Is there social capital in service exchange tools?: Investigating timebanking use and social capital development

Computers in Human Behavior 81 (2018) 274e281

Contents lists available at ScienceDirect

Computers in Human Behavior journal homepage: www.elsevier.com/locate/comphumbeh

Full length article

Is there social capital in service exchange tools?: Investigating timebanking use and social capital development Chien Wen (Tina) Yuan a, *, Benjamin V. Hanrahan b, John M. Carroll b a b

Department of Advertising and Public Relations, Fu Jen University, 510, Zhongzheng Rd., Xinzhuang Dist., New Taipei City, 24205, Taiwan College of Information Sciences and Technology, Pennsylvania State University, 332, State College, P.A., 16801, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Available online 22 December 2017

Timebanking is a peer-to-peer service exchange tool in which the services exchanged are valued by the time it takes to provide them. This study complements previous work by empirically examining if timebanking is positively related to social capital development. Grounded in the theoretical model of social capital, the study incorporated self-efficacy and timebanking activities (requests and offers) as predictor variables, trust and reciprocity as dimension variables, and sense of community as outcome variable of social capital. Using data from a survey distributed across the timebanks nationwide (N ¼ 429), our findings provided evidence of positive relationships among self-efficacy, requests on timebanks, trust, and sense of community. We conclude that timebanking use is a promising way to develop social capital. Our study contributes to the understanding of social capital development on different mediated platforms. Unlike social network site use, it does not require members to personally construct the networks to reap associated social benefits. Timebanks help connect people, pool resources from the community, and pave the way for trust and reciprocity among members. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Timebanking Service exchange tools Social capital Social networks Coproduction Community engagement

1. Introduction While people might be familiar with the commercial business models of Uber or Airbnb, timebanks are a unique type of nonprofit peer-to-peer system grounded in a similar logic of service exchange. One key difference, however, is that instead of using money, timebanks use time credits for exchanges (Bellotti et al., 2014; Cahn, 2000). This study examined hOurworld.org, which is one of the largest timebank platforms in the U.S. and operates based on the value of coproduction among members in the community. Timebanks that feature coproduction recognize every member's contribution with equal value using time credits and foster active participation in bettering the exchange outcome (Cahn, 2000; Ostrom, 1996). With time credits, users make exchanges of otherwise idle resources, such as services or skills. For example, Joe can help pick up a prescription for another member on his way to a local drug store and earn time credits; then Joe may spend the time credits on

* Corresponding author. E-mail addresses: [email protected] (C.W. (Tina) Yuan), [email protected]. edu (B.V. Hanrahan), [email protected] (J.M. Carroll). https://doi.org/10.1016/j.chb.2017.12.029 0747-5632/© 2017 Elsevier Ltd. All rights reserved.

requesting a piano lesson from another member in the community. Each service is posted and reacted to on the timebanking system; each contribution is treated equally, and its value depends solely on how much time is spent on the service. The underlying principle of timebanking is generalized reciprocity, where the exchange does not have to be mutual but rather, members pay it forward to whoever needs it in the community (Whitham & Clarke, 2016). Grounded in generalized reciprocity, timebanks facilitate the creation of social networks in local communities with each member committing to both providing and requesting services despite the fact that they may not know one another. Given the logic of equal contribution and the encouragement of social interaction, timebanks are proposed to strengthen social connections and community participation (Ostrom, 1996; Ozanne, 2010; Seyfang, 2003). An engaged and sustainable community is a good source of social capital (Putnam, 2000), and the principle of generalized reciprocity that connects and strengthens community networks paves for the development of social capital for members to reap its benefits (Whitham & Clarke, 2016). Due to its premises and potential, timebanking has gained popularity since its introduction. Timebanks are active in 32 countries worldwide and there are around 500 timebanks across

C.W. (Tina) Yuan et al. / Computers in Human Behavior 81 (2018) 274e281

600 communities in different states in the U.S.1. Much previous research discussed the social capital benefits of timebanking from a theoretical and analytic angle (e.g., Cahn, 2000; Ozanne, 2010; Seyfang, 2004; Whitham & Clarke, 2016). Given its popularity and potential, it is important to empirically examine how social capital is developed through timebanking practice. A large body of research investigates the use of social network sites like Facebook for social capital development because they allow their users to keep in touch with old friends and maintain relationships with acquaintances; studying social capital embedded in people's known networks on social media, be it of close friends or of acquaintances, yields valuable theoretical and technological implications in the existing literature (e.g., Burke, Kraut, & Marlow, 2011; Ellison, Steinfield, & Lampe, 2011; Valenzuela, Park, & Kee, 2009). A study on how social capital can be potentially cultivated on distributed and less known networks on service exchange platforms like timebanks is valuable. It may suggest potential access to a wide variety of resources without participants scrupulously building personal networks to access them. It carries promising implications for the underprivileged, the young, and the old because they may face the issue of relatively homogeneous and small social networks, and have limited access to the types of resources that they may need (Collom, 2008; Lasker et al., 2011; M. B.; Marks, 2012). Given the significance, we propose to study timebanking from a social capital perspective to flesh out the benefits of timebanking participation. Previous studies predominantly theorized the potential of timebanking use for social capital and its associated outcomes. According to Seyfang (2003), people join timebanking for several motivations, such as volunteering, offering informal support to one another in the community, interacting with one another, and earning time credits. The outcomes include implementation and enhancement of public safety (Ostrom, 1996), social policy (Glynos & Speed, 2012), elderly healthcare (Lasker et al., 2011), youth transition support (M. B. Marks, 2012), environmental conservation (Seyfang & Smith, 2002), etc. Less work investigated social capital using quantitative empirical data. So far, only Collom (2008) used timebanking transaction data and social network analysis to study social capital for the older adults based on a stand-alone timebank. Our study complements previous ones on the following regards: 1) we focus on a more general timebanking user pool from timebanks nationwide with diverse demographic backgrounds; 2) an empirical investigation of timebanking use and social capital development is established through survey data; and 3) we examine the impacts of technological mediation on social capital by including different timebanking platforms, such as websites and mobile applications. Our framework draws on social capital perspective that includes predictors, dimensions, and outcome of social capital to provide a comprehensive way of investigating social capital development on timebanks (Narayan & Cassidy, 2001). This study contributes to the theoretical understanding of social capital development among distributed, less known networks mediated by timebanking, especially for underrepresentative populations. In addition, the study carries practical implications. The majority of timebanking services are mediated through web platforms. Previous work that investigated different timebanking platforms suggests that mobile applications may better support timebanking participation and engagement because 1) smart phone adoption rate is high, 2) mobile applications are highly integrated in people's

