untouched ideas after a short time period helped the participants to fade out undesired and unused phrases. To rate and categorize ideas different backgrounds were utilized for keeping the overview of the complete picture. While multiple users could control the interface simultaneously the study revieled that resizing the global view was difficult. This could result in confusion for participants who did not directly manipulate the table. Therefore, instead of making it a global action, scaling should be implemented as an individual control for each participant.

WordPlay presented a system to include the computer as an addi- tional participant via a linked semantic knowledge database for asso- ciations and additional idea generation. Offering the microphone and voice recognition, WordPlay simplified the idea generation and log- ging process.


4 INFLUENCE OF VISUAL FEEDBACK ON SOCIAL BEHAVIOR

A different approach to influence the meeting scenario is to pro- vide feedback during the sessions. Whereas the previous chapter focused on devices to be actively used by the participants, in this section I compare various forms of displaying feedback about social behavior to the user. Several studies already proved that providing information about their performance during discussions had an impact on the user’s behavior and could be used to influence social dynamics ([1, 6, 14, 20, 21]).

In this chapter I provide an overview about the ideas, implementa- tions, studies and results of four selected papers. Each has a special focus and investigates different aspects and influence on the meeting process and outcome. To implement the individual systems, different factors were taken into account which are presented in this chapter. The way the information is presented, the location and type of the dis- play or what aspects of the meetings are visualized were considered. The common intention of all the methods followed in these publica- tions is to improve the information exchange and to balance the par- ticipation level.

4.1 Influencing Group Participation with a Shared Display

DiMiccio et al. evaluated the influence of a shared display on group decision making, regarding critical information involved [6]. The idea was to display the individual speaking time and turns, in order to balance participation and improve information sharing without loosing the critical information. The system is referenced as Shared Display in this paper.

For their study they used two wall displays to give feedback to ev- ery participant about their individual speaking time. Each individual’s speech was recorded with a microphone and processed to be displayed in real time. To make a fair judgment on the participation of each member only the essential information was captured and used for calculation. In the meeting room two displays were placed sideways to the participants, so that they were not disturbed in their discussions. The displays showed histograms with the individual participation level compared to each other (see figure 4). The categories were: over-, under- and normal participation. Additionally a row of circles above the histogram shows the chronology and duration of the speakers.

The general setup of the user study was a face-to-face meeting between four people. Overall there were two groups each discussing problems in two rounds. One group had to complete the tasks without the aid of the shared display. The other group also used no display in the first discussion round and were supported by the display during their second decision making session. Every group was provided with general information about the topic. Furthermore each participant got extra information, which was not disclosed to the other members. The goal was to motivate knowledge exchange and communication. Critical information, which was crucial for determining the right decision, was only provided to one participant per group. This fact was not explained to the groups, but the individuals recognized the importance of this fact during the discussions.

Evaluating the results of the user study, the authors compared their findings between the different groups and the two sessions. The study showed that the displays did not disturb the decision making process. Since they were not located in the direct view of the participants, the information was not the center of the attention, but could be recognized on demand. With the feedback of the shared displays the second group did change their discussion behavior during the second task. This was not observed compared to the first group without display. The study indicated that participants, who over-participated during the first task, regulated their speaking time due to the feedback. This only applied for subjects without critical information. On the other hand, members with critical information knew their importance for the decision making and did not reduce their participation share. This means that the quality of the decision did not suffer from the usage of this setup. Participants with a low participation level were not encouraged, since they found the feedback not accurate compared to their own impression, resulting them to ignore displayed information.

Overall this paper concluded that while over-participation was reg- ulated to a better level, the crucial information holders did not restrict their participation in order to ensure that these facts are shared with the other members.

4.2 Metaphorical Group Mirror on Table or Wall

Groupgarden [21] is an implementation of ambient feedback via metaphors. Instead of displaying bar graphs the idea was to give the user feedback via images and to motivate the group to collaborate in a playful way [4]. Additionally the paper examined the influence of the location of the display.

