a social network approach to understanding ... - Wiley Online Library

A R T I C L E

A SOCIAL NETWORK APPROACH TO UNDERSTANDING COMMUNITY PARTNERSHIPS IN A NONTRADITIONAL DESTINATION FOR LATINOS Brian A. Eiler Center for Cognition, Action and Perception and University of Cincinnati

Daniele A. Bologna University of Cincinnati

Lisa M. Vaughn Cincinnati Children’s Hospital Medical Center and University of Cincinnati College of Education

Farrah Jacquez University of Cincinnati

Cincinnati, like other new migration areas, has recently experienced tremendous growth in the Latino immigrant population. Because greater health disparities exist for Latinos compared to both majority and other minority groups, it is essential to understand how migratory patterns and healthcare infrastructure are related. In this study, social network analysis (SNA), which quantitatively assesses and evaluates network formation and network relationships, was used to investigate the structure of the Greater Cincinnati Latino health network. Referral and collaboration networks were assessed for 29 individuals serving the Latino community. Results indicated the desired collaboration network was nearly twice as dense as either the physical or the mental health referral networks. The physical network was also denser than the mental health network. Similar results were found when analyzing network centralization. Taken together, Research reported in this article was supported by the United Way of Greater Cincinnati. We also thank our participants and research team, as well as the Editor-in-Chief and an anonymous reviewer. Please address correspondence to: Brian A. Eiler, Department of Psychology, ML 0376, 5145 Edwards C1, University of Cincinnati, Cincinnati, OH 45221-0376. E-mail: [email protected] JOURNAL OF COMMUNITY PSYCHOLOGY, Vol. 45, No. 2, 178–192 (2017) Published online in Wiley Online Library (wileyonlinelibrary.com/journal/jcop). C 2017 Wiley Periodicals, Inc. DOI: 10.1002/jcop.21841

Social Network Approach to Understanding Community Partnerships

r

179

results indicate a need for additional strategic partnerships between Latino-serving providers and the Latino-serving community. Specific C 2017 Wiley Periodicals, Inc. recommendations are briefly discussed.

INTRODUCTION Cincinnati has experienced tremendous growth in the Latino immigrant population. In fact, there was a 63% increase in the Latino population across Ohio between 2000 and 2010 (U.S. Census Bureau, 2012). Because of significant health disparities for Latinos compared to both majority and other minority groups, it is essential to understand how migratory patterns and healthcare infrastructure are related, particularly in nontraditional migration destinations like Cincinnati (Jacquez, Vaughn, Pelley, & Topmiller, 2015). Furthermore, healthcare disparities are partially attributable to differences in healthcare access, lack of health insurance, and quality of care (e.g., Carrasquillo, Carrasquillo, & Shea, 2000; Gresenz, Derose, Ruder, & Escarce, 2012). Healthcare infrastructure research allows both providers and community partners to work together to develop specific interventions and policies to address local contextual issues and service gaps that contribute to disparities for Latinos. As such, quantifying the structure that scaffolds healthcare experiences for Latinos in nontraditional areas can document how changes in community institutions and agencies may reduce negative health experiences facing Latinos. Social network analysis (SNA) is a quantitative approach that investigates network formation and relationship structure between those that comprise the network (for review, see Hanneman & Riddle, 2005). SNA provides a visual illustration of the structure and patterns underlying social interactions by placing actors (individuals or organizations) on a graph, with connections between them representing relationships of interest (Hanneman & Riddle, 2005). Over the last several decades, interest in SNA has developed in several fields including organizational and community psychology (e.g., Granovetter, 1985; Sparrowe, Liden, Wayne, & Kraimer, 2001). SNA has been employed to understand diverse phenomena including sexually transmitted infections (De Rubeis, Wylie, Cameron, Nair, & Jolly, 2007), drug use (Young, Halgin, DiClemente, Sterk, & Havens, 2014), happiness (Fowler & Christakis, 2008), innovation (Gloor, Paasivaara, Schoder, & Willems 2007), organizational structure (Cross, Kase, Kilduff, & King, 2013), and smoking cessation (Mercken, Snijders, Steglich, Vertianen, & de Vries, 2010). This attention has highlighted SNA’s interactionist framework, whereby explanatory power and analysis are placed in relationships, rather than indi¨ uy ¨ uk, ¨ & Milanov, 2010). Using the structure of a social vidual attributes (Zaheer, Gözub network to identify relevant components (nodes) and investigate how specific types of interactions (ties) between the nodes relate to outcome variables can provide insight into how interacting levels of organization relate to community specific contexts such as healthcare. Healthcare system research using SNA has steadily gained acceptance (e.g., Lee et al., 2011). For example, SNA has been employed to both explore community partnershipbuilding processes (Lasker, Weiss, & Miller, 2001) and evaluate coalition building (Luque et al., 2010). Overall, social networks facilitate successful public healthcare planning and implementation (Harris & Clements, 2007). However, little research is available that has investigated social networks in contexts in which healthcare infrastructure is lacking for