1 http://abcnews.go.com/blogs/headlines/2014/01/saving-money-helpingothers-with-timebanking/.

275

daily lives that could alleviate temporal and spatial barriers, and 3) mobile applications increases access and usage with higher mobility, immediacy, and social presence (Han, Shih, Bellotti, & Carroll, 2015; Han, Shih, Rosson, & Carroll, 2014). In this study, the authors want to explore whether users of different timebanking platforms, including its web version and mobile application, engage in timebanking in a way that contributes to different social capital development. RQ: How do mobile timebanking application users differ from web-timebanking users in social capital development? 1.1. Defining social capital Social capital is a multidimensional and multilevel construct. It is the sum of tangible and intangible resources derived from people's social connections in their network, which encompasses: 1) network structure; 2) the relationship people have with others in the network, which allows them to access resources they wish to use; and 3) the actual resources in quest/obtained, such as access to novel information, mobilization for collective action, or tangible goods (Adler & Kwon, 2002; Bourdieu, 2010; Portes, 1998). As an expansive, all-inclusive concept, social capital has different interpretations, appropriations, and operationalizations among scholars in different fields because each has its own academic focus (Adler & Kwon, 2002; Burt, 1997; Coleman, 1989; Lin, 1999). For example, the unit of analysis varies: scholars from sociology and economics highlight the micro level of social capital in terms of individual access or strategic position in the network for job opportunities (Burt, 2001; Granovetter, 1973). Other scholars from political science or education take social capital at a macro or collective level that explores how collective assets like social cohesion and community engagement are formed through trust among network members (Coleman, 1989; Putnam, 2000). In order to theoretically and practically study and operationalize social capital, researchers proposed a framework that includes predictors, dimensions, and outcome of social capital (See Fig. 1) (Narayan & Cassidy, 2001). In the model, predictors like empowerment and communication are factors that contribute to the development of social capital. We propose that self-efficacy can be considered as a form of empowerment for social capital predictor because the concept refers to individuals' beliefs in their capacity to deal with issues that happen to them (Bandura, 1977). The model also proposed that communication is the other determinant; given the practice of timebanking, we propose to use requests and offers as proxy for communication on timebanks, as people need to engage in communication during negotiating their exchanges. Next, the dimensions or the forms of social capital involve trust and reciprocity in the model. Last, the social outcome of social capital in timebanking use is the sense of community. We elaborate each factor in the model in detail and how we draw on Narayan and Cassidy's (2001) model in the following sections. 1.1.1. Social outcome of social capital: sense of community Sense of community reflects community members' attachment and commitment towards a community, by which members develop the common goal of fulfilling one another's needs (McMillan & Chavis, 1986). It is an outcome for social capital because it entails community participation, a necessary factor for utilization of a range of community assets (Chavis & Wandersman, 1990). Timebanking is a suitable platform for social capital development because it is a community-driven technology. The operation of timebanks is rooted in local community so that exchanges among members are possible. Therefore, the study used this variable as the outcome of social capital.

276

C.W. (Tina) Yuan et al. / Computers in Human Behavior 81 (2018) 274e281

1.1.2. Predictors of social capital: self-efficacy and timebanking activities In the model, empowerment is proposed as a predictor of social capital and operationalized as perceived impact one can have on oneself and the community (Narayan & Cassidy, 2001). In a similar vein, self-efficacy refers to the perception of control one has over the activities and contexts in which one is situated, along with the expectation of achievements one can reach (Bandura, 1977). Selfefficacy is positively related to the effort and ability one is willing to spend on overcoming the setbacks they encounter. This sense of control and motivation can be extended beyond individual level to community level because they are established as important factors in community psychology (Chavis & Wandersman, 1990). Members in the community feel they are able to deal with negative issues or stressors in the community through their active participation and engagement, which in turn builds a shared sense of community (Bandura, 1977; Chavis & Wandersman, 1990; Chavis, Lee, & Acosta, 2008). According to Cahn (2000), every member in timebanks is an asset with the capability to actively contribute to the community and meet another member's needs. Whatever the service type, a contribution is valued by the time spent on the task, redefining the value of work and empowering the members, especially those who are underprivileged or unemployed (M. B. Marks, 2012; Seyfang, 2004; Seyfang & Smith, 2002). People who provide services to others in ways the others need are valuable members (Collom, Kyriacou, & Lasker, 2012; Seyfang, 2003). Conventionally undervalued work, such as housework or domestic chores, weighs the same with work that is conventionally valued higher, like computer-related issues. As people can contribute however they can, such equality leads to a sense of self-efficacy (Lasker et al., 2011). Moreover, people have alternative ways to earn time credits and may use them to develop new skillsets for self-improvement (Lee, 2009). Taken together, timebanking is suggested to reduce social issues like unemployment and poverty (Seyfang, 2003). The associated outcomes are that people are not only active participants but empowered agents in carrying out timebanking exchanges. We hypothesize that through timebanking practices, people derive higher self-efficacy, which leads to higher sense of community and social capital. H1. Self-efficacy in timebanking practices is positively associated with sense of community. In Narayan and Cassidy (2001)’s model, communication emerged as the other significant predictor of social capital. Communication is positively associated with relationship development among people, which paves the way for social capital development (Burke et al., 2011; Ellison et al., 2011). On timebank platforms, the major activities are requesting and providing services for one another. Members negotiate exchanges after they respond to others' service requests using additional technologies like emails or through face-to-face interactions. Social interaction and communication are entailed by such exchanges. From the requests and offers on timebanks, members develop much awareness about resources and needs in the local community so that a sense of community can be strengthened (Carroll, 2014). Bellotti et al. (2014) suggest that differences exist in requests and offers in timebanking practices with more offers and fewer requests on most timebank platforms. People may feel vulnerable and in debt to others by posting their needs and therefore highlighting their deficiency. However, people who request services in timebanking are not passive recipients of assistance or charity; instead, by requesting services, they give others the opportunity to contribute, to learn, and to earn time credits to fulfill their future