The original Groupgarden used a wall display, but for the location study the concept was additionally adapted to a table display. For comparison of the tasks, the number of ideas generated during brainstorming meetings were counted and presented. To generate the feedback and to control the application, an operator was necessary (Wizard of Oz [13]). As an individual visualization of each user, flowers with petals were displayed. The number of ideas were represented by the number of petals and the size of the flower. As group feedback a tree was implemented, which grew according to the balance of the participants and the overall number of ideas. An unbalanced distribution of ideas resulted in a small leafless tree, whereas even participation created a prosperous tree. In addition bad weather warned the users about bad behavior and going off topic (see figure 5).

The user study of the brainstorming sessions was divided into two parts. At first groups were selected to complete two brainstorming tasks, each without and with Groupgarden displayed on the wall. Groups consisted of three participants and could choose how to arrange their chairs during the meeting. They had to chose if they preferred to be turned to the display or to sit face-to-face to each other. Based on the fact that in brainstorming meetings face-to-face is an important factor, they conducted a second study to investigate the influence of the display location. So the focus of the second study was a comparison between the original wall display and a table display adaptation. The table setup was a face-to-face meeting around a table with the display, where every participant had their flower in front of him/her and the tree in the middle.

With the use of Groupgarden the number of ideas generated by each participants were balanced, in contrast to the groups with no feedback. Over-, as well as under-participation were adjusted. The playful character of the implementation motivated the users to improve their individual as well as the group result. Even participants with fewer ideas in the task without Groupgarden tried to enhance their contribution in order to achieve a better outcome, rather than giving up. Since the tree reflected the balance of the group, it drew attention of users with a higher amount of ideas to look for smaller flowers, encourage them to participate and self-regulate their dominance. The second study unveiled, that the location of the display had no direct influence on the amount of ideas generated or their distribution. In both cases (wall and table) the participation was well balanced and supported by Groupgarden. Some participants stated that the table solution was better for maintaining eye contact and that it supported collaboration. Other users chose the wall display over the table since it was less distracting and served as ambient feedback. Depending on personal preferences either the wall or the table solution were favored.

Groupgarden regulated the group participation in a positive way, because over- and under-participation was well adjusted via feedback through metaphors. Investigating the location of the display, the find- ings were dependent on personal preferences only, but had no influ- ence on the overall improvement.

4.3 Visualizing Level of Agreement during Discussions

Based on the research and suggestion of Karahalios [12] the influence of the level of agreement should be investigated. The original work consisted of two types of Social Mirrors: The Conventional Clock to display speaking turns and length and the Conventional Vote to visualize the level of agreement. The paper ”The Effects Visual Feedback on Social Behavior during Decision Making Meeting” [1] describes an implementation of a new Social Mirror, combining these aspects.

The implementation consisted of a personal computer for each participant, a shared table display and a wall display. Each user had to start by creating their own avatar they could refer to. The mapping of the avatars to their owners was displayed on the table display. The wall display and the personal computers were used for visualization of behavior and level of agreement. The size of each avatar circle represented the individual speaking time and participation level calculated through automatic speaker detection. The interactivity level between two participants was indicated with lines between the avatars. Proximity of avatars signaled a good level of agreement, while distance stood for disagreement (see figure 6). With the user interface on the PC the participants controlled the position of their avatars to show their opinions. In this paper, I use Social Mirror to refer to the latter described system.

To investigate the Social Mirror system, a study with two groups in total was done with the focus on providing active feedback about individual opinions and agreement. Each group had four members, one external facilitator and an additional operator providing a mind map of the discussion. Roles were predefined for every participant to state four positions. Through discussion of a given problem, a collaborative decision had to be made. The system was briefly introduced, but only regarding the possibility to show which concept they prefer and which opinion they share. It was not explained how the size of the avatar or the lines between them should be interpreted and influenced. Similar to the other studies one task was held without feedback and a second task with the aid of the Social Mirror. Based on different background designs the users could vote for their preferred decision or display their satisfaction with the outcome of the discussion.