Journal of Community Psychology DOI: 10.1002/jcop

180

r

Journal of Community Psychology, March 2017

a specific population, such as Latinos in the nontraditional migratory destination of Cincinnati, Ohio. Cincinnati as a Nontraditional Destination for Latinos In the United States, the Hispanic population has grown at a rate of 57.9% between 1990 to 2000, and in the Midwestern states, which includes Ohio, this growth has exceeded 80% (Guzmán & McConnell, 2002). In Hamilton County, Ohio (i.e. Cincinnati), between 2000 and 2011, growth was beyond 125%, and the Latino population in the two counties immediately North of Cincinnati have grown over 215% (Mora et al., 2014). Previous research has classified traditional destinations for Latinos as ones in which Latinos comprise 25% or more of the population (Gresenz et al., 2012). New or nontraditional destinations for Latinos have been classified as ones in which Latinos comprise less than 20% of the total initial population, and have subsequently experienced growth beyond 2.5% in the subsequent decade (Gresenz et al, 2012). Cincinnati’s Latino population was less than 20% in 2000 and has experienced growth rates well above the 2.5% threshold from 2000 to 2010 (for discussion and growth map, see Jazquez et al., 2015). Indeed, we are not the first to classify Cincinnati as a nontraditional destination for Latinos (e.g. Jacquez et al., 2015), and this classification is consistent with the logic and criteria established by Gresenz and colleagues (2012). Researchers have deemed investigating how Latinos fare in these new areas a critical task (Guzmán & McConnell, 2002) and have called for community interventions to be continually adapted and evaluated to adjust to the rapid influx of new individuals (Harari, Davis, & Heisler, 2008). Latinos living in nontraditional destinations, including Cincinnati, Ohio, encounter unique experiences, such as discrimination and social exclusion (Engstrom, 2006; Millard & Chapa, 2004; Pérez, Fortuna, & Alegria, 2008) and poorer health outcomes, greater probability of unmet healthcare needs, and lower satisfaction with healthcare compared to majority groups (Gresenze et al., 2012; Stull, Broadway, & Erickson, 1992). For example, Latinos in nontraditional destinations are less likely to have a routine source of care, have health insurance, visited the doctor in the past 12 months, live within 5 miles of a community health center, or live within 10 miles of a safety net hospital (Cunningham, Banker, Artiga, & Tolbert, 2006). The Census Bureau estimates the Hispanic/Latino population of Greater Cincinnati to be approximately 30,000; however, the true population (including undocumented individuals) is estimated to be closer to 70,000 (Zandvakili, Passty, von Hofe, & Mueller, 2010). In Cincinnati, Latinos tend to occupy specific neighborhoods, which are often geographically disconnected (Jacquez et al., 2015). A lack of needed infrastructure is inherent to nontraditional destinations and can include health clinics, social organizations, bilingual services, and legal aid bureaus (Waters & Jimémez, 2005). This issue is exacerbated by geography in that it can limit communication and resources if characterized by detachment. Moreover, this geographic disconnection for Cincinnati Latinos constrains social acceptance and health outcomes (Jacquez et al., 2015). In sum, Latinos living in Cincinnati experience lower perceived healthcare quality, worse mental and physical health, heightened healthcare barriers, and reduced healthcare access compared to the rest of the community (Health Foundation of Greater Cincinnati, 2006; Jacquez et al., 2015). The social cohesion and support provided by other Latinos in cities with a more established history of Latino immigration, such as bilingual store workers or healthcare Journal of Community Psychology DOI: 10.1002/jcop


r

181

providers who provide culturally competent services, are often insufficient in nontraditional destinations (Waters & Jimémez, 2005). This means that Latinos in nontraditional destinations likely face ambiguity in social hierarchies such as class, racial, and ethnic group membership, whereas Latinos in more traditional areas may experience reduced uncertainty in terms of belonging (Waters & Jiménez, 2005). In essence, lack of social infrastructure, especially in terms of social services (i.e. healthcare) may implicitly suggest that Latinos are not welcome in the community. As such, it is imperative to assess the existing infrastructure in these emerging communities to establish and address the aforementioned issues constructively. Cincinnati has struggled to establish a healthcare infrastructure to support the rapidly increasing Latino population, beckoning a call for what Carpenter, Li, and Jiang (2012) refer to as network development research. By employing an SNA approach, practitioners may identify, discuss, and solve organizational and community issues by addressing specific types of relationships amongst individuals and Latino-serving organizations embedded in the community. The current study used SNA to quantitatively document the need for changes in healthcare infrastructure to more fully support the Latino community. We examined three networks related to the health infrastructure for Latino immigrants: the physical health network, the mental health network, and the desired collaboration network, in two ways. Our first objective was to use referral networks to assess the structure of the physical and mental health networks. Second, we assessed the level of desired collaboration within these networks. In conjunction with the referral networks, this offers insights into opportunities for community based interventions.