needs (Cahn, 2000; Whitham & Clarke, 2016). In some purely coproduced cases like seeking a chess player or a tennis partner for a game, the two coproduce a desired outcome without anyone in debt to the other (Carroll, Chen, Yuan, & Hanrahan, 2016). Given the fact that these two timebanking activities may be perceived and practiced differently by members, the second set of hypotheses investigates which type of timebanking practice is positively associated with sense of community: H2. a) Requesting; b) providing services in timebanking is positively associated with sense of community. 1.1.3. Dimensions of social capital: trust and reciprocity The original model proposed a list of dimensions, or the forms, of social capital other than trust and reciprocity. Narayan and Cassidy (2001) pointed out that it should not be considered as an exhaustive list; rather it is important to include relevant dimensions to contextualize the types of social capital needed for each study. Given the focus of this study, we focus on trust and reciprocity, as they are related to timebanking practices and norms. Trust refers to the willingness and confidence in believing that community members will behave as expected (McMillan, 1996). Trust facilitates collaboration on common issues without requesting specific sanctions or rewards (Putnam, 2000). Trust is also associated with shared norms in the community because these norms provide clear guidance to the members as to how to properly engage in the community and provide assistance to one another, which facilitates shaping social capital in a community (Nahapiet & Ghoshal, 1998). Another dimension of social capital is reciprocity, which involves the expectation that members will offer mutual support when others are in need (Onyx & Bullen, 2000). Reciprocity helps consolidate shared norms and community development in that members act for both individual and collective benefits, supporting prosocial behaviors (Reno, Cialdini, & Kallgren, 1993). Timebanking practices are prosocial in the sense that they require interactions and commitment among members. Unlike direct reciprocity where the same two members help each other, the mechanism of timebanking is grounded in the pay-it-forward logic, or called generalized reciprocity, in which the recipient of a service may pay it back to another person in the community in a different way and at a different time (Cahn, 2000). Therefore, the generated benefits are not only shared by the ones who engage in the exchange per se but by the community as a whole (Cahn, 2000; Whitham & Clarke, 2016). Based on the sociology of helping and the social exchange principles, reciprocity is coproduced in general terms among members in the community, which sustains participation and reinforces the sense of community (Bellotti et al., 2014; Collom, 2008; Collom et al., 2012; Seyfang, 2003). With diverse requests and offers on timebanks, the potential pool of resources is accessible and available to the whole community. At the individual level, people do not necessarily have to rely on their immediate social networks for any type of support but may resort to this extended network for resources. At the collective level, trust and reciprocity generated by the connections and interactions among members may strengthen the sense of community, a form of social capital (Adler & Kwon, 2002; Lin, 1999; Putnam, 2000). With timebanking, the connected social networks in the community help build the social infrastructure of social capital. Timebanking usage could positively mediate less known networks and cultivate trust among members. Believing that others will honor their requests and offers as well as that they will pay the service forward may facilitate community building. In other words, trust, reciprocity, and timebanking use may have positive associations on forming the sense of community.

C.W. (Tina) Yuan et al. / Computers in Human Behavior 81 (2018) 274e281

H3. Trust in timebanking practices is positively associated with sense of community.

Table 1 Zero-order correlations.

H4. Reciprocity in timebanking practices is positively associated with sense of community.

1. 2. 3. 4. 5. 6.