Although the users were not informed about the visualization of their speaking length and talking turns, they had the impression, that the use of the Social Mirror led to a change in their social behavior. Because of the awareness of the level of agreement from other participants, they could individually adjust their actions. The non verbal feedback that the participants could give to the group by moving the avatar closer or further away from other user was helpful to recognize the opinions about the statements. This also gave insights about the preferences of the participants for the facilitator and pointed out extreme positions. Additionally each member was constantly engaged in forming an opinion, share their thoughts and work towards decisions.

In contrast to previously shown studies the task without Social Mirror had a better balance in the participation level. The questioning after the tasks showed, that the participants could realize the corre- lation between avatar size and speaking time although previously not explained. Since the focus of attention was on displaying and controlling the level of agreement via their avatar, and since the visualization of the other factors were not explicitly mentioned before, the participants only concentrated on the given suggestions of usage. Therefore the outcome of the study differs from the other experiments.

The evaluation of Social Mirror showed that giving feedback about the level of agreement can be a useful feature to influence the social behavior in decision making meetings and support no verbal expres- sion of agreement and disagreement.

4.4 Enhancing Collaboration using Sociometric Feedback

An approach for portable meeting support is Meeting Mediator (abbreviated with MM) [14]. MM uses sociometric signals to provide real time feedback on mobile phones for co-located and distributed meeting scenarios.

For the studies of Kim et al. a combination of two smart devices were used. The visual feedback to each participant was displayed on a personal smart phone. Every user was represented by a square in one corner of the display. A circle in the center was connected to the user square with a line and changed its color from white to green according to the degree of good collaboration. The thickness of the lines represented the speaking time of the individual participants and pulled the circle towards their corner indicating over-participation (see figure 7). To control the visual feedback sociometric badges (worn around the neck) were used which transferred the information via Blue tooth to the smart phone. The sociometric badges collected not only speaking time, but also other speech characteristics like speaking speed or tone of voice. Additionally, body movement, proximity to other badges or face-to-face interaction could be detected and used for evaluation. Sociometric badges are described in detail in [15].

For the user study, major points of interest other than improving the interactivity, were the influence of dominant participants on the group dynamics and the reduction of differences between co-located and distributed meetings. To test the effects of MM, groups of four people were built. Groups which had the feedback of MM and groups without feedback had to solve two tasks. Each in a co-located setup and in a distributed meeting scenario where the group members were separated. The tasks were derived from the ”Twenty Question” game, where one player is the answerer who chooses a secret to guess and the other players have 20 questions with ”yes”- or ”no”- answers to find the solution. During the first phase ten questions including answers were presented to the group. Through brainstorming and collaboration the members had to generate appropriate ideas for the given set of questions and corresponding answers. During the second phase (problem solving) the additional ten questions could be asked to an external ”answerer” who had the solution to the quiz. Indicators for a good collaboration was the number of questions needed to obtain the correct conclusion.

The results of the evaluation state that MM had a positive effect on the group participation. The difference between dominant and non dominant participants could be reduced. The color changing circle helped to balance the change of speaking turns. Furthermore MM was able to identify dominant people and with this detection it could be shown that dominant participants had an influence on the group dynamic. For groups with one or even multiple dominant speakers, MM could make distributed meetings more like co-located collaboration. However MM could not reduce the difference between the two task settings for groups with only non dominant participants. This emphasized the important factor which dominant people play in group collaboration.

In summary the study of MM showed, that it could detect social in- teraction between co-located and distributed meetings and give feed- back to influence and improve group collaboration especially in sce- narios where dominant speaker are participating.


5 OVERVIEW

In the previous chapter I presented the chosen applications in detail in- cluding the overall system and their individual user studies. Although the common goal was to design applications to improve idea genera- tion and collaboration during meetings, several different approaches were chosen to address problems in these scenarios. For a better overview the general setup of the systems are summarized and com- pared within their category.

5.1 Interactive Multitouch Systems

Basically the main differences between the three interactive multi- touch table systems were the input devices, the way of interaction and the support of the brainstorming process (see table 1). Building on common knowledge, devices like pen or keyboard were chosen or familiar gestures from smart phone usage were implemented. Only a few new actions had to be learned which simplified the operation of the system and avoided distraction from the task. The plain surface of the table interface supported face-to-face interaction to improve collaboration and awareness.