SNA OVERVIEW Networks are comprised of a set of nodes linked by relationships (Laumann, Galaskiewicz, & Marsden, 1978; Phelps, Heidl, & Wadhwa, 2012). In a social network, both overall network structure and individual node position matter. Thus, current network structure constrains local interactions, which gave rise to macroscopic network structure to begin with (Eiler, Kallen, & Richardson, in press). In developing networks, this means that individual connections between nodes may have characteristic influences on the network as a whole. Because SNA tends to attribute explanatory power to interactions between components, rather than intrinsic properties of the components themselves, these structural properties of a network are most relevant. With respect to the current study, this implies that improving health disparities for Latinos in Cincinnati may be accomplished by increasing positive interactions and collaborations between actors comprising networks that serve the Latino community. Integrative Networks A node’s position in a network may produce distinct advantages, such as increased individual and group performance (Sparrowe et al., 2001) or uneven influence over decision making (Friedkin, 1993). In terms of mental and physical health, policy decisions may affect both specific patient populations and other hospitals because medical care facilities do not operate independently of one another (Lee et al., 2011). Where healthcare infrastructure is emerging or nonexistent for a population, influential nodes must be included in decisions that affect healthcare policy due to the space they occupy in the emerging network. Because these individuals or organizations serve as conduits Journal of Community Psychology DOI: 10.1002/jcop

182

r


between an underserved population and healthcare providers, understanding how influence is perceived and distributed within the network is needed to implement successful and targeted intervention. As such, not only does SNA provide statistical tools, but it also serves as an alternative theoretical lens through which one may include and evaluate the effect that individuals, organizations, and community partners have on outcomes of interest. Furthermore, both individual- and community-level variables may be used as units of analysis, a distinct advantage over many other types of research paradigms. Networks may differ in how collective and individual outcomes manifest (e.g., relationships, resource sharing, other collaborative factors). In networks that facilitate collaboration; such as with the current sample, relationships between organizations and organizational partners increase a community’s capacity via communication, shared resources, and cooperation (Wells, Ford, McClure, Holt, & Ward, 2007). In other words, increases in social capital are leveraged to facilitate network level outcomes. In these integrative networks, political maneuvers such as resource and information hoarding are not desirable because they directly oppose network level outcomes. Instead, organizations and individuals that openly share information and resources, and actively look to support one another to accomplish common goals, enable the emergent outcomes associated with collaborative and integrated networks through interaction. For example, if building infrastructure is a common goal, one way to start is to identify and build relationships amongst community members and organizations that facilitate a more integrative network structure. Related to our first objective, the current project aimed to identify highly cooperative individuals and organizations because these nodes are key to developing collaborative and integrated infrastructure networks. We operationalized cooperation as willingness to refer Latino individuals with physical or mental health issues to others within the network. As cooperation facilitates sharing one another’s resources and expertise, we also identified perceptions of the most desired collaborators to increase effective community service. Network Properties of Interest Two primary structural patterns of a network are density and centralization, often used as measures of embeddedness (Adler & Kwon, 2002). Embeddedness refers to contextual influences on a node—the influence of social ties and positional influences in network structure on participants’ actions (Carpenter et al., 2012). This is the mechanism through which nodes acquire resources and social capital, derived from being implanted within the larger ecology of the network. For example, if a doctor is highly embedded, then he or she may be connected to many other doctors within the hospital. Because hospitals are embedded within a larger healthcare network, the doctor is able to leverage others’ resources as well as the resources of the hospital through their connections. Of interest to us, embeddedness refers to overall network closure in which strong ties between nodes, in addition to high network cohesion, impart a structure of trust and reciprocity in which cooperation characterizes interaction patterns (Carpenter et al., 2012). To this end, network density quantifies the number of ties in the network as a function of the total possible connections (Wasserman & Faust, 1994). In addition, centralization provides a measure of interaction concentration in the network, as opposed to a distributed structure of connections. For the current project, it was expected that measured networks would be highly dense because shared information and resources characterize an integrative network. In particular, we anticipated the physical health network would be more connected (i.e., Journal of Community Psychology DOI: 10.1002/jcop


r

183

dense) because treating physical health issues has a more established history than mental health treatment in Cincinnati. It is widely accepted that successful mental health services require a coordinated effort between a wide variety of specialists, not just counseling, throughout the community (Provan, Isett, & Milward, 2004). Thus, we also expected mental healthcare networks to be integrative in nature, though perhaps less dense than the more established physical health network. Finally, we expected density to be highest in the desired collaborator network because all individuals and organizations surveyed serve the same community.