2. Material and method 2.1. Study procedure and participants Data were collected from January to May of 2016. The survey was hosted by a University-affiliated web survey tool and included questions about participants’ use of timebanking, including their requests and offers, as well as timebanking-enabled social outcomes like self-efficacy, trust, reciprocity, and sense of community. Only respondents with mobile timebanking experience answered questions related to mobile application usability. The survey was conducted in English and took around 20 min to complete. The Institutional Review Board reviewed and approved the study because there was neither risk nor ethical concerns posed to the participants. Participants were recruited through one of the largest timebanking organizations, hOurworld, which has more than 10,000 members across over 654 timebanks in different communities in the U.S. Each hOurworld timebank is operated in a local community, which could match to municipal boundary like township or borough, because it aims at facilitating face-to-face interactions and exchanges among members in the community. The recruiting message was posted to the hOurworld portal and distributed from the portal to its affiliated local timebank websites. The respondents received 1 h of hOurworld time credit for completing the survey. Overall, we received a total of 520 survey responses (5.2% response rate); excluding incomplete responses, we had 429 valid entries. Among them, 84 respondents (19.58%) reported to use the mobile Timebanking application. The average respondent is female (78.6%), Caucasian (82.6%), and married or in a committed relationship (55.3%), whose mean age is 52.45 (SD ¼ 14.38), ranging from 20 to 88. Most respondents (79.2%) held a college degree or above. 2.2. Measures Most of the measures used in the survey were validated scales from previous studies. Scale items were modified to fit the context of this study. Unless otherwise described, all scales were measured on Likert scale of 1 (strongly disagree) to 5 (strongly agree). Zeroorder correlations of the variables can be found in Table 1. 2.2.1. Trust The original Interpersonal Trust Scale (Rotter, 1967) has 25 items and we drew on four to capture users’ trust in hOurworld members and their services (M ¼ 3.75, SD ¼ 0.64; a ¼ 0.88). Questions include “hOurworld members usually can be relied on to keep their promises; Most hOurworld members are honest in describing their offers; Most hOurworld members will commit to fulfilling the tasks, even if they think you are ignorant of their specialty; Most hOurworld members respond to requests honestly.”

277

Requests Offers Trust Self Efficacy Sense of Community Reciprocity

1

2

3

4

5

6

1 .698** .127* .009 .184** .202**

.698** 1 .045 .035 .157** .110*

.127* .045 1 .129* .436** .340**

.009 .035 .129* 1 .192** .025

.184** .157** .436** .192** 1 .158**

.202** .110* .340** .025 .158** 1

**

p < .01, *p < .05 (2-tailed).

Other hOurworld members and I are both providing the same amount of help and support to one another; Other hOurworld members are providing more help and support to me than I provide in return; Neither I nor other hOurworld members are providing help and support to each other.” We reverse-coded the first and the third item that did not indicate equal reciprocity among members and took the mean of all items (M ¼ 3.39, SD ¼ 0.53). 2.2.3. Self-efficacy Self-efficacy was measured using Schwarzer's (2014) scale on a 1 (very unlikely) to 7 scale (very likely), which aims at capturing a broad sense of individual competence to deal with various difficult situations. The original scale contains 10 questions and we drew on four of them, which factored into a single component (M ¼ 5.58, SD ¼ 0.94; a ¼ 0.82). Questions include “I can always manage to solve difficult problems if I try hard enough; If someone opposes me, I can find the ways and means to get what I want; I am confident that I could deal efficiently with unexpected events; Thanks to my resourcefulness, I can handle unforeseen situations.” 2.2.4. Sense of community According to McMillan and Chavis (1986), sense of community (SCI) encompasses four factors, including membership, influence (both member over the community and the community over members), integration and fulfillment of needs, and shared emotional connection. The original SCI has 24 items (Chavis et al., 2008), eight of which were used to measure the construct on a 1 (very unlikely) to 7 scale (very likely) (M ¼ 4.97, SD ¼ 1.03; a ¼ 0.87), including “Other hOurworld members and I value the same things; Being a member of hOurworld makes me feel good; I can trust people in hOurworld; Being a member of hOurworld is a part of my identity; Fitting into hOurworld is important to me; I care about what other hOurworld members think of me; It is very important to me to be a part of hOurworld community; I feel hopeful about the future of hOurworld community.” 2.2.5. System usability scale (SUS) We used this widely adopted scale to assess respondents’ feedback about the mobile application. SUS contains 10 items and a score above 68 is considered above average (Lewis & Sauro, 2009). A sample question is “I thought the application was easy to use.” Our results suggest that the respondents thought the mobile timebanking application has good usability (M ¼ 84.79, SD ¼ 20.35). 3. Results

2.2.2. Reciprocity The scale of Buunk, Doosje, Jans, and Hopstaken (1993) (a ¼ 0.92) was adapted to measure perceived reciprocity and social support one can provide to and receive from other hOurworld members. The items are “I am providing much more help and support to other hOurworld members than I receive in return;

3.1. General timebanking and mobile timebanking use Fifty-nine point seven percent of the respondents have been participating in timebanks for more than one year. In terms of receiving timebank services, 37.1% of our respondents reported