To reduce social loafing Firestorm decided to use colored note cards to indicate the author of the ideas. Through this participants were al- ways aware of their contribution. This fact was not directly addressed nor needed in the Digital Post-its setup because of the limitation to only two users.

For the categorization, Firestorm decided to use container and hier- archy structures on the table, whereas the table of Digital Post-its was only used as shared working space. The grouping could be performed on the wall display to get a better overview of the intermediate results. Wordplay used different backgrounds with patterns aiding the users.

Parallel input was supported by individual input devices or working spaces to prevent production blocking. Every action could be per- formed independently from the other participants, but the result was displayed in real-time for the entire group. Solely WordPlay had prob- lems with their implemented ”zoom” function, since it affected the whole viewport and disturbed the others.

Although WordPlay did not compare their performance with tradi- tional brainstorming techniques it was selected for this paper because of the fact that it combines speech recognition and a semantic knowl- edge database to enhance idea generation.

5.2 Ambient Feedback Systems

In contrast to the previously mentioned applications, the following systems focus on the presentation of feedback to the participants of the meetings in order to influence group dynamics. The selection of publications was chosen so that a broad range of different influencing factors, their respective visualization and their impact on the sessions could be presented. Although the common goal was the same, each system had it’s individual focus and design. An overview of the different visualization indicators and system setups is summarized in table 2.

To verify the effects on brainstorming and decision making, individ- ual user studies were developed. The Groupgarden study compared a conventional brainstorming session with their system. Shared Display gave each participant identical information as well as additional facts. One of the members got special knowledge about the given problem, to swing the decision in the right direction. For the Social Mirror study, each user had their own position and had to decide between four stated concepts. Meeting Mediator divided the ”Twenty Question” game into an idea generation and decision making phase. This shows that every application had a special focus area of investigation.

To provide feedback, information of various types was presented. For visualization Shared Display selected bar graphs to display the par- ticipation level of each user and animated circles for speaking turns. This simple design allowed easy comparison. Participants suffering from under-participation ignored the feedback and could not be moti- vated. Important to note is that the holders of critical information did not reduce their participation level because of the relevance of their input.

Representing the number of ideas as growing flowers (metaphors), Groupgarden showed its positive impact on the group dynamics. Under- and over-participation could be regulated. Feedback on group performance and the chance to achieve a greater tree motivated team members to collaborate in order to reach a better result. The investi- gation of the location of the display showed that the placement, either on the wall or on the table, did not directly influence the outcome.

The focus of the Social Mirror implementation was to display the level of agreement. Participants constantly formed their opinion and could not only indicate their position but also react to the feedback of others. With the aid of Sociometric Badges Meeting Mediator was able to take additional information into account. In addition to the talking turns and duration, speech energy and movement could be de- tected. This enabled the device to identify dominant participants and their influence on the group dynamics.

Only MM offered the possibility to be utilized in co-located as well as distributed meetings and compared both scenarios. The MM study concluded that it reduces the differences between these two types, due to the consideration and visualization of social signals. Despite the differences, all systems showed a positive impact on the behavior of the individuals, the participation level and the general collaboration.


6 DISCUSSION

All the previously presented papers showed that smart devices have an influence on meetings. Diverse implementations and design choices addressed various problem areas of collaboration and were able to increase the users’ satisfaction compared to conventional methods. Although not every solution amplified the outcome in terms of the number of ideas generated or chosen solution, the electronic support added individual benefits.

Based on the observation of the described systems general design principals can be derived which should be considered for future works. Firstly face-to-face interaction should be enabled during brain- storming and meetings since facial expressions and body language have a big influence on our behavior [7]. Therefore table displays provide the solution to provide ambient feedback (see chapter 4.2), as well as allowing multiple users to switch between collaboration and individual idea generation. Secondly the system must not distract from the main task. Information and interaction should be available but not demand dedicated concentration or time from the user to understand the message or operate the device. Thirdly, systems should be easy to learn and understand. This applies to the operation of the input devices as well as the information visualization. In the presented papers, well-known brainstorming techniques were adapted to simplify the handling of the systems. Feedback should not be overloaded with information and a simple and intuitive design supports quick interpretation by the participants. These principals were taken into account by all the presented systems but realized in different ways and to a certain extend.

