Extