New conference attendees often lack existing social networks and thus face di culties in identifying relevant collaborators or in making appropriate connections. As a consequence they often feel disconnected from the research community and do not derive the desired bene ts from the conferences that they attend. In this paper we discuss Snag'em, a social network game designed to support new conference attendees in forming social connections and in developing an appropriate research network. Snag'em has been used at seven professional conferences and in four student settings and is the subject of active research and development. The developers have sought to make the system engaging and competitive while preventing players from `gaming' it and thus accruing points while neglecting to form real-world connections. We brie y describe the system itself, discuss its impact on users, and describe our ongoing work on the identi cation of critical hub players and important social networks.
Social networking is an essential task at any academic
conference or professional venue. One of the primary goals of
attendees is to seek out relevant work, identify potential
collaborators, and to maintain existing connections. Many
of these contacts are made by building upon existing
relationships and by expanding the attendees existing social
network. New conference goers however, particularly
students and historically underrepresented groups, lack these
foundational networks and thus face di culties making
connections. Based on Tinto's Theory of University Departure,
increased interaction with other students, faculty, sta and
community supporters can increase the retention rate of
minority populations and sense of community within secondary
and post-secondary academic communities [
In academia, sense of community has a strong positive
correlation with retention [
Snag'em was designed as a pervasive game to encourage
valuable professional networking and promote sense of
community. The system's pervasive features are designed to
help players translate their in-game networks directly into
real world peer groups. The system was originally created
for the 2009 Students and Technology Academia Research &
Service (STARS) conference. This conference is unusual in
that it is an academic conference designed speci cally to
engage with minority and female undergraduates majoring in
computing elds. Students who attend the conference
participate in competitions and attend training sessions to
support engagement and research. Studies conducted at prior
conferences has shown that while students were engaged in
the training sessions and vigorously involved in learning they
did not develop the lasting social connections that can arise
out of conferences. Snag'em was designed to engage
students in social networking through gami cation of the
process. Prior research has shown that social games can help
people to engage in otherwise challenging or uncomfortable
situations [
Snag'em functions as a large human scavenger hunt. Players are assigned a set of relevant tags (e.g. \I'm a games researcher", or \I'm interested in data-mining"). They are then assigned a set of missions (e.g. \Find someone who specializes in HCI") which they must complete by identifying and engaging with an appropriate individual. The system was developed in PHP with a MySQL backed and provides a web-based front end for players to edit their pro le and to record interactions. We have also developed a mobile version of Snag'em which allows players to access the game via tablets and smartphones. The game itself is designed for easy deployment to new conferences and we are presently adding features that will allow us to automatically populate the database with initial tags.
Figure 1 shows a snapshot of the mission browser screen from the web version of Snag'em. Contact is registered when the players enter a 4-digit ID from the other person. In addition to missions the systems also allows players to record notes about one-another for future reference (e.g. \I should e-mail my proposal to him after the conference") and to send one-another messages. A sample message from the mobile interface is shown in Figure 2. Snag'em can also be congured to suggest speci c individuals that students should make contact with based upon their mutual interests or social connections.
The system logs all player interactions including tag updates, missions completed, notes made, messages, sent, connections added, and so on. This provides a rich dataset of information that we can use to analyze social patterns at conferences and to improve the impact of the intervention. In addition to the raw logs the game contains a number of features to support easy analysis. The developers have created a set of badges that allowed administrators to easily track the number of people playing via the mobile or web interfaces as well as the number of missions completed. The badge system also provides a simple visual record of the types of features (i.e. notes, tags, avatars) each player is using. The badge systems also allows administrators to note the frequency of use, time of day that players are online and so on. To date, Snag'em has been used at seven academic conferences. It has also been deployed to help incoming freshman and transfer students connect at four academic institutions. In 2009, for example, Snag'em was used by new students in the College of Computing and Informatics at the University of North Carolina at Charlotte. Students were able to play the game during the freshman orientation week with kiosks available for students to sign up located in the College of Computing and Informatics. SNAG'EM was used alongside other social activities to get students acquainted with each other, the faculty, and the CCI campus.
We have studied the impact of Snag'em on users and found
that playing the game improved conference attendees' sense
of community [
In analyzing the game mechanisms we have focused primarily on the STARS 2009 dataset. As mentioned above STARS is primarily targeted at undergraduate students speci cally females and underrepresented minorities. We deployed the system via the conference infrastructure and set up a table near the registration booth. The game was active during the rst two full days of the conference. The conference had 280 attendees 60.0% of whom were female (N=168) and 70% of which (N=196) were students. Roughly 28% of the conference-goers played the game (N=80) of whom 50% were female. In previous analysis 35.0% of the players were classi ed as active. It is important to note that this data was collected on an earlier version of SNAG'EM where players could snag each other only once, and only a single mission was available at a time. Because completing missions was signi cantly more di cult in this version of the game, players were classi ed as active if they completed at least two missions. An additional 50% of the players were classi ed as Interested, meaning they did more than just register for the game or that they completed one mission. Our analysis of this data was focused primarily on the mission and scoring systems. In 2009 the mission system was relatively simple and focused solely on guiding students to locate a single individual with a desired tag. Players were then guided to record the match via the ID system discussed above. Both the missions generated and points received were determined by the state of the current network. When generating missions we attempted to ensure that they were of varying di culty, and were relevant to the current user. In this iteration of the system the missions could only be satis ed by identifying someone whom the user had not previously snagged. The target tags were selected from the full set listed in the system. Easy missions were assigned high frequency tags (more than 12 of the non-adjacent users), while medium missions were assigned tags that are present in 14 of non-adjacent users and hard missions required tags present in less than 41 of the non-adjacent community.