A direct comparison of the presented applications is difficult, because each study focused on their particular design goals. Hence they compared their systems against the traditional brainstorming scenario to identify their advantages and improvements. To under- stand the influence of the individual design choices a study should be carried out were the implementations are compared against each other. A general setup with equal starting conditions and tasks should be selected to achieve comparable results. An adaptation of the ”Twenty Questions” game from [14] could be suitable for interactive brainstorming support and ambient feedback applications, since it combines an idea generation and decision making phase. This might help to identify the impact of chosen input devices or visualization types on the meeting process and outcome.

Based on the results of such a study a combination of interactive and feedback systems could be derived, which might offer even more benefits for meetings and collaboration. Devices providing feedback as well as interactive solutions for brainstorming support need dis- plays for visualization and interaction. During the storming phase of a brainstorming task multitouch table offer face-to-face interaction and enable parallel input to reduce production blocking. In the norming phase categorization and discussion could be supported with ambient feedback. The information can be used to reflect on behavior, balance participation or to display level of agreement. Digital Post-its already combine personal working spaces on a multitouch table and a wall display to provide an overview of the generated ideas. It could be a possibility to include feedback on the display about the progress and participation level of the group with flowers and trees or to make use of Sociometric Badges to enable distributed meeting scenarios. Even novel technologies like Google Glass or Apple Watch might offer new opportunities for visualization or input methods.


7 CONCLUSION

In conclusion this paper provides an overview of existing approaches supporting brainstorming or decision making during meetings. Inter- active systems improve the idea generation process, whereas smart de- vices with ambient feedback can influence the behavior of groups. A big advantage of smart devices is their capability to gather information about the environment with sensors and transfer them to displays for visualization. Existing methods and techniques are enhanced with this technology eliminating the problems described in chapter 2.

Building on already established concepts for meeting support smart devices can add additional benefits, increasing the contribution and satisfaction of the participants. Multitouch table systems allow face- to-face interaction which improves collaboration and support creation and categorization of ideas during brainstorming tasks. Generally am- bient feedback provided by smart devices result in a more balanced participation level and has a positive effect on the social behavior, without distracting the participants.

Despite the different intentions of the presented studies common approaches have been identified and summarized in chapter 5. Some of the previously stated design choices could be transferred into other systems to compare the resulting application with their original solution.

Given the proven impacts of the presented concepts, future studies and implementations should be based on the design ideas and findings to combine their benefits and further improve the process and outcome of meetings.


