Shared identity and personal ties in influencing cooperative behavior

Hao Jiang1 & John M. Carroll2
  1. PhD Candidate, Pennsylvania State University, Email: hjiang@ist.psu.edu
  2. Pennsylvania State University, Email: jcarroll@ist.psu.edu

INTRODUCTION

As information technology pervades modern society and creates new forms of social life (e.g. online communities, virtual organizations, etc), concepts rooted in sociology and economics begin to emerge in information studies. These concepts are key to understanding information technology in its social context and to building IT that benefits communities. Social capital is one of these concepts. Researchers in community informatics and related domains have realized how important social capital is to understanding the relationship between IT and people's social lives, and to building robust communities assisted by technologies (Blanchard & Horan, 1998; Hampton & Wellman, 2003; Simpson, 2005; Daniel, Schwier & McCalla, 2003; Quan-Haase & Wellman, 2004; Chiu, Hsu & Wang, 2006).

IT has been studied in community context since the Internet was available to the public (Carroll & Rosson, 1996; Gurstein, 2000; Lee, Vogel & Limayem, 2003). Social capital, as a genetic element of social interaction, exists in all types of communities; studies of the interactions between social capital and IT have been urged (Foth, 2003; Pigg & Crank, 2004; Simpson, 2005; Stoecker, 2005; Williams, 2008). Community life is built on everyday social exchange (Berger & Luckmann, 1967). Along with the macro-level social elements such as social institutions, authority systems, trust systems and so on, micro-behavioral components of social life such as social exchange among social actors give us insight on how communities build and where they project (Coleman, 1990). The study we introduce in this paper focuses on detailed, concrete social interactions of this kind: social exchange among social actors and their resource allocation behavior in the face of different social relations.

Social relations are considered sources of social capital. However, how social relations shape social capital is not clear. Furthermore, social relations are diverse in types and forms, so do they contribute to social capital in a unified way? Finding out the impacts of social relationships on social capital is what we are after in this study.

We simulated a social exchange scenario with a three-person economic game. With this online game, we are able to see how different types of social relationships, which are considered a major source of social capital, influence the distribution of resources and behaviors of social actors. We first briefly discuss the constructs in which we are interested: social capital, shared identity and social ties; then we elaborate on the experiment we conducted. Next we report on data collection and data analysis. Lastly, we present results and discuss future direction.

SOCIAL CAPITAL AND SOCIAL RELATIONSHIPS

Social actors need resources to achieve any goal, whether it serves an immediate need or a long-term interest. However, resources people possess are not only limited in type (so we have to exchange with other people for resources we do not possess), but also in amount (so that we have to allocate them with discretion). When we want to achieve something but lack the resources required, we must get them through exchange (Blau, 1986; Coleman, 1990). This is the building block of social interaction. Social exchange and interaction can become very complex and eventually constitute social systems (Coleman, 1990). Any given community's development requires resources as well. The well-being of its members and the development of a community are tightly connected.

In conceptualizing the ability of a social system to produce or nurture public goods, scholars borrowed the term "capital" from economics and coined the term social capital. Since the 1980s when it was introduced and conceptualized in sociology and economics (Bourdieu, 1985; Coleman, 1988; for review, see Portes, 1998), social capital and its ramifications have been developed in many social sciences: management, politics, and information science, just to name a few. Social capital is considered an important resource in society that helps enable social actions of different kinds, and that benefits society in many ways (Coleman, 1988; 1990). In Coleman's (1988) discussion of social capital, the Jewish diamond market in New York City shows a high rate of cooperative behavior, enabled by a high level of trust; its close connections among parents provide guardianship for youngsters. These are just two examples of how important social capital is to local communities.

Research in information science also started to pay attention to IT and social capital with the advent of the Internet, although the exact term was not used in early studies. Carroll and Rosson (1996) introduced their effort in developing the Internet-enhanced Blacksburg community. Blanchard and Horan (1998) studied IT-assisted local communities and found that social capital and social engagement increase when online life and local life are connected and when the online community supports communities of interest. In addition, they also identified some types of online communities: education, government and political participation, and sharing general community knowledge. Hampton and Wellman (2003) concluded similarly that IT could assist traditional community with locally oriented content, such as online discussion groups supporting people living in close proximity. In the knowledge management field, many studies found a positive relationship between social capital and knowledge sharing behavior. Wasko and Faraj (2005) found positive connections between knowledge sharing and social capital in the form of social approval, respect, and so forth (Wasko & Faraj, 2005). Chiu, Hsu and Wang (2006) surveyed 310 online community members and confirmed the positive effect of social capital on willingness to share knowledge. Besides findings coinciding with the conclusions of Wasko and Faraj (2005), they also found community-related outcome expectation and personal outcome expectation could influence the motivation of knowledge sharing in virtual communities.

Key authors (Bourdieu, 1985; Coleman, 1988) emphasized the non-contractual factors of social systems to reduce transaction costs and to enable social exchanges that would be impossible otherwise. Compared to contractual situations, social capital tends to be low cost in maintaining social exchanges and interactions because it is voluntary. However, being non-contractual, social capital is not protected by strong reinforcement and therefore expected outcomes are not guaranteed. Scholars found social capital is fragile in the sense that it is usually a byproduct of social exchange and lacks nutrition to grow and be maintained in a persistent way (Coleman, 1990; Putnan, 1993). Studies have shown that, for example, families moving away from a community can easily destroy established social capital (Hagan & Wheaton, 1996; Portes, 1998) and the causes of moving are many: new jobs, unanticipated life events, and so on. In communities that operate on IT, the bond and relational reinforcement are even weaker, so it is harder to maintain social capital at a consistently high level.