The di culty of the mission determined the base score which was then modi ed by a connectedness factor. This factor was greater than 1 if adding this connection expanded your \Friends of friends," that is, the number of vertices less than 2 edges distant from the user. The connectedness factor was less than 1 if you completed the mission using the ID of a person you were already adjacent to, In this way we hoped to encourage players to branch out.
When developing the system we had hoped that players would develop social networks that exhibited breadth (i.e. meeting lots of people), depth (i.e. getting to know some individuals well), and mutuality (i.e. snags in both directions). We therefore hoped that users' immediate neighborhoods would be large and relatively dense with multiple snags between some people and bidirectional connections. When analyzing the STARS 2009 dataset, however, we found that this was not the case. Rather the game mechanics encouraged players to make a relatively large number of unrelated connections which, in turn, produced relatively broad and shallow social neighborhoods with very few inbound arcs. In fact some players actually opted to hide their IDs so that no other player could gain points by using them to complete a mission. As a consequence the attendees were actually less likely to engage in the deep and meaningful conversations required or to form lasting connections.
In response to these results we have overhauled the scoring system. This included changing the connectivity bonus to reward players based upon the size of the largest clique that they participate in. Players are now rewarded more for expanding this clique, thus deepening their social networks, than they are for adding an unrelated individual to their friends of friends. We have also allowed players to resnag the same individual for multiple missions with a low penalty for re-snags, and have begun to reward players with points for allowing themselves to be snagged to help others complete a mission. We have not yet analyzed the e ects of these changes on a the dataset.
We have used two measures of importance when identifying critical players. The rst is the simple interaction frequency as measured by the number of outgoing arcs from a player in the network. The second is membership in maximal cliques, that is, cliques which are not part of a larger clique. Players that participate in a large number of maximal cliques are hubs. We were able to identify three distinct user communities in the STARS 2009 dataset that centered on these hubs. A sample community graph is shown in Figure 3. We also found that the activity of these hub players was highly correlated with the activity of the other players in the community (r=0.827). A graph of these spikes is shown in Figure 4. More speci cally, on any day where one or more of the hub players were active, we observed spikes in the number of interactions taking place across users. We were able to observe a similar e ect (r = 0.659) on days when the developers had a booth/kiosk available.
We also performed an analysis of hub players using the
UNCC Student Orientation dataset described above. In this
dataset 91 of the 1290 potential students registered to play
Snag'em of which 22% (N=20) were female [
Our prior research has focused on identifying key players using graph methods. We plan to continue examining these key players in future work and to modify the mission selection criteria to better engage players that have not been active recently. Our chosen method of community detection, based upon maximal cliques, is both computationally expensive on large networks and can change substantially based upon small shifts in the network. Using a simpler, less volatile measure to identify community centers would allow us to adapt the gameplay based upon those communities more e ciently. This would in turn enable us to encourage new players to speci cally seek out these active players in an e ort to better engage them from the start. Di erent community detection algorithms might identify di erent hub players, or provide di erent ways of scoring missions that help to foster larger communities. Further development in this area might facilitate play in the absence of an instigating `active player' or outside of areas with an active game station or kiosk.
One open question is how to better identify hub players during the game, and modify mission selection criteria to engage inactive players or players who don't need motivation to network. These `social elites' are important to attract, as they are precisely who we should be encouraging our players to network with. If we are better able to build and analyze our networks, we may be able to o er features to these social elites that would attract them to Snag'em as a system more than the gami cation aspects would. We hope to explore techniques for reliably generating edges and tags for users based on existing data sources like conference proceedings or citations. This would reduce the burden of entry on new players, particularly elites, and make it more likely for those users to participate in networking (if not gameplay) using SNAG'EM.
We also plan to expand our in-game evaluation of Snag'em itself. We are presently adapting the system to poll players for their opinions as the system is used. This will better help us to identify the immediate impact of the system on users' social connections. We will be deploying some of these new features of the system during the 2014 Educational Datamining Conference in London as well as subsequent conferences in 2014 and 2014.
This research was supported by the NSF GRFP Fellowships No. 0900860 & No. 1252376 and BPC Grant No. 0739216 and No. 1042468 Thanks to all developers who have worked on the SNAG'EM project. The authors also wish to thank Shaghayegh Sahebi for her expert advice.