REFERENCES

[1] M. Brandon, S. Epskamp, T. de Groot, T. Franssen, B. van Gennep, and T. Visser. The effects visual feedback on social behavior during decision making meetings. In Proceedings of the 1st International Conference on Human Interface and the Management of Information: Interacting with Information - Volume Part II, HCII’11, pages 219–228, Berlin, Heidel- berg, 2011. Springer-Verlag.
[2] S. Buisine, G. Besacier, A. Aoussat, and F. Vernier. How do interac- tive tabletop systems influence collaboration? Comput. Hum. Behav., 28(1):49–59, Jan. 2012.
[3] A. Clayphan, A. Collins, C. Ackad, B. Kummerfeld, and J. Kay. Firestorm: A brainstorming application for collaborative group work at tabletops. In Proceedings of the ACM International Conference on Interactive Tabletops and Surfaces, ITS ’11, pages 162–171, New York, NY,
USA, 2011. ACM.
[4] S. Deterding. Gamification: Designing for motivation. interactions,
19(4):14–17, July 2012.
[5] M.DiehlandW.Stroebe.Productivitylossinbrainstorminggroups:To-
ward the solution of a riddle. Journal of personality and social psychol-
ogy, 53(3):497, 1987.
[6] J.M.DiMicco,A.Pandolfo,andW.Bender.Influencinggroupparticipa-
tion with a shared display. In Proceedings of the 2004 ACM Conference on Computer Supported Cooperative Work, CSCW ’04, pages 614–623, New York, NY, USA, 2004. ACM.
[7] P. Dourish and V. Bellotti. Awareness and coordination in shared workspaces. In Proceedings of the 1992 ACM conference on Computer- supported cooperative work, pages 107–114. ACM, 1992.
[8] R.Gumienny,L.Gericke,M.Wenzel,andC.Meinel.Supportingcreative collaboration in globally distributed companies. In Proceedings of the 2013 Conference on Computer Supported Cooperative Work, CSCW ’13, pages 995–1007, New York, NY, USA, 2013. ACM.
[9] O. Hilliges, L. Terrenghi, S. Boring, D. Kim, H. Richter, and A. Butz. Designing for collaborative creative problem solving. In Proceedings of the 6th ACM SIGCHI Conference on Creativity &Amp; Cognition, C&C ’07, pages 137–146, New York, NY, USA, 2007. ACM.
[10] S.HunterandP.Maes.Wordplay:Atable-topinterfaceforcollaborative brainstorming and decision making. Proceedings of IEEE Tabletops and Interactive Surfaces, pages 2–5, 2008.
[11] S. Hunter, P. Maes, S. Scott, and H. Kaufman. Memtable: An integrated system for capture and recall of shared histories in group workspaces. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’11, pages 3305–3314, New York, NY, USA, 2011. ACM.
[12] K. Karahalios. Social mirrors as social signals: Transforming audio into graphics. IEEE Comput. Graph. Appl., 29(5):22–32, Sept. 2009.
[13] J. F. Kelley. An empirical methodology for writing user-friendly natural language computer applications. In Proceedings of the SIGCHI Confer- ence on Human Factors in Computing Systems, CHI ’83, pages 193–196, New York, NY, USA, 1983. ACM.
[14] T. Kim, A. Chang, L. Holland, and A. S. Pentland. Meeting mediator: Enhancing group collaboration using sociometric feedback. In Proceed- ings of the 2008 ACM Conference on Computer Supported Cooperative Work, CSCW ’08, pages 457–466, New York, NY, USA, 2008. ACM.
[15] D. O. Olgu ́ın, B. N. Waber, T. Kim, A. Mohan, K. Ara, and A. Pent- land. Sensible organizations: Technology and methodology for automat- ically measuring organizational behavior. Trans. Sys. Man Cyber. Part B, 39(1):43–55, Feb. 2009.
[16] A. Osborn. Applied imagination; principles and procedures of creative problem-solving. Charles Scribner’s Sons, New York, NY, USA, 1963.
[17] L. Schmitt, S. Buisine, J. Chaboissier, A. Aoussat, and F. Vernier. Dy- namic tabletop interfaces for increasing creativity. Comput. Hum. Behav., 28(5):1892–1901, Sept. 2012.
[18] S. D. Scott, K. D. Grant, and R. L. Mandryk. System guidelines for co- located, collaborative work on a tabletop display. In Proceedings of the Eighth Conference on European Conference on Computer Supported Co- operative Work, ECSCW’03, pages 159–178, Norwell, MA, USA, 2003. Kluwer Academic Publishers.
[19] E. E. Smith and S. S. Kight. Effects of feedback on insight and problem solving efficiency in training groups. Journal of Applied Psychology, 43(3):209, 1959.
[20] J. Sturm, O. H.-v. Herwijnen, A. Eyck, and J. Terken. Influencing social dynamics in meetings through a peripheral display. In Proceedings of the 9th International Conference on Multimodal Interfaces, ICMI ’07, pages 263–270, New York, NY, USA, 2007. ACM.
[21] S. Tausch, D. Hausen, I. Kosan, A. Raltchev, and H. Hussmann. Group- garden: Supporting brainstorming through a metaphorical group mirror on table or wall. In Proceedings of the 8th Nordic Conference on Human- Computer Interaction: Fun, Fast, Foundational, NordiCHI ’14, pages 541–550, New York, NY, USA, 2014. ACM.
[22] C. A. Valhouli. The internet of things: Networked objects and smart devices. The Hammersmith Group Research Report, 20, 2010.