Social relations are considered a major source of social capital (Coleman, 1988, 1990; Blau, 1986; Scott, 1988; Burt, 1997; 2000, 2001). However, the question that interests us more is how social relations affect (generating, maintaining, and disseminating) social capital. The social relations we participate in are rather diverse in forms and types. Two types of social relations are fundamental; namely, identity-based (shared identity) relations and social ties. Identity bond or identification with a social group entails a connection between a person and a group; that is to say, identification with a social group does not require bonding with any individual members of that social group (Ashforth & Mael, 1989). Personal ties, on the other hand, are dynamics between the very persons who are connected. Friendship is a relationship of this kind. Both types of relation enrich our social experience and activities. In our professional and personal life, we receive help from and have fun with people from both types of relations. We interact with people we know very well and are closely connected with: parents, brothers and sisters and close friends; we also engage in activities with people whom we have not met before and may never meet again in our lives.

In research, shared identity and social ties have been the focus for decades. Social ties can bring various resources, from financial to informational, and from social and to emotional. Job searching was studied in social network research and personal ties were shown to lead to successful job searching (Granovetter, 1973; Scott, 1988; Lin & Dumin, 1986; Burt, 2000); social ties are also found to provide emotional support (Ellison, Steinfield & Lampe, 2007; Gilbert & Karahalios, 2009; Cross & Borgatti, 2004). On the other hand, studies show that shared identity increases helping behavior in social interactions (e.g. Haslam et al., 2005; Levine et al., 2005).

Through critically analyzing existing theories and studies, researchers have raised concerns that shared identity and social ties may affect social capital in different ways, and that the term social capital in different contexts may refer to constructs that are qualitatively different underneath (Jiang & Carroll 2009). Both relations are considered able to mobilize resources among social actors (Granovetter, 1973; Burt, 1997, 2000, 2001; Tajfel et al, 1971; Kramer, 1995; Brewer, 1995), but since they have different origins and function differently in influencing human behavior and decision making, for us it is of interest to see if and how these different types of social relations influence social capital and how they influence it jointly if at all.

According to social identity theory and related studies (Ashforth & Mael, 1989; Brown, 2000; Brown & Williams, 1984; for recent review, see Amoit et al., 2007) a relationship can be built between a person and a social group when one identifies with it (e.g., "I am a community informatics researcher"); and to have and maintain identification with a group, a person does not have to have a personal connection with other individual members. Therefore, it is a relationship based on the connection between a person and a social group. A social group can be a social unit, such as a team consisting of several people, or a community with a large number of members, or even an entity as large as a nation.

During social interaction, a social categorization process will take place - a process that helps us to understand other people by classifying them into different categories based on certain perceivable traits - and further social categorization of the self also known as self-categorization - associating the self with a prototype of group - will result, identifying with the social group and hence a social identity forms (Tajfel et al., 1971; Hogg & Terry, 2000). These processes are cognitive possesses that help a person to build a self-concept. Also, they can lead to behavioral consequences, which eventually help construct and reinforce social relations.

PROPOSITIONS

The most important implication of social identity theory to us is its prediction of in-group favoring over out-group and perceived interchangeable characteristics of its members (Tajfel et al, 1971; Brewer, 1979; Brown, 1969; Brown & Williams, 1984; Brown, 2000). Ashforth and Mael (1989) summarized three consequences of social identity. First, identifying with a group will increase commitment to the group; second, it will increase group cohesion, cooperation, and altruism; and third, group identification will reinforce perceived group distinctiveness, prestige, and salience. Considering all these effects of shared identity, we come to our first proposition (Proposition 1): that shared identity will help increase social capital within a social group by guiding members to choose group-favoring actions.

We are also interested in social ties, which are connections between social actors. Compared with relations based on shared identity, ties are more personal and are maintained on an interpersonal basis, measured through the amount of communication time, emotional intensity, intimacy, and reciprocity of service (Granovetter, 1973). Realizing that ties vary in strength, and that weak ties can be of great value, researchers in social network analysis have developed a typology of social capital, which characterizes resources mobilized by ties of different strength or locus. An often-cited typology distinguishes bridging social capital and bonding social capital (Putnam, 2000; Elison et al., 2007). Ties can also be classified as internal or external, given a network as a larger context. In general, weak ties tend to bring resources not locally possessed (e.g. new information, jobs, new technologies and opportunities), because weak ties usually connect people with more diverse backgrounds. Strong ties, by contrast, often are built on intense interaction, high intimacy, and similarity, therefore bonding people and bringing social/emotional support.

There are a few challenges related to the distinction between strong and weak social ties. For example, there is no qualitative difference between strong ties and weak ties; they are only relative measures. A weak tie can be seen as strong if the frame of reference changes. The distinction between internal and external sounds qualitatively different, but the same challenge still stands - if the frame of reference changes, an external tie can become an internal tie (Adler & Kwon, 2002; Jiang & Carroll, 2009).

Personal ties are considered as conduits between two connected individuals. Drawing from this point, we have our second proposition (Proposition 2): personal ties will keep resources within connected social actors in a social group. In terms of social capital, therefore, ties presented in a social group may compete for resources with the rest of the group. In terms of distribution of exchanges, this means that a social group with personal ties will demonstrate uneven resource distribution.

Propositions 1 and 2 are derived from studies of shared identity and social ties respectively, and they point to somewhat conflicting effects on social capital in social groups. Shared identity is hypothesized to produce high social capital for groups; social ties lead to unevenly distributed exchanges within a social group and potentially compete for resources with the rest of the group, resulting in possibly less social capital for the group. Considering that shared identity consolidates the connection between a person and the collective, and produces interchangeable characteristics in the collective's members, linking with Proposition 2 we can draw our third proposition (Proposition 3): in a social context with high shared identity, shared identity will tail off the concentrating effect of social ties on resource allocation.

STUDY DESIGN

To explore these propositions, we invited participants to play an online economic game. In sociology and economics, economic games in lab settings have been used intensively to explore human decision-making and cooperative behavior (Simon, 1959; Dawes, 1980; Liebrand et al., 1986). In fact, laboratory experiments are the major source of knowledge in the social sciences (Falk & Heckman, 2009). Lab experiments and economic games have been questioned for lacking real-world relevance, given that true social life takes place on a larger stage than life in a lab, and reacts to different stimuli. However, we see laboratory studies as magnifiers of certain social elements of researchers' interest. Lab studies, when carefully designed, can well mimic interaction in the real social world rather than creating an alien universe. As we will see later, our study is designed in a way that leads participants to play a realistic game. They are not forced to decide with whom they want to cooperate, which would seem to be very unrealistic. Recently many lab experiments have been reported to specifically study social relations (e.g., Rothstein & Eek, 2006; Rand et al., 2009; McCubbins, Paturi & Weller, 2009; Centola, 2010). Rothstein and Eek (2006) experimentally tested social trust, asking people of different cultural backgrounds to react to political corruption stories. Rand et al. (2009) devised a public good game to test how reward and punishment can influence cooperative behavior. They found that positive interactions can promote public cooperation. Centola (2010) also used an experiment to study how behavior diffuses through social networks. His experiment was not in the lab, but the social network was an artificially structured online community. McCubbins et al. (2009) used a lab experiment to study conditions under which groups can solve coordination games.

The experiment we designed, like many other economic game experiments, allows us to dig into micro-level social behaviors that account for macro-level social movements and construction of social systems. Without knowing the former, the existence of social systems at higher levels would remain mysterious. Knowing how individuals react to social environments not only gives us a better understanding of how social systems take shape and how they work, but also opens windows for practitioners and action researchers to provide interventions on a more accessible ground.

Social capital, the dependent variable in the study, is a very broad concept. It is impossible to use one or two variables to capture a comprehensive construct of social capital. In the game we designed, we used cooperation or willingness to cooperate as an indicator of social capital, and the measure we used is transactions between players. We manipulated the independent variables (shared identity and personal ties) with light intervention, which we will see in detail. Cooperation and the willingness to cooperate are appropriate in our case for a few reasons. First, cooperation is seen by many researchers (Coleman, 1988; Putnam, 1995; Knack & Keefer, 1997; Woolcock & Narayan, 2000; Fukuyama, 2001) as a major form of social capital; second, cooperative behavior, willingness to cooperate, or cooperative norms in fact underlie a wide range of forms of social capital, such as observance of social norms, taking part in collective action, sharing resources, and so forth.

Baseline Design

We conceived a 3-person, online economic game. In this game, we created a background story of a developing human habitat evolving from Stone Age to Gold Age. In each game, the human habitat starts off as an ancient tribe in a very primitive setting. With a limited amount of initial gold, members of a habitat invest in various projects allowing them and the habitat to survive and develop: seeding techniques, fishing technology, framing skills, weapon making, and so on. As these investments return, part of the gain will go to the habitat as the resource of its evolution, and another part will be the profit for individual investors. However, to successfully invest in a project, a player needs to find another person in the same session to co-invest with, which means they share equal cost (the price of a project) and return. Co-investment is project-wise, which means for each new investment, players can choose a different player to co-invest with.

Each investment varies in price and return. However, a remarkable difference is that among all the investments, some give more return to the habitat and less to individual investors; some give more return to the investors and less to the habitat. This difference is introduced into the game to ensure measurable collective/individual favoring behavior. The project pool will replenish periodically, so for each individual player after each investment being made, the pool will be refilled to ensure an equal number of habitat- and investor-favoring investments. Players can cancel co-investment requests they sent. Those receiving co-investment requests can either accept, meaning the projects are successfully invested, or they can decline.

The entire experiment takes about 60 minutes or less, depending on how quickly the habitat reaches Gold Age. This one-hour experiment includes participant consent, reading instruction materials, playing the game, and filling out a post-game survey.

Manipulations

Throughout the game, there is no face-to-face contact for all three participants in the same session. One of the participants is kept in a separate room the whole time. Based on the idea of social identity theory, merely assigning people into groups will lead to social categorization and in-group/out-group effect (Tajfel, 1970). Manipulation of shared identity (independent variable) is achieved by letting participants in the same session discuss a name for their team before the game. This discussion is carried out online, so the participants do not see each other. We also have groups that do not receive this option.

Among three persons in each group, we expect two of them to have some personal contact that can help them build personal ties (independent variable). We achieve this by letting the first two participants coming to the lab chat a few minutes face-to-face before the third participant arrives. The manipulation we use is light, so that the participants will not feel unnatural. The results show that our manipulation, although not heavy, does affect the participants' behavior. Table 1 shows a summary of variables and manipulation of the study.

Table 1 -Variable summary
table1

PARTICIPANTS

We recruited 51 university students from a large northeast American university where our lab is located. All participants received monetary compensation. Most participants were enrolled in information or computer science-related majors. Before they participated in a session, we made sure they did not know each other.

In piloting the game, we found that participants did not engage in the game very much when they were all compensated equally, each person receiving $10 for their participation. The task itself is not cognitively demanding. Some participants reported to us that when they played the game, they randomly selected people to co-invest with, because no matter what their project and co-investor choices were, the game would eventually end and it made no difference to them what they got from the game.

To make the game more appealing and attractive to players, we decided to compensate them based on their performance. The total compensation of each player consisted of two parts: individual compensation and group compensation. For the individual part, players who made the highest personal fund got $9; for second place, $7; and $5 for third place. Each group gained a group compensation, which eventually was divided by three and added to players' individual compensations. Group compensation was based on group performance, which was a function of number of transactions a group took to finish the game. The fewer transactions a group had, the higher performance the group produced; low-performance groups got $0, high-performance groups got $9, and middle-performance groups got $6. Thus, for example, at the end of the game, a player in first place in a middle-performance group will get $9 + $6/3 = $11 in total.

Some researchers have raised the issue that basing compensation on participants' performance may bias the game. Originally we compensated our participants with an equal $10 each. However, participants not paying attention to what they do in the game will undermine the study. Furthermore, in the compensation mechanism we used in data collection, we also included a collective part that can help offset incentive for competition led by the possible bias, if any. What we are looking for is the difference between groups and players in different relations. If all participants have the same baseline, we will be able to see the difference. To be more specific, for example, we expect to see that groups with high shared identity will have more community-favoring investments than groups without. Also, to make their compensations tied to performance, the game will be a bit more fun and make participation more natural.

DATA COLLECTION

We ran the study from late November 2010 to mid-December the same year. By the end of the semester we collected data from 17 groups (51 individuals). We collected all transactions among three players in each session. The transactions included players' co-investing requests, along with the cancellations, acceptances, and declinations of those requests. With these data, we know what types of investment (community-favoring vs. individual-favoring) they chose and with whom they co-invested.

Twelve groups had group chat sessions; 5 groups did not receive this option. Seventeen of the groups included pairs who had a face-to-face chat; 34 did not. The unequal size of these latter two categories was due purely to logistic issues: We ran sessions of groups with chat exclusively at first and sessions of groups without chat later. However, the number of participating groups dropped during the final exam week. At the group level, the number of groups not receiving chat treatment was small, but in most of our data analysis we paid attention to dyadic and individual data. In this way, statistical power was increased with increased data units.

Among these 51 participants, 46 were information science majors, 2 from the computer science department, 1 from psychology, 1 from human resources at the College of Liberal Arts, and 1 was studying economics. They have been using computers and the Internet at least for 6 years and the longest users have used computers for about 20 years. The average number of years using computers was 12.9 years (σ = 3). Participants were 21.3 years old on average (σ = 2.02).

ANALYSIS AND RESULTS

Experience in the game

In the post-game survey, we asked players their general feelings about the game. With these questions, we wanted to make sure that the game gave participants a positive experience, not necessarily in favor of the study, but at least not getting in the way of collecting data. These questions were included to gather information about participants' experience and to see if the system and procedure led to breakdowns or pitfalls that may compromise the study. Participants used a 1 to 5 (strongly disagree to strongly agree) Likert-scale for each question in Table 2.

Table 2 - Post-game survey: player experience
table2

The result (Figure 1) shows that most players had a positive (good) experience playing the game.

Figure 1 - Player experience
figure1

We also collected open-ended responses about their general feeling about the game. It confirms that most participants enjoyed the game and had fun during their play, for example:

"Better than I expected. The several seconds of pause during each investment session is quite reasonable and useful."

"The game was simple, fast paced, and interesting."

"It was pretty fun. Hard to work with other players without actually being able to talk to them, but it worked pretty well regardless."

"It was enjoyable, I would probably play again."

Transactional data

In current design, the game ends when the habitat evolves to the Gold Age, which requires 7500 points. This is the only condition for ending the game, so groups that invest in more community-favoring (CF) investments will reach the Gold Age sooner and the transactions of these groups will be fewer than groups that choose more investor-favoring (IF) investments (Table 3).

Table 3 - Summary of transactions of each team
table3

T-tests on total transactions (Table 3, column 2) and on total requests (Table 3, column 3) show that groups with chat sessions did have fewer total transactions and fewer requests sent during the game, and thus higher performances for these groups (total transactions: diff = mean (no) - mean (yes) > 0, p = 0.014; total requests: diff = mean(no) - mean(yes) > 0, p = 0.014). Since we unfortunately have an unequal sample in the second category, we box-plotted group transaction data in Figure 2 to visually show the difference between the two categories for visual investigation.

Figure 2 - Box plots of transactions of group with and without chat session
figure2

Figure 2 shows 4 box plots of transactional data. Charts A, B and D show a very similar pattern, namely that groups without a pre-game chat have a considerably higher number of transactions (total transaction, Chart A; total requests, Chart B; total investor-favoring requests, Chart D), and Chart C shows that groups with chat had relatively more community-favoring requests than groups without pre-game chat.

Figure 3 shows the comparison between groups with chat and without chat on their Community/Investor-favoring requests, expressed as a percentage (Column 5 and 6, Table 3). Charts C and D in Figure 2 and Charts A and B in Figure 3 present similar results: that groups with a chat session chose more community-favoring investments than groups without chat sessions did, in both numbers and percentages. The t-test and the box plots confirmed our Proposition 1, that members in a high shared-identity setting tend to endorse collective-favoring behaviors than members in a low-identity setting.

Figure 3 - Box plots of Community- and Investor- favoring request in percentage
figure3

Another explanation for the difference between groups with pre-game chat and those without is due to the relatively high success of transactions in groups with pre-game chat - in other words, the small number of failed transactions (declination and cancellation). Again, since the size of groups is small, we resort to graphic presentation to help us interpret the data (see Figure 4).

Figure 4 - Unsuccessful transactions of groups with pre-game chat and groups without
figure4

Table 4 shows requests between individual players as a percentage. We are interested in the effect of social ties on how people select cooperative partners and their preference for pro-social behavior. To do so, we looked at requests sent by players with tie coded as P1 and P2. Player 3, the stranger, is coded as P3. For example, column 2 in Table 4 means requests sent by player 1 to player 2 (P1_P2), and column 3 means requests sent by player 1 to player 3 (P1_P3), both in percentage. In total, we have 34 pairs for each category: pairs with face-to-face conversation and pairs without. Having these ratios calculated, we performed a paired t-test between pairs with face-to-face conversation and pairs without. To be clear, the test is performed between, for example, player 1 to player 2 (persons who are connected) and player 1 to player 3 (persons without ties). The result shows a significant difference between these two groups: the request ratio between pairs with primed personal ties is higher (p = 0.0001), which means transactions between people having ties are significantly higher than transactions between people without previous ties, and thus supports our Proposition 2.

Table 4 - Transactions between players
table4

To further analyze transactions in groups, we calculated the variance of transaction ratios on columns 2, 3, 4 and 5 (Table 4). To see if shared identity has an effect on personal ties in concentrating resources between connected persons, we ran a t-test between pairs with ties and those without. The result is diff < 0 with p = 0.0375. Since the sample size on the variation of the low shared identity group is small, we list the descriptive statistics in Figure 5.

Figure 5 - Descriptive statistics on low shared identity and high shared identity groups' request variance
figure5

Another t-test was performed to see among pairs with personal ties, if those from sessions with pre-game chat had less concentrated transactions. To do this, we looked at whether the difference between the percentage of transactions sent from players to their connected players was lower than those sent to the stranger, and whether such transactions were fewer in settings with pre-game chat than in settings without. First, we calculated the ratio difference (column 7 for P1 and 8 for P2) between requests players sent to their connected players (column 2 for P1 and 4 for P2) and those players sent to strangers (column 3 for P1 and 5 for P2). Then we performed a t-test grouped by groups with pre-game chat and groups without. The result confirmed that pairs with ties from groups having chat sessions had a lower percentage of total requests than pairs with ties from groups without chat sessions (p = 0.0122). Therefore, Proposition 3 is supported.

DISCUSSION

Our three propositions are drawn from existing theories and studies. Proposition 1 specifies the positive effect of shared identity on social capital, in that shared identity will encourage group favoring actions; Proposition 2 is about social ties and their concentrating effect on resource allocation; Proposition 3 postulates an effect of shared identity over social ties on reducing social ties' concentrating effect described in Proposition 2. Data analysis in the previous section confirmed all three propositions.

Social groups, from small teams that contain only a few people to large entities that represent nations, unavoidably operate via interactions between social actors on a day-to-day basis. Social relationships that influence how people interact with each other impact cooperation among people. Cooperation and willingness to cooperate are forms of social capital that we can see in everyday life. From intense collaboration in small social groups in real time, to construction of culture and norms in a large community over time and space, sharing resources and taking part in collective actions of different kinds sometimes determine success or failure. Cooperation can take many forms. Sharing resources with others for public goods, observing social norms (not littering, not line jumping), and volunteering in community, can all be considered as cooperation in various forms.

All forms of resources possessed by social actors are limited. Physical and financial resources are obvious examples; other resources, such as intelligence and emotional energy, though considered reproducible over time, at any given period or moment are very limited. If one gives one's attention to something or someone, one cannot give it to other things or people at the same time. With the limitations imposed on the resources that we can allocate, how social actors allocate resources becomes extremely important and of interest, and finding out what factors can influence our resource allocation behavior can give clues to both researchers and social practitioners to better understand and build communities and information technologies for communities.

The game we designed in this study is a public good game with a more complex social relation setting and background story. The game is designed to simulate a social context where multiple social relationships co-exist and people can have various choices that may conflict with each other. Economic games such as prisoners' dilemma and public good games are often used in sociology and economics to help researchers understand human behavior in different conditions. Many public good games have been created to simulate situations where selfish choices are detrimental to public well being, such as preventing global warning and air pollution. In these cases, collective and individual interests may compete. For preventing air pollution and saving non-reproducible resources, immediate benefit can be very little and may not directly affect action takers; and at the same time, actions that go against public good (i.e. producing more carbon dioxide) may give irresistible immediate returns. This is the tension that public good games try to simulate. Public good games have been tested under different conditions: if the transaction is repeated or just a one-shot choice; if the identities of players are revealed in the game or not; and if punishment or reward is presented, contingent on the choice players making (Rand et al., 2009).

It is true in communities, no matter online or offline, that the same tension exists between the immediate and long-term public goods and benefits for the certain individual members of the community. Since the resources - financial, mental, emotional, and other - that a person possesses are very limited, where and in what amounts we commit our resources can have subtle but deep consequences upon the development of communities we participate in or even those we are not part of. If we consider the larger context of social groups in a connected ecology, the distribution of resources becomes a zero-sum game, where if one puts something in one sub-group, then other groups lose it.

For individuals we face similar dilemmas. Should I go to the town hall meeting this evening or play tennis with my friends? Should I give my opinion on a research topic that a colleague posted on the discussion forum or write one more section of my own article? Questions like these are hard to answer without context, and it may be trivial at times for individuals, but taken together they contribute to building the larger social context, the community.

In the game we designed, besides the tension between public goods and individual benefits, we also brought in another element, social relations, which in real social exchange always play a significant role. This component was manipulated to form personal ties and shared identity in groups. We found support for all three propositions. The analysis we employed is very simple.

The first proposition focuses on the effect of shared identity in social groups with social ties. Our experiment pointed out that groups with shared identity (groups having a pre-game chat session to determine a group name) show higher performance in terms of significantly fewer total transactions (p = 0.014) and fewer total requests sent (p = 0.014) against groups without the treatment. Groups with a high volume of community-favoring investments will reach the Gold Age quickly and a high rate of successful investments also helps increases group performance by reducing incomplete transactions, such as cancelling and declining requests. In real-world social exchange, transactions can be costly. Unsuccessful social exchange will waste a social agent's time, as well as financial, physical, mental, and emotional resources, so keeping transactions successful at a high rate becomes critical, especially in a world where resources are scarce.

In the data we collected, the fastest habitat to reach the Gold Age generated only 98 total transactions, and the lowest one produced 339 total transactions, which is more than three times the former. In this setting, we consider willingness to cooperate and endorsement of community-favoring investments as forms of social capital, which lead to fast growth of public goods and a high rate of successful transactions. Critics have commented that a high volume of transactions in groups without chats prior to the game may be that the result of players in these groups needing more transactions to create a social climate and reciprocity, while groups with previous chats had already established a stage of reciprocal understanding and strategy. Social interaction history, in other words, can enhance the interactions that follow. We totally agree and that is what we are arguing for. However, we doubt that the chat in our design can add an impact significant enough to create strategic guidance for the players to carry out. We explicitly asked the players to discuss a name for their group and their actual conversations took a very short time. As soon as they reached a name, the chat was shut down. Some of chats only took about 30 seconds with less than 20 messages exchanged. We did not see any conversation related to the tasks at hand. The higher number and percentage of community-favoring investments in groups with chat sessions also points to the fact that the high performance of these teams is not only due to reciprocity, but also due to their pro-social choice.

The second proposition postulates that personal ties can constrain resource exchange between connected persons and thus limit its externality (Coleman, 1988). This is also supported by our study in that players have a significantly higher ratio of requests sent to persons with whom they have ties than to strangers with whom they are not connected (p = 0.0001). Many existing studies point out benefits brought by social ties such as job searching, information flow, and career advantage in a given social network. In our study, we consider social ties in collective settings, and the experiment we conducted confirmed that social ties tend to concentrate resource exchange between connected social agents. More importantly, in a collective setting social ties can lock resource exchange in social ties. Social ties are important and our study shows that the influence of ties is very significant in terms of enforcing the transactions on social ties. However, in our analysis, we did not analyze the reciprocity between connected persons (e.g. P1 to P2) and between strangers (e.g. P1 to P3). A social tie can be either symmetric or asymmetric. Our analysis did not look into if transactions in a given group between player 1 and player 2 show higher reciprocity. Nevertheless, from the perspective of individuals, our study did suggest that people tend to exchange with those they already know.

Our third proposition is supported by the fact that in groups with shared identities (groups with pre-game chat sessions) the difference in percentage between players' requests sent to connected persons and to strangers is significantly lower than in groups without shared identity (p = 0.0122). This means that shared identity can help a social group or a community become more evenly distributed its social exchange among members. Put another way, shared identity can reduce the effect of personal ties in keeping exchanges between connected persons. In real life, unevenly distributed social exchange can lead to serious consequences. Shared identity can increase the likelihood that members are perceived as interchangeable, and thus it will matter less with whom one exchanges. This can be a reason why shared identity can help resources reach people without ties. This effect of shared identity in social groups is different from the one we mentioned in the first proposition, which is the pro-social effect led by identifying and committing to a social group one belongs to.

These three propositions give us much food for thought about social capital and social relationships. Social relations do bring benefits to people. In the past, research and studies have been arguing for positive outcome of social relations. The result of our study suggests that social relations can lead to different distribution of resources in a social group. This will enrich our understanding of social capital with a more refined lens, a view that looks at possible different consequences brought by social relationships that are different in nature.

It does not downplay the importance of social ties in social life. No one can escape from interpersonal relations with others. But the study does remind us to think deeper about social relationships in creating social capital or benefits in other terms. When we take all results of this study into consideration and put them into a more integrated picture, we find that shared identity and social ties as social relations both generate social capital, but in social interactions shared identity tends to increase the externality of resources one processes, making resources easily accessible to more people. Social ties work on another dimension, which is the activation of resources - making resource access more stable for people connected. Put another way, shared identity increases resource availability to a wide range of people in a social setting; and social ties make resources available to connected persons with high probability.

This study is an empirical exploration of how shared identity and social ties, as two distinct social relationships, can influence social capital distribution in different ways. Although the experiment is criticized for its lack of social relevance, we want to emphasize its ability to help researchers test relationships, even causal relationships, between variables. Rothstein and Eek (2006) stressed the same rationale for using experiment in their trust-related experiment. The analysis is rather simple and straightforward. The findings we reported here are drawn from a lab study we conducted. The study did show validity and the results did support our propositions. However, social relations and social capital are much more complicated than the study presented in this paper. The study reported does not claim that it describes a comprehensive view of social capital and social relationships; rather, it is intended to provoke the attention and interest of related research communities and to call for more studies exploring social relationships in detail. Only when more future studies are carried out will we be able to possibly draw a better picture of refined relationships among these components that underpin social life and social structures.

LIMITS AND FUTURE DIRECTION

The results we presented are from simple statistical analysis, and the data size is different in two conditions. The results from our analysis are significant, but we are looking for more data collection in the future. Also, we have not looked at reciprocity between players. In our data analysis, even though we are interested in social ties, we mainly took into account one-way transactions (e.g. whether people send more requests to players they know more). We believe more insights will result from looking at how reciprocal those connections are (e.g. if people send more requests to other players they know more, and then find that those other players do likewise). Ties can be either symmetrical or asymmetrical. Nonetheless, analyzing reciprocity will undoubtedly enrich our understanding of how ties can play roles in social groups.

In terms of experiment design, one limit of the game we used was that it was constrained to three persons and their communication was very limited. Furthermore, our study was designed in a way that it enforced cooperation (players need to co-invest). This design, in one way, guarantees we collected valid data in terms of transactions, but it does limit options for players. However, this design also represents real-world cases in which no exit options are available.