278

C.W. (Tina) Yuan et al. / Computers in Human Behavior 81 (2018) 274e281

receiving 1e5 timebank services within the last 12 months, 18.5% 6e10 services, 8.1% 11e20 services, 10.9% 21e100 services, and 3% more than 100 services. For offering timebank services, 31.6% of our respondents reported to offer 1e5 services within the last 12 months, 17.4% 6e10 services, 13.9% 11e20 services, 17.6% 21e100 services, and 4.3 percent more than 100 services. The average percentage of the total reported hours for requesting services is 61.14% (SD ¼ 66.90%), whereas for offering services is 70.61% (SD ¼ 63.39%). In terms of tools they used to access timebanks, our respondents reported to use their personal computers or laptops at least weekly. Smartphones were used to access the website or mobile application at least monthly. Tablets were rarely used to access either timebank website or mobile application. Comparing mobile-application and web-platform timebank users, the results indicate that mobile-application users both request (t(392) ¼ 2.882, p ¼ .004) and provide (t(372) ¼ 4.423, p < .00) more services than web-platform users. Among the mobile application users, according to paired-samples t-test, there is no difference in terms of number of services provided before and after using mobile timebanking application t(81) ¼ 0.245, p ¼ .807; nor in number of services requested t(79) ¼ 0.457, p ¼ .649. 3.2. Hypotheses testing To test our model, hierarchical (blocked) ordinary linear regression (OLS) was conducted to assess the effects of each block. The variables in the first block included control variables like age, education, gender (1 ¼ M; 2 ¼ F), and relationship status (1 ¼ single; 2 ¼ married or in a committed relationship). The second block had predictor variables of social capital, including self-efficacy and two forms of timebanking activities (requests and offers). The last block contained the measures of trust and reciprocity. Each model was tested for collinearity and the variance inflation factor (VIF) value ranged from 1.02 to 2.08, indicating that multicollinearity was not an issue (Allison, 2012). Each of our models account for reasonable portion of variance. Stepwise addition of each block significantly improved model fit. In the output of each block, we present the overall R2 for the regression including the block and if any proceeding block. We also have F value (F change) to show the significance of any increase in R2 resulting from adding that block to the previous model. Here we discuss the results of the OLS model (see Table 2). In the first block, gender shows a trend among all the control variables, males (b ¼ 0.25, p ¼ .07) tend to have more sense of community through timebank use. The total variance of the first block explained by the regression model was below 10%. The block of control variables had less explanatory power. In the second block, we found that making requests on timebanks is positively associated with sense of community (b ¼ 0.12, p ¼ .04) but providing services is not (b ¼ 0.03, n.s.). Self-efficacy is positively associated with sense of community (b ¼ 0.20, p ¼ .001). Adding the block of predictor variables significantly improved the regression model (R2 ¼ 0.09; F [3, 330] ¼ 8.23, p < .001). Last, the third block improved the overall predictability (R2 ¼ 0.24; F [2, 328] ¼ 32.43, p < .001), with only trust (b ¼ 0.64, p < .001) emerging as variable positively associated with sense of community. H1 dealt with self-efficacy as a result of timebanking use and sense of community and was supported. H2a and H2b stated that requesting and offering services in timebanking is positively associated with sense of community. Based on the analysis, our second set of hypotheses was partially supported, as only requesting services on timebanks was positively associated with sense of

Table 2 OLS regression predicting sense of community. Model

1

2

3

Control Variables Age Gender Education Relationship

.00 (.00) -.25þ (.13) -.05 (.05) .06 (.06)

.00 (.00) -.25þ (.13) -.04 (.04) .03 (.06)

.00 (.00) -.16 (.12) -.03 (.04) .01 (.06)

.20**** (.06) .12* (.06) .03 (.05)

.15* (.05) .06 (.05) .08 (.05)

Predictor Variables Self-efficacy Requests Offers Dimension Variables Trust Reciprocity R2 F full model df full model R2 change F change df change

.64**** (.08) -.00 (.10) .02 1.43 4, 333

.09 4.40**** 7, 330 .07 8.23**** 3, 330

.24 11.28**** 9, 328 .15 32.43**** 2, 328

Notes: Regression coefficients are unstandardized, controlling for all other variables. Standard errors in parentheses. R2 change refers to the unique contribution of each block of variables controlling for the previous variables entered in the regression. Statistical significance is derived from two-tailed t tests. ****p < .00, ***p < .005, ** p < .01, *p < .05, þp < .10.

community. Last, H3 and H4 examined trust (H3) and reciprocity (H4) resulting from timebanking practices and their association with sense of community. H3 was supported. The descriptive analysis showed that mobile-application users requested and provided more services than web-platform users. We further compared if these two groups had different levels of trust, selfefficacy, and sense of community but found no significant results. 4. Discussion The goal of this study was to investigate whether service exchange tool use, specifically with the case of non-profit timebanking use, was related to users’ social capital development. This study has several contributions. First, it is among the first that empirically investigates timebanking use and social capital development using a general user pool. Second, this is also the first study to investigate whether different timebanking platforms (webplatform vs. mobile application) mediate social capital development distinctively. Third, a theoretical model analyzing social capital by its predictor variables, dimension variables, and outcome variable was employed (Narayan & Cassidy, 2001). Using survey data collected across U.S. timebanks, we hypothesized that selfefficacy, requests, and offers as predictor variables, trust and reciprocity as dimension variables would be positively associated with sense of community, the outcome variable of social capital. After taking into account of several demographic variables, our results show that among the predictor variables, self-efficacy and requests were positively associated with sense of community; among dimension variables, trust emerged to have positive association with sense of community. 4.1. Control variables The associations between timebanking use and social capital variables were not moderated by education, relationship status, or age. Only gender moderated the association, with male users having stronger sense of community, a measure of social capital, by

C.W. (Tina) Yuan et al. / Computers in Human Behavior 81 (2018) 274e281

using timebanks. Previous literature about timebanking participation has reported that females were much more likely to join than males (e.g., Collom, 2008; Lasker et al., 2011; Shih, Bellotti, Han, & Carroll, 2015). It is also suggested that timebanking mechanism works in favor of females because it recognizes informal or nonmarket labor to which females are often the major contributors (Callison, 2003; Glynos & Speed, 2012; Piscicelli, Cooper, & Fisher, 2015). Based on our results, we suggest that a distinction between membership and actual usage may be important in predicting social capital development. While findings about membership may suggest that females are more likely to acquire social resources from timebanks, actual usage and engagement are better predictors of social capital. For example, scholars (Collom, 2008; Lasker et al., 2011) pointed out that although females consisted of the majority of the local timebank members in their studies, males had higher reciprocation rate as well as more transactions, leading to more social capital development. In order to promote overall timebanking use and social capital development, it may be useful to think about gender-specific technological support to boost membership to actual use in female users. Alternatively, it may be helpful to incorporate other user characteristics, such as motivations or personality traits to further unpack gender and individual differences in timebank use and social capital development. 4.2. Predictor variables of social capital For the predictor variables of social capital development, the results confirm that self-efficacy is positively associated with sense of community. The premise of timebanking on recognizing all community members as valuable contributors strengthens the beliefs that members have for themselves and the community as a whole. Such beliefs are positively associated with developing individual and community resources. While previous studies showed that the concept of self-efficacy predicts learning outcomes (Zimmerman, 2000), performance and behavioral change (Bandura, 1977), or health improvement (R. Marks & Allegrante, 2005), the present findings complement previous literature by suggesting that high self-efficacy is a good predictor variable for social capital development too. As for the other two predictor variables, making request, in contrast to making offers, is the specific timebanking activity that contributes to social capital. Research shows that it is more likely for timebank members to offer services than request ones due to the tendency of altruism and the avoidance of being in debt (Bellotti et al., 2014; Shih et al., 2015). The metaphors of “banking” and “keeping balance of time credits” reinforce this imbalance. However, our results show that it is making requests that contributes to social capital development. Given the fact that participating in timebanking and other service exchange tools requires both initiators and joiners to complete the exchanges and/or the interactions (Carroll & Bellotti, 2015; Carroll et al., 2016), it is worth for timebank platforms to leverage designs that encourage users to take the initiative of requesting services on timebanks. If making requests seems more positive and favorable via designs, members may actively invite others to join the exchanges, which may increase social interactions, enhance generalized reciprocity, and potentially lead to greater social capital development. The results also confirm that it is necessary to separate these two distinct activities on timebanks and other service exchange tools because the benefits associated with these two activities may differ. Technological affordances of different timebank platforms should leverage their interface designs by facilitating users to make requests comfortably without casting the impressions of making impositions on others or feeling vulnerable for themselves. Carroll

279

et al. (2016)’s design proposes the idea of equal coproduction/ contribution is a case in point. 4.3. Dimension variables of social capital Trust is usually cultivated from close networks consisting of family members or close friends because it requires familiarity, bonds of friendship, and common faith and values (Coleman, 1989; Granovetter, 1973; Lin, 1999; Putnam, 2000). Our results suggest that it does not require a close network or even known networks on timebanks to develop trust among members and for them to reap associated benefits. This is promising especially for those who are underprivileged or have smaller network because participating in timebanking allows them to acquire resources they cannot easily access otherwise. It facilitates redistribution of social resources and paves for social justice (Ostrom, 1996; Seyfang & Smith, 2002; Seyfang, 2003). Another dimension associated with trust is reciprocity. Our results do not indicate a significant association between reciprocity and social capital. We surmise that it is due to the limitation of its operationalization. In theory, timebanking promotes generalized reciprocity where members pay received favors forward by offering others services. The lack of existing measurement on generalized reciprocity led us to use one for mutual reciprocity between two parties. While this decision had the benefits of securing the validity of the measurement, a scale that better fits the practice of generalized reciprocity in timebanking or other non-profit service exchange tools constitutes a good venue for future research. 4.4. Sense of community: an outcome of social capital We used sense of community as an outcome variable for social capital because timebanking is operated in local communities to encourage interactions and exchanges among community members. Our results confirm that this online timebanking participation could be transferred to offline social capital. Thus, this study supports that the affordances of timebanking, be it technological one or social one, can influence users’ social capital development. Likewise, our model also verifies the multidimensional aspect of social capital. This study uses timebanking as an example of a peer-topeer non-profit service exchange tool. As technologies grounded in the mechanism of service exchange or shared economy continue to burgeon, we hope our findings contribute to the larger understanding of social capital development in these tools. 4.5. Platform differences: websites vs. mobile applications Our research question taps into whether using different timebanking platforms results in different levels of social capital. Previous research points out that the affordances of mobile timebanking application, including reducing transaction time, supporting real-time coordination, and enhancing awareness of opportunities based on time and place, may contribute to more timebanking transactions and a higher sense of community (Han et al., 2015). Our results show that mobile timebanking users found the application to be of high usability and conducted more transactions than regular online-platform timebanking users but the level of sense of community did not differ between the two groups. The results also show that among the mobile users, their usage pattern did not change before and after using the application, suggesting that they were a group of active timebanking members irrespective of different technological mediations. Also, given the fact that mobile users only constituted 19.58% of our survey respondents, it may require a higher sample size to see the variances.

280

C.W. (Tina) Yuan et al. / Computers in Human Behavior 81 (2018) 274e281

4.6. Limitations and future directions We have to note that the relationship between timebanking use and social capital development was modest, as revealed by the percentage (24%) of variance of the control, predictor, and dimension variables explained by the regression models. The effect size, Cohen's f2, is .32, which suggests a decent magnitude in behavioral science. Several factors contribute to the result. First, social capital is a concept with multiple levels that combine relational and structural aspects, which can be difficult to quantify and measure in a cross-sectional survey because social capital requires time to develop (Streeten, 2002). A longitudinal study is helpful in unpacking social capital development. Secondly, factors associated with social capital may entail multiple aspects themselves; for example, reciprocity can refer to emotional support such as empathy or caring, instrumental support such as small favor and monetary support, or informational support such as advice or solutions to a problem (Ferlander, 2007). Also, we used a generalized self-efficacy scale in our study. According to Bandura (2006), the construct has better explanatory and predictive value when the measurements reflect situational demands. Future study should contextualize the needs in different populations and/or communities so as to precisely reflect community members' social capital development. Last, the adoption and use of timebanks may not be as widespread and frequent as other tools like Facebook. It could be just as alarming if a sole technological platform can predict social capital development with high variance, not to mention that there are other factors like people's socialization, political participation, civic engagement, or other technology use that may lead to social capital too (Adler & Kwon, 2002; Coleman, 1989; Ostrom, 1996; Putnam, 2000). There are several limitations in this study. First, the issue of sampling bias may influence the results in the sense that our respondents may belong to the group of relatively active timebanking users by attending to the recruitment notifications and completing the survey. Secondly, although our sample was collected from timebanks across the U.S., our data was cross-sectional and so cannot be used to make any causational statements about our findings. For example, it may well be that people with higher selfefficacy and trust are more likely to join timebanking. A longitudinal, contextualized study of social capital development on timebanking can offer valuable insight. Last, there are different types of timebanks with purposes like improving local economy, skill learning, or civic engagement; future study should draw attention to examine social capital development among these service exchange tools. 5. Conclusion The major contribution of our study is to show that timebanking use is positively associated with social capital development. To address the complexity of social capital as a theoretical concept, we integrate multiple levels to its operationalization, using Narayan and Cassidy's (2001) model, including its predictors, different dimensions, and the outcome. Our results suggest that in order to reap benefits from timebanking use, actual engagement, specifically posting requests, is helpful instead of mainly keeping a profile or offering services. Users' self-efficacy and trust are also positively associated with social capital. The study shows that online timebank participation can lead to offline social resources, suggesting a promising use of timebanking. These implications can be helpful for the underprivileged, the young, or the older generation. We did not find differences in terms of technological platforms for mediating social capital development on timebanks.

Appendix

Fig. 1. The original model of Narayan and Cassidy (2001, p. 90).

References Adler, P. S., & Kwon, S.-W. (2002). Social capital: Prospects for a new concept. Academy of Management Review, 27(1), 17e40. Allison, P. D. (2012). When can you safely ignore multicollinearity. Statistical Horizons, 5(1). Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191e215. https://doi.org/10.1037/0033-295X.84.2.191. Bandura, A. (2006). Guide for constructing self-efficacy scales (pp.307-337). In F. Pajares, & T. C. Urdan (Eds.), Self-efficacy beliefs of adolescents. Information Age Publishing Inc. Bellotti, V., Cambridge, S., Hoy, K., Shih, P. C., Handalian, L. R., Han, K., et al. (2014). Towards community-centered support for peer-to-peer service exchange: Rethinking the timebanking metaphor. In Proceedings of the SIGCHI conference on Human factors in computing systems (pp. 2975e2984). https://doi.org/ 10.1145/2556288.2557061. Bourdieu, P. (2010). The forms of capital (1986). In I. Szeman, & T. Kaposy (Eds.), Cultural theory: An anthology (pp. 81e93). John Wiley & Sons. Burke, M., Kraut, R. E., & Marlow, C. (2011). Social capital on Facebook: Differentiating uses and users. In Proceedings of the SIGCHI conference on Human factors in computing systems (pp. 571e580). https://doi.org/10.1145/1978942.1979023. Burt, R. S. (1997). A note on social capital and network content. Social Networks, 19(4), 355e373. Burt, R. S. (2001). Structural holes versus network closure as social capital. Social Capital: Theory and Research, 31e56. Buunk, B. P., Doosje, B. J., Jans, L. G. J. M., & Hopstaken, L. E. M. (1993). Perceived reciprocity, social support, and stress at work: The role of exchange and communal orientation. Journal of Personality and Social Psychology, 65(4), 801. https://doi.org/10.1037/0022-3514.65.4.801. Cahn, E. S. (2000). No more throw-away people: The co-production imperative. Washington DC: Essential Books. Callison, S. (2003). Ntial Books.-production imperativeaken, L. E. M. (1993). Perceived reciprocity, social supporLocal Economy, 18(3), 264e267. https:// doi.org/10.1080/0269094032000111048d. Carroll, J. M. (2014). The neighborhood in the Internet: Design research projects in community informatics. New York: Routledge. Carroll, J. M., & Bellotti, V. (2015). Creating value together: The emerging design space of peer-to-peer currency and exchange. In Proceedings of the 18th ACM conference on computer supported cooperative work & social computing (pp. 1500e1510). https://doi.org/10.1145/2675133.2675270. Carroll, J. M., Chen, J., Yuan, C. W., & Hanrahan, B. V. (2016). In search of coproduction: Smart services as purely reciprocal. IEEE Transactions on Computer, 49(7), 26e32. Chavis, D. M., Lee, K. S., & Acosta, J. D. (2008). The sense of community (SCI) revised: The reliability and validity of the SCI-2. In Paper presented at the 2nd international community psychology conference. Lisboa, Portugal. Chavis, D. M., & Wandersman, A. (1990). Sense of community in the urban environment: A catalyst for participation and community development. American Journal of Community Psychology, 18(1), 55e81. https://doi.org/10.1007/ BF00922689. Coleman, J. S. (1989). Social capital in the creation of human capital. University of Chicago Press. Collom, E. (2008). Engagement of the elderly in time banking: The potential for

C.W. (Tina) Yuan et al. / Computers in Human Behavior 81 (2018) 274e281 social capital generation in an aging society. Journal of Aging & Social Policy, 20(4), 414e436. Collom, E., Kyriacou, C., & Lasker, J. N. (2012). Equal time, equal value: Community currencies and time banking in the U.S. Surrey. England: Ashgate Publishing. Ellison, N., Steinfield, C., & Lampe, C. (2011). Connection strategies: Social capital implications of Facebook-enabled communication practices. New Media & Society, 13(6), 873e892. Ferlander, S. (2007). The importance of different forms of social capital for health. Acta Sociologica, 50(2), 115e128. https://doi.org/10.1177/0001699307077654. Glynos, J., & Speed, E. (2012). Varieties of co-production in public services: Time banks in a UK health policy context. Critical Policy Studies, 6(4), 402e433. https://doi.org/10.1080/19460171.2012.730760. Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78(6), 1360e1380. Han, K., Shih, P. C., Bellotti, V., & Carroll, J. M. (2015). It's time there was an app for that too: A usability study of mobile timebanking. International Journal of Mobile Human Computer Interaction, 7(2), 1e22. https://doi.org/10.4018/ijmhci.2015040101. Han, K., Shih, P. C., Rosson, M. B., & Carroll, J. M. (2014). Understanding local community attachment, engagement and social support networks mediated by mobile technology. Interacting with Computers. https://doi.org/10.1093/iwc/ iwu040. Lasker, J., Collom, E., Bealer, T., Niclaus, E., Keefe, J. Y., Kratzer, Z., & Schulman, J. (2011). Time banking and health: The role of a community currency organization in enhancing well-being. Health Promotion Practice, 12(1), 102e115. https://doi.org/10.1177/1524839909353022. Lee, L. (2009). Valuing the time, leading the future: A case study of the time banking program of an educational institution in Taiwan. International NGO Journal, 4(6), 324e328. Lewis, J. R., & Sauro, J. (2009). The factor structure of the system usability scale. Human Centered Design, 84e103. Lin, N. (1999). Building a network theory of social capital. Connections, 22(1), 28e51. Marks, M. B. (2012). Time banking service exchange systems: A review of the research and policy and practice implications in support of youth in transition. Children and Youth Services Review, 34(7), 1230e1236. https://doi.org/10.1016/ j.childyouth.2012.02.017. Marks, R., & Allegrante, J. P. (2005). A review and synthesis of research evidence for self-efficacy-enhancing interventions for reducing chronic disability: Implications for health education practice (part II). Health Promotion Practice, 6(2), 148e156. https://doi.org/10.1177/1524839904266792. McMillan, D. W. (1996). Sense of community. Journal of Community Psychology, 24(4), 315e325. McMillan, D. W., & Chavis, D. M. (1986). Sense of community: A definition and theory. Journal of Community Psychology, 14(1), 6e23. Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organizational advantage. Academy of Management Review, 23(2), 242e266. https:// doi.org/10.5465/AMR.1998.533225. Narayan, D., & Cassidy, M. F. (2001). A dimensional approach to measuring social

281

capital: Development and validation of a social capital inventory. Current Sociology, 49(2), 59e102. https://doi.org/10.1177/0011392101049002006. Onyx, J., & Bullen, P. (2000). Measuring social capital in five communities. The Journal of Applied Behavioral Science, 36(1), 23e42. https://doi.org/10.1177/ 0021886300361002. Ostrom, E. (1996). Crossing the great divide: Coproduction, synergy, and development. World Development, 24(6), 1073e1087. Ozanne, L. K. (2010). Learning to exchange time: Benefits and obstacles to time banking. International Journal Of Community Currency Research, 14, A1eA16. Piscicelli, L., Cooper, T., & Fisher, T. (2015). The role of values in collaborative consumption: Insights from a product-service system for lending and borrowing in the UK. Journal of Cleaner Production, 97, 21e29. https://doi.org/10.1002/ asi.23552. Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24(1), 1e24. https://doi.org/10.1146/annurev.soc.24.1.1. Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. Simon and Schuster. Reno, R. R., Cialdini, R. B., & Kallgren, C. A. (1993). The transsituational influence of social norms. Journal of Personality and Social Psychology, 64(1), 104e112. https://doi.org/10.1037/0022-3514.64.1.104. Rotter, J. B. (1967). A new scale for the measurement of interpersonal trust1. Journal of Personality, 35(4), 651e665. https://doi.org/10.1111/j.1467-6494.1967.tb01454.x. Schwarzer, R. (2014). Self-efficacy: Thought control of action. Taylor & Francis. Seyfang, G. (2003). ‘With a little help from my friends:’ Evaluating time banks as a tool for community self-help. Local Economy, 18(3), 257e264. https://doi.org/ 10.1080/0269094032000111048c. Seyfang, G. (2004). Time banks: Rewarding community self-help in the inner city? Community Development Journal, 39(1), 62e71. https://doi.org/10.1093/cdj/ 39.1.62. Seyfang, G., & Smith, K. A. (2002). The time of our lives: Using time banking for neighbourhood renewal and community capacity building. London: New Economics Foundation. Shih, P. C., Bellotti, V., Han, K., & Carroll, J. M. (2015). Unequal time for unequal value: Implications of differing motivations for participation in timebanking. In Proceedings of the 33rd annual ACM conference on Human factors in computing systems (pp. 1075e1084). https://doi.org/10.1145/2702123.2702560. Streeten, P. (2002). Reflections on social and antisocial capital. Journal of Human Development, 3(1), 7e22. https://doi.org/10.1080/14649880120105362. Valenzuela, S., Park, N., & Kee, K. F. (2009). Is there social capital in a social network site?: Facebook use and college students' life satisfaction, trust, and participation1. Journal of Computer-mediated Communication, 14(4), 875e901. https:// doi.org/10.1111/j.1083-6101.2009.01474.x. Whitham, M. M., & Clarke, H. (2016). Getting is giving: Time banking as formalized generalized exchange. Sociology Compass, 10(1), 87e97. https://doi.org/10.1111/ soc4.12343. Zimmerman, B. J. (2000). Self-efficacy: An essential motive to learn. Contemporary Educational Psychology, 25(1), 82e91. https://doi.org/10.1006/ceps.1999.1016.