METHOD SNA was used to uncover the network structure of interrelationships between direct and indirect healthcare service providers for the Greater Cincinnati Latino community along three collaboration dimensions: physical health, mental health, and desired service partners. All study related procedures were approved by the University of Cincinnati Institutional Review Board. Participants To identify potential nodes, we used a snowball technique that included three steps: initial sample generation, snowball generation one, and snowball generation two, coupled with study dissemination (e.g., Fiesler, Fleck, & Meckel, 2010). The initial sample was derived from Internet searches, historical document analysis (e.g., meetings, health fairs, previous datasets), input from two of the authors1 familiar with the network, and inquires from known network members. Because of the dynamic nature of the network, extracting individuals and organizations from existing network members was deemed critical for our analyses. Step 1 generated 34 potential network members. In Step 2, we emailed individuals identified in Step 1 and asked them to identify any additional individual or organizational network members that should be included. This process generated one additional individual, who was included. Thus, the final sample included 35 potential nodes. Potential participants (n = 35) were contacted in Step 3 and sent the study survey via e-mail. A total of 29 individuals returned complete responses; therefore, our final response rate was 83%, which is well above recommended response rates for interpretable SNA results (Kossinets, 2006). All 29 participants were compensated for their time with a $10 gift card. To treat missing data, we took the available case, or reconstruction, approach ˇ (Znidarˇ siˇc, Ferligoj, & Doreian, 2012) to ensure that the data accurately represented all 35 individuals (for discussion, see Stork & Richards, 1992). This reconstruction approach allowed us to keep data in which nonrespondents were chosen as someone to refer a Latino immigrant to (i.e., in-degree ties). This provided a more accurate picture

1 As should be expected by the small size and their inclusion in Step 1 of the network generation, a small subset of

the final list of participants are known by the third and fourth authors. This is primarily due to the high density of shared interest and the lack of infrastructure for Latinos in Cincinnati, thus interactions between individuals serving this community is likely and unavoidable. This does not constitute a confound or a conflict of interest; however, in that the third and fourth authors would have been identified as part of the network by others (in Step 2), even if they had not been identified in Step 1. Finally, the authors do not represent disproportionately influential nodes in any of the networks, thus they have been included in all analyses. Journal of Community Psychology DOI: 10.1002/jcop

184

r


of the network by allowing us to keep some information about nonresponders rather than excluding them. Of the 29 respondents, 31% self-identified as Latino and 52% were at least semifluent in Spanish. Respondents provided diverse services: adult and child mental health therapy; primary healthcare providers (e.g., physicians); legal services; conversational English sessions; homework assistance; child passenger safety education; housing assistance; religious services; job search; academic research; and others. And some participants provided multiple services. Additionally, respondents worked either directly for or with numerous Latino serving organizations: Good Samaritan Free Health Center; Cincinnati Children’s Hospital; Central Clinic; Cincinnati Public Schools; AMIS; Northern Kentucky University; the Diocese of Southern Ohio; The Healing Center; Su Casa; LULAC; Legal Aid Society of Southwest Ohio; Santa Maria Community Services; and others. In our analyses, we included organizations that serve the Latino community, although we did not give organizations an opportunity to fill out the network survey because identifying an individual to fully represent an organization is difficult, if not impossible. Instead, we included these organizations because often community members refer someone to an organization (e.g., hospital or clinic) for physical or mental health services without knowing an individual provider’s name at that location. Network Survey We used a web-based software, SurveyGizmo, to electronically collect data. We used UCINET 6.545 for Windows and NetDraw 2.147 Graph Visualization Software to visualize networks and compute network and node statistics. All potential relations were randomly presented, and individuals were given an opportunity to write in additional individuals they felt should be represented in the network. We also collected demographics including type of services provided, sex, ethnicity, primary language that services are provided in, and country of origin. Referral networks. Participants were asked to select all individuals and organizations to which they would (a) “refer a Latino Immigrant who has physical health needs” and (b) “refer a Latino Immigrant who has mental health needs.” Collaboration network. Participants were then asked to select all individuals and organizations that “by working with more, you would be able to better serve the Greater Cincinnati Latino Community.” RESULTS Network Visualizations First, we show a full visualization of each of the three networks of interest. Next, we show key players for each network, operationalized as the top 20% most influential individuals and organizations. For all networks, black nodes represent individuals, while grey nodes represent organizations. Only organizations and individuals included in analyses are shown. Node size is scaled to a node’s in-degree (number of times a node was chosen), for both individuals and organizations in all visualizations. Each graph is displayed using an algorithm that is biased toward making edges equal lengths and treats nodes as repulsive (NetDraw). Thus, the shape of the network should not be used to infer structural Journal of Community Psychology DOI: 10.1002/jcop


r

185

Figure 1. The overall physical health network for all individuals (black) and organizations (grey) is shown on the left. Note. The top 20% of influential individuals (black) and organizations (grey) in the physical health network is depicted on the right. Size represents in-degree for both.

Figure 2. The mental health network for all individuals (black) and organizations (grey) is shown on the left. Note. The top 20% of influential individuals (black) and organizations (grey) in the mental health network is on the right. Size represents in-degree for both.

properties. In line with our expected results, we found these networks to be integrative in nature, as can be seen from a visual inspection of the network graphs displayed below. Network Results A network’s density is defined as the proportion of possible ties that are present in a given network (Hanneman & Riddle, 2005). This property provides insight into the availability of social capital—the higher the network density, the greater the ability for nodes to leverage each other’s resources. In a dense network, there are multiple pathways that one might take to leverage resources because nodes are highly and redundantly interconnected. An examination of Figure 4 below clearly demonstrates that the desired collaborator network is nearly twice as dense as either the physical or mental health network. Furthermore, the density of the physical health network is greater than the mental health network. Taken together, these data suggest a strong desire for collaboration among individuals who serve the Latino community, yet actual partnerships are lacking. Journal of Community Psychology DOI: 10.1002/jcop

186

r


Figure 3. Desired collaborator network for all individuals (black) and organizations (grey) is shown on the left. Note. The top 20% of influential individuals (black) and organizations (grey) in the desired collaborator network is visualized on the right. Size represents in-degree for both.

Figure 4.

Network density and network degree centralization as a function of referral type.

This also indicates that existing social capital has been better leveraged for physical health issues than for mental health issues, clearly indicating a disparity. We also calculated the network degree centralization index. This is an index of how central all nodes are compared to the centrality of the most central node, interpreted as a measure of social capital distribution (Freeman, 1979). Here, we use in-degree centralization (how often one is chosen) for the physical and mental health networks because how often a referral is given is less meaningful in the current sample. For the collaborator network, we use out-degree centralization because we are interested in who people desire to work with, rather than who they actually work with. In both cases, degree centralization is higher if the network is more integrative and collaborative. Results show a similar pattern, as network density, in that centralization is highest for the desired collaborator network, followed by the physical health network, and subsequently the mental health network. Both of these findings are in line with our expectations and indicate a desire for greater collaboration among service providers, with the most immediate need arising in Journal of Community Psychology DOI: 10.1002/jcop


r

187

the mental health network. Most importantly, this is a direct signal of a need for additional infrastructure for both physical and mental health for Latinos in Cincinnati, evidenced by the drastic discrepancy between actual referrals and desired collaborations that would more positively affect community health.

DISCUSSION The current study examined three networks of direct and indirect service providers for physical and mental health to assess the healthcare infrastructure in a nontraditional destination for Latino immigrants. We also examined potential for collaboration by asking participants to indicate who would help them best serve the Latino community with increased interaction. Overall, we found that networks differed in their structural properties. Specifically, the desired collaborator network was found to be nearly twice as dense as both physical health and mental health referral networks. A similar pattern emerged for network centralization. Together, the density and centralization of all three networks indicate an integrative and collaborative structure. We take this as an indication that the current infrastructure does not suffer from problems with resource sharing or visibility. Furthermore, the top 20% of influential individuals and organizations were also highly interconnected. A high level of connection implies that both individuals and organizations are working together to provide services to the community. Thus, any resource issues are likely attributable to population growth, which has outpaced growth in organizations and support groups for Latinos. As such, several recommendations in line with emerging integrative networks are outlined below. First, we suggest that because individuals are clearly aware of other providers and have a strong desire for increased interaction, there must be some barrier that has hindered collaboration, likely opportunity for contact. We hypothesize that these relationships will continue to strengthen with time, even with the addition of new infrastructure, as predicted by network formation theories of cooperation (Santos & Pecheco, 2005). Importantly, this means establishing a community-wide and cross-sectional task force or coalition to address the lack of infrastructure is valuable. This type of organization has the potential to create a structured system of interaction opportunities—infrastructure that facilitates building community connectivity. Second, support should be focused on establishing new organizations and individual change agents rather than being devoted to developing relationships that do not already exist. This is especially true considering the geographical dispersion of Latinos in Greater Cincinnati and the current trajectory of Latino immigration to the area. In sum, infrastructure must be developed through additional agents capable of imparting change and facilitated by structured opportunities for interaction at the community level. We also established that, in general, organizations were referred to more often than individuals. Because healthcare organizations do not operate in isolation of one another, this brings us to our third recommendation—that healthcare organizations can most effectively leverage each other’s resources by instituting common policy. Distributing this knowledge to highly influential individuals would also strengthen this effort. Additionally, this implies that policy or practice changes, with respect to the Latino community at one organization, are likely to influence changes at others because these organizations are highly redundant. Journal of Community Psychology DOI: 10.1002/jcop