ACKNOWLEDGEMENTS

We thank the reviewers who commented and provided suggestions for our revision. Those comments and suggestions helped us refine the paper and more importantly helped us realize the limitations and promising future of this study.

REFERENCES

Adler, P. and Kwon. S. (2002). Social Capital: Prospects for a New Concept. Academy of Management Review, 27(1), 17-40.
Amiot, C., De la Sablonnière, R., Terry, D., Smith, J. (2007). Integration of Social Identities in the Self: Toward a Cognitive-Developmental Model. Personality and Social Psychology Review , 11(4), November. SAGE Publications.
Ashforth, B. E. and Mael, F. (1989). Social Identity Theory and the Organization. The Academy of Management Review, 14(1), 20-39.
Blanchard, A. and Horan, T. (1998). Virtual Communities and Social Capital. Social Science Computer Review, 16(3), 293-307.
Bowles, S. and Gintis H. (2002). Social Capital and Community Governanc'. The Economic Journal, 112(483), F419-F436.
Brewer, M. B. (1979). In-Group Bias in the Minimal Intergroup Situation: A Cognitive-Motivational Analysis. Psychological Bulletin, 86(2), 307-24.
Brown, M. E. (1969). Identification and Some Conditions of Organizational Involvement. Administrative Science Quarterly, 14(3), 346-355.
Brown, R. (2000). Social Identity Theory: Past Achievements, Current Problems and Future Challenges. European Journal of Social Psychology, 30(6), 745-778.
Brown, R. and Williams, J. (1984). Group Identification: The Same Thing to All People?' Human Relations, 37(7), 547-564.
Burt, R. S. (1997). The Contingent Value of Social Capital. Administrative Science Quarterly, 42(2), 339-365.
Burt, R. S. (2000). The Network Structure of Social Capital. Research in Organizational Behavior, 22, 345-423.
Burt, R. S. (2001). Structural Holes Versus Network Closure as Social Capital. in N. Lin, K. Cook and R.S. Burt, Social Capital: Theory and Research. (pp. 31-56). Sociology and Economics: Controversy and Integration series. New York: Aldine de Gruyter.
Carroll, J. & Rosson, M. (1996). Developing the Blacksburg Electronic Village. Communications of the ACM, 39(12), 68-74.
Centola, D. (2009) The Spread of Behavior in an Online Social Network Experiment, Science, 329(5996), 1194-1197.
Chen, Y., and Sherry X-L (2009). Group Identity and Social Preferences. American Economic Review, 99(1), 431-57.
Chiu, C.-M., M.-H. Hsu, E. T. and Wang G. (2006). Understanding Knowledge Sharing In Virtual Communities: An Integration Of Social Capital And Social Cognitive Theories. Decision Support Systems , 42(3), 1872-1888.
Cohen, D. J. and Prusak L. (2001). In Good Company: How Social Capital Makes Organizations Work. Ubiquity, 1(42), 3-3.
Coleman, J. (1988). Social Capital in the Creation of Human Capital. American Journal of Sociology, 94(s1), S95-S95.
Coleman, J. (1990). Foundations of Social Theory . Harvard University Press.
Cross, R. and Borgatti S.P. (2004). The Ties That Share: Relational Characteristics That Facilitate Information Seeking. Social Capital and Information Technology, 137-161.
Daniel, B., Schwier, R. A. & McCalla, G. (2003). Social Capital in Virtual Learning Communities and Distributed Communities of Practice. Canadian Journal of Learning and Technnology, Volume 29.
Dawes, R. (1980). Social Dilemmas, Annual Review of Psychology, (31), 169-193.
Ellison, NB., Steinfield, C., Lampe, C (2007). 'The Benefits Of Facebook Friends: Social Capital and College Students Use of Online Social Network Sites', Journal of Computer-Mediated Communication , 12(4): 1143-1168.
Emerson, R. M. (2003). Social Exchange Theory, Annual Review of Sociology, 2(1), 335.
Fukuyama, F. (2001). Social Capital, Civil Society and Development, Third World Quarterly, 22(1), 7-20.
Gilbert, E. and Karahalios (K. 2009). Predicting Tie Strength with Social Media. Proceedings of the 27th International Conference on Human Factors in Computing Systems. ACM: Boston, MA, USA.
Granovetter, M., (1985). Economic Action and Social Structure: The Problem of Embeddedness. American Journal of Sociology, 91(3), 481-481.
Granovetter, M.S. (1973). The Strength of Weak Ties. American Journal of Sociology, 78(6), 1360-1360.
Granovetter, M. (1982). The Strength of Weak Ties: A Network Theory Revisited. In P. Marsden and N. Lin, Social Structure and Network Analysis (pp. 105-130). New York: John Wiley and Sons.
Gurstein, M. (2000). Community Informatics: Enabling Communities with Information and Communications Technologies. London: Idea Group Publishing.
Hagan J, MacMillan, R. and Wheaton B. (1996). New Kid in Town: Social Capital and the Life Course Effects of Family Migration in Children. American Sociological Review, 61, 368-385.
Hampton, K., & Wellman, B. (2003). Neighboring in Netville: How the Internet Supports Community And Social Capital in a Wired Suburb. City & Community, 2(4), 277-311.
Handy, C. (1995). Trust and the Virtual Organization. Harvard Business Review,73(3), 40-50.
Haslam, S.A., O'Brien, A., Jetten, J., Vormedal, K. and Penna, S. (2005). Taking the Strain: Social Identity, Social Support, and the Experience of Stress, British Journal of Social Psychology, 44(3), 355-370.
Hogg, M. A., & Terry, D. J. (2000). Social Identity and Self-Categorization Processes in Organizational Contexts. Academy of Management Review, 25, 121- 140.
Huysman, M. and Wulf V. (2004). Social Capital and Information Technology. MIT Press.
Huysman, M. and Wulf V. (2006). IT to Support Knowledge Sharing in Communities, Towards a Social Capital Analysis. Journal of Information Technology, 21(1), 40-51.
Jiang, H. and Carroll, J.M. (2009). Social Capital, Social Network and Identity Bonds: A Reconceptualization. Proceeding of The 4th International Conference of Community and Technology, 51-60.
Jackman, R. and Miller. R.A. (1998). Social Capital and Politics. Annual Review of Political Science, 1, 47-73.
Kavanaugh, A. L. and Patterson S.J. (2001). The Impact of Community Computer Networks on Social Capital and Community Involvement. American Behavioral Scientist, 45(3), 496-496.
Kraut, R., Kiesler, S., Boneva, B., Cummings, J., Helgeson, V., and Crawford, A (2002). Internet Paradox Revisited. Journal of Social Issues, 58(1), 49-74.
Kraut R; Patterson M; Lundmark V; Kiesler S; Mukopadhyay T and Scherlis W. (1998). Internet Paradox: A Social Technology That Reduces Social Involvement and Psychological Well-Being? The American Psychologist, 53(9), 1017-31.
Lee, F. S. L., Vogel, D., and Limayem, M. (2003). Virtual Community Informatics: A Review and Research Agenda. Journal of Information Technology Theory and Application, 5(1), 47-61.
Leonard, R. and Onyx J. (2003). Networking Through Loose and Strong Ties: An Australian Qualitative Study. International Journal of Voluntary and Nonprofit Organizations, 14(2), 189-203.
Levine, M., Prosser, A., Evans, D. and Reicher, S. (2005). Identity and Emergency Intervention: How Social Group Membership and Inclusiveness of Group Boundaries Shape Helping Behavior. Personality and Social Psychology Bulletin, 31(4), 443-453.
Liebrand, W. B. G., Wilke, H. A. M., Vogel, R., and Wolters, F. J. M. (1986). Value Orientation and Conformity in Three Types of Social Dilemma Games. Journal of Conflict Resolution, 30, 77-97.
McCubbins, M.D., Paturi, R. and Weller, N. (2009). Connected Coordination: Network Structure and Group Coordination. American Politics Research, 37, 899-920.
Nahapiet, J. and Ghoshal S. (1998). Social Capital, Intellectual Capital, and the Organizational Advantage. The Academy of Management Review, 23(2), 242-266.
Onyx, J. and Bullen P. (2000). Measuring Social Capital in Five Communities. Journal of Applied Behavioral Science, 36(1), 23-42.
Paldam, M.(2000). Social Capital: One or Many? Definition and Measurement. Journal of Economic Surveys,14(5), 629-653.
Portes, Alejandro. (1998). 'Social Capital: Its Origins and Applications in Modern Sociology', Annual Review of Sociology 22: 1-24.
Putnam, R. D. (1995). Bowling Alone: America's Declining Social Capital. Journal of Democracy, 6(1), 65-78.
Putnam, R. D. (1995). Tuning In, Tuning Out: The Strange Disappearance of Social Capital in America. Political Science and Politics, 28(4), 664-683.
Putnam, Robert (2001). Social Capital. Measurement and Consequences. Canadian Journal of Policy Research, Spring, 41-51.
Quan-Haase, A., Wellman, B. (2004). How Does the Internet Affect Social Capital. Social Capital and Information Technology, 113-135.
Rand, D.G., Dreber, A., Ellingsen, T., Fudenberg, D. and Nowak, M.A. (2009). Positive Interactions Promote Public Cooperation. Science, 325(5945), 1272-1275.
Ren, Y., Kraut, R., Kiesler, S., (2007). Applying Common Identity and Bond Theory to Design of Online Communities. Organization Studies, 28(3), 377-408.
Rothstein, B., Eek, D. (2006). Political Corruption and Social Trust - An Experimental Approach. Presented at the American Political Science Association Meeting, Philadelphia, August 31-September 3.
Scott, J., (1988). Social Network Analysis. Sociology, 22(1), 109-109.
Tajfel H. (1970). Experiments in Intergroup Discrimination. Scientific American, 223, 96-102.
Tajfel, H., Billig, M.G., Bundy, R.P., and Flament, C. (1971). Social Categorization and Intergroup Behavior. European Journal of Social Psychology, 1(2), 149-178.
Tajfel, H. (1982). Social Psychology of Intergroup Relations, Annual Reviews in Psychology, 33(1), 1-39.
Van Dijk, F. Sonnemans J. and van Winden, F. (2002). Social Ties in a Public Good Experiment. Journal of Public Economics, 85, 275-299.
Wasko and Faraj, S. (2005). Why Should I Share? Examining Social Capital and Knowledge Contribution in Electronic Networks OF Practice. MIS Quarterly, 29(1), 35-57.
Wellman, B., A. Q. Haase, Witte, J., Hampton, K. (2001). Does the Internet Increase, Decrease, or Supplement Social Capital? Social Networks, Participation, and Community Commitment. American Behavioral Scientist, 45(3), 436-455.
Williams, K. (2005). Social Networks, Social Capital, and the Use of Information and Communications Technology in Socially Excluded Communities: A Study Of Community Groups in Manchester, England. Unpublished doctoral dissertation, University of Michigan. Retrieved from http://hdl.handle.net/2027.42/39370
Williams, K., Durrance, J.C. (2008) Social Networks and Social Capital: Rethinking Theory in Community Informatics. The Journal of Community Informatics, 4(3).