188

r


The physical health referral network did not immediately present issues related to social capital or resource sharing. In fact, both highly influential organizations and highly influential individuals were densely connected. Said differently, both organizations and individuals are aware of each other and seem to be willing to leverage resources in the service of the community. This means that access may not be an issue for physical health disparities. Furthermore, this demonstrates that both indirect service providers and direct service provider organizations are influential and should both be included in important decision-making processes that affect the Latino community. Taken together, this suggests indirect providers should be supported and valued as fundamental to direct service provision—the structure of the community must not be reduced to context alone, but directly involved in decisions that shape the community itself, at least with respect to physical health needs. As such, our fourth recommendation is that influential individuals within the community should be viewed as active agents in change-making processes, even outside of their own institutions, because they wield a social power beyond their job titles. The referral network for mental health issues showed a different pattern of results that illuminates opportunity for improvement. Aside from the fact that the most influential individuals and organizations either do not provide or provide inadequate mental health services, they also were highly disconnected. Unlike the physical health network, influential individuals are not connected to influential organizations. Furthermore, critical individuals and organizations are not connected to one another. For example, one organization is completely isolated and individuals are sparsely connected to each other. This can be interpreted in two ways. On one hand, it could be the case that they are aware of issues with existing service provision. Alternatively, this could mean they do not refer potential patients to influential organizations because they do not offer relevant services in the first place. One might argue the sparse connections could be a result of competition, but this is unlikely because if competition were a driving force for disconnection, then a similar pattern in the physical health network would have been observed. Furthermore, the desired collaborator network shows double the connectivity, evidence that directly contradicts a competition argument for perceived disconnection. Instead, we argue that supportive infrastructure is lacking for mental health services for the Latino population, and because of the lack of infrastructure, individuals are forced to refer potential patients and clients to inadequate providers. Therefore, the last recommendation we have is that the mental health infrastructure should be considerably, but strategically, increased. Limitations and Future Research The current study is not without limitations. First, we were able to analyze only healthcare networks in a single community. Ideally, this study would have investigated a set of similar healthcare communities, thus providing ample comparisons and a more reliable understanding of the social phenomena of interest. As such, future research should use common methodology (i.e., SNA) to examine how healthcare issues pervade nontraditional destinations for Latinos more broadly. Our research can be used as a prototypical case for future research aimed at this goal and results should be compared across communities to determine if similar underlying principles govern the formation of infrastructure across nontraditional destinations for Latinos. Second, our methodology allowed only for cross-sectional comparison. To understand changes in healthcare infrastructure in nontraditional locations, future research should use longitudinal methods in which SNA data are collected over time. This would be Journal of Community Psychology DOI: 10.1002/jcop


r

189

particularly useful in situations in which data collection follows community interventions, policy change, personnel adjustments, or other alterations relevant to the system. In addition, by collecting this type of data, change models (e.g., Markov chain modeling) can be used to describe influence changes and simulate potential influences (e.g., agent based modeling) before actually invoking change (for review of social influence network theory, see Friedkin & Johnsen, 1999). Conclusion In sum, we have identified several needs to address the lack of health-related resources supporting the Latino community in Cincinnati, Ohio. There are two factors that drive connection development in networks: network development opportunities in ongoing networks and an entity’s strategic inducements to shape the network (Ahuji, 2000). For Cincinnati, network development is less actionable because positive network characteristics are likely to remain or even strengthen over time. Because network development is path dependent and dynamic, and therefore future network states depend on developmental context and formative history (Eiler et al., in press), and those measured are already collaborative, we are confident that with time and additional resources and infrastructure these networks will continue this positive trend. Yet network development strategy provides intervention opportunity in that organizations and their partners may aim to develop incentives that encourage strategic collaboration in the Latino network. This leads us to an additional recommendation—that money primarily devoted as internal to a funding organization might be reallocated to establish more interpersonal resources within the larger community network. For example, supporting travel to local Latino healthcare-oriented networking events may provide opportunities to expand relationships within the existing network, but supporting travel to national conferences or recruiting additional mental health specialists may have a greater effect because additional resources (i.e., nodes) should be added to the existing network. National conference attendance and increasing the pool of mental health specialists would be especially successful if both influential individuals and organizations were united in this effort. Finally, because academic researchers are often more successful than community agencies at obtaining funding aimed at increasing service provision across communities, it would be highly beneficial for both organizations and individuals to partner with community-based researchers focused on implementing interventions in nontraditional destinations for Latino immigrants. Most importantly, because SNA can quantify community relationships and how the structure of those relationships relate to outcomes, community-based researchers are uniquely able to document intervention efficacy targeting changes in community infrastructure.

REFERENCES Adler, P. S., & Kwon, S. W. (2002). Social capital: Prospects for a new concept. Academy of Management Review, 27, 17–40. doi:10.5465/AMR.2002.5922314 Ahuja, G. (2000). The duality of collaboration: Inducement and opportunities in the formation of interfirm linkages. Strategic Management Journal, 21, 317–343. doi:10.1002/(SICI)10970266(200003)21:33.0.CO;2-B Journal of Community Psychology DOI: 10.1002/jcop

190

r


Carpenter, M. A., Li, M., & Jiang, H. (2012). Social network research in organizational contexts: A systematic review of methodological issues and choices. Journal of Management, 38, 1328–1361. doi:10.1177/0149206312440119 Carrasquillo, O., Carrasquillo, A. I., & Shea, S. (2000). Health insurance coverage of immigrants living in the United States: Differences by citizenship status and country of origin. American Journal of Public Health, 90, 917–923. doi:10.2105/AJPH.90.6.917 Cross, R., Kase, R., Kilduff, M., & King, Z. (2013). Bridging the gap between research and practice in organizational network analysis: A conversation between Rob Cross and Martin Kilduff. Human Resource Management, 52, 627–644. doi:10.1002/hrm.21545 Cunningham, P., Banker, M., Artiga, S., & Tolbert, J. (2006). Health coverage and access to care for Hispanics in “new growth communities” and “major Hispanic centers.” Washington, DC: Kaiser Commission of Medicaid and the Uninsured. Retrieved from http://www.kff.org/uninsured/7551.cfm De Rubeis, E., Wylie, J. L., Cameron, D. W., Nair, R. C., & Jolly, A. M. (2007). Combining social network analysis and cluster analysis to identify sexual network types. International Journal of STD & AIDS, 18, 754–759. doi:10.1258/095646207782212234 Eiler, B. A., Kallen, R. W., & Richardson, M. J. (in press). Interaction dominant dynamics, timescale enslavement and the emergence of social behavior. In R. Vallacher, A. Nowak, & S. Read (Eds.), Frontiers in Social Psychology: Computational Models in Social Psychology Engstrom, D. W., (2006). Outsiders and exclusion: Immigrants in the United States. In D. W. Engstrom & L. M. Piedra (Eds.), Our diverse society: Race and ethnicity implications for the 21st century American society (pp.19–36). Washington, DC: NASW Press. Fiesler, C., Fleck, M., & Meckel, M. (2010). Corporate social responsibility in the blogosphere. Journal of Business Ethics, 91, 599–614. doi:10.1007/s10551-009-0135-8 Fowler, J. H., & Christakis, N. A. (2008). Dynamic spread of happiness in a large social network: Longitudinal analysis over 20 years in the Framingham Heart Study. BMJ, 337, a2338. doi:10.1136/bmj.a2338 Freeman, L. C. (1979). Centrality in social networks conceptual clarification. Social Networks, 1, 215–239. doi:10.1016/0378-8733(78)90021-7 Friedkin, N. E. (1993). Structural bases of interpersonal influence in groups: A longitudinal case study. American Sociological Review, 58, 861–872. doi:10.2307/2095955 Friedkin, N. E., & Johnsen, E. C. (1999). Social influence networks and opinion change. Advances in Group Processes, 16, 1–29. doi:10.1016/S0378-8733(96)00298-5. Gloor, P. A., Paasivaara, M., Schoder, D., & Willems, P. (2007). Finding collaborative innovation networks through correlating performance with social network structure. International Journal of Production Research, 46, 1357–1371. doi:10.1080/00207540701224582 Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness. American Journal of Sociology, 91, 481–510. Retrieved from: http://www.jstor.org/ stable/2780199 Gresenz, C. R., Derose, K. P., Ruder, T., & Escarce, J. J. (2012). Health care experiences of Hispanics in new and traditional U.S. destinations. Medical Care Research and Review, 69, 663–678. doi:10.117/1077558712457242 Guzman, B., & McConnell, E.D. (2002). The Hispanic population: 1990-2000 growth and change. Population Research and Policy Review, 21,109–128. Hanneman, R. A., & Riddle, M. (2005). Introduction to social network methods. Riverside, CA: University of California, Riverside. Retrieved from: http://faculty.ucr.edu/hanneman/ Harari, N., Davis, M., & Heisler, M. (2008). Strangers in a strange land: Health care experiences for recent Latino immigrants in Midwest communities. Journal of Health Care for the Poor and Underserved, 19, 1350–1367. doi:10.1353/hpu.0.0086



r

191

Harris, J. K., & Clements, B. (2007). Using social network analysis to understand Missouri’s system of public health emergency planners. Public Health Reports, 122(4), 488–498. Retrieved from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1888499/ Health Foundation of Greater Cincinnati. (2006). 2005 Greater Cincinnati Hispanic/Latino Health Survey. Retrieved from https://www.healthfoundation.org/greater-cincinnati-communityhealth-status-survey-2005#gh Jacquez, F., Vaughn, L. M., Pelley, T., & Topmiller, M. (2015). Healthcare experiences of Latinos in a nontraditional destination area. Journal of Community Practice, 23, 76–101. doi:10.1080/10705422.2014.983629 Kossinets, G. (2006). Effects of missing data in social networks. Social Networks, 28, 247-268. Lasker, R. D., Weiss, E. S., & Miller, R. (2001). Partnership synergy: A practical framework for studying and strengthening the collaborative advantage. The Milbank Quarterly, 79, 179-205. doi:10.1111/1468-0009.00203 Laumann, E. O., Galaskiewicz, J., & Marsden, P. V., (1978). Community structure as interorganizational linkages. Annual Review of Sociology, 4, 455–484. doi:10.1146/annurev. so.04.080178.002323 Lee, B. Y., McGlone, S. M., Song, Y., Avery, T. R., Eubank, S., Chang, C., . . . Huang, S. S. (2011). Social network analysis of patient sharing among hospitals in Orange County, California. American Journal of Public Health, 101, 707–713. doi:10.2105/AJPH.2010.202754 Luque, J., Tyson, D. M., Lee, J., Gwede, C., Vadaparampil, S., Noel-Thomas, S., & Meade, C. (2010). Using social network analysis to evaluate community capacity building of a regional community cancer network. Journal of Community Psychology, 38, 656–668. doi:10.1002/jcop.20386 Mercken, L., Snijders, T. A. B., Steglich, C., Vertiainen, E., & de Vries, H. (2010). Smoking-based selection and influence in gender-segregated friendship networks: A social network analysis of adolescent smoking. Addiction, 105, 1280–1289. doi:10.1111/j.1360-0443.2010.02930.x Millard, A. V., & Chapa, J. (2004). Apple pie and enchiladas: Latino newcomers in the rural Midwest. Austin: University of Texas Press. Mora, D. C., Grzywacz, J. G., Anderson, A. M., Chen, H., Arcury, T. A, Mar´ın, A. J., & Quandt, S. A. (2014). Social isolation among Latino workers in rural North Carolina: Exposure and health implications. Journal of Immigrant and Minority Health, 16(5), 822–30. doi:10.1007/s10903013-9784-x. Pérez, D. J., Fortuna, L., & Alegria, M. (2008). Prevalence and correlates of everyday discrimination among U.S. Latinos. Journal of Community Psychology, 36, 421–433. doi:10.1002/jcop.20221 Phelps, C., Heidl, R., & Wadhwa, A. (2012). Networks, knowledge, and knowledge networks: A critical review and research agenda. Journal of Management, 38, 1115–1166. doi:10.1177/0149206311432640 Provan, K. G., Isett, K. R., & Milward, H. B. (2004). Cooperation and compromise: A network response to conflicting institutional pressures in community mental health. Nonprofit and Voluntary Sector Quarterly, 33, 489–514. doi:10.1177/0899764004265718 Santos, F. C., & Pacheco, J. M. (2005). Scale-free networks provide a unifying framework for the emergence of cooperation. Physical Review Letters, 95(098104), 098104-1-098104-4. doi:10.1103/PhysRevLett.95.098104 Sparrowe, R. T., Liden, R. C., Wayne, S. J., & Kraimer, M. L. (2001). Social networks and the performance of individuals and groups. Academy of Management Journal, 44, 316–325. doi:10.2307/3069458 Stork, D., & Richards, W. D. (1992). Nonrespondents in communicational network studies. Group & Organization Management 17, 193–209. doi:10.1177/1059601192172006 Stull, D. D., Broadway, M. J., & Erickson, E. P. (1992). The price of a good steak: Beef packing and its consequences for Garden City, Kansas. In L. Lamphere (Ed), Structuring diversity:


192

r


Ethnographic perspectives on the new immigration (pp. 35–64). Chicago: University of Chicago Press. U.S. Census Bureau. (2012). U.S. Census Bureau projections show a slower growing, older, more diverse nation a half century from now. Retrieved from http://www.census.gov/newsroom/releases/archives/population/cb12-243.html Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications. Cambridge, UK: Cambridge University Press. Waters, M. C., & Jiménez, T. R. (2005). Assessing immigrant assimilation: New empirical and theoretical challenges. Annual Review of Sociology, 31, 105–125. doi:10.1146/annurev.soc.29.010202.100026 Wells, R., Ford, E. W., McClure, J. A., Holt, M. L., & Ward, A. (2007). Community-based coalitions’ capacity for sustainable action: The role of relationships. Health Education and Behavior, 34, 124–139. doi:10.1177/1090198105277851 Young, A. M., Halgin, D. S., DiClemente, R. J., Sterk, C. E., & Havens, J. R. (2014). Will HIV vaccination reshape HIV risk behavior networks? A social network analysis of drug users’ anticipated risk compensation. PLoS ONE, 9, 1–10. doi:10.1371/journal.pone.0101047 ¨ uk, R., & Milanov, H. (2010). It’s the connections: The network perZaheer, A., Gözübuy¨ spective in interorganizational research. Academy of Management Executive, 24, 62–77. doi:10.5465/AMP.2010.50304417 Zandvakili, S., Passty, B., von Hofe, R., & Mueller, A. (2010). Economic impact of Hispanic community in the Cincinnati MSA. Cincinnati: University of Cincinnati. ˇ Znidarˇ siˇc, A., Ferligoj, A., & Doreian, P. (2012). Non-response in social networks: The impact of different non-response treatments on the stability of blockmodels. Social Networks, 34, 438– 450. doi:10.1016/j.socnet.2012.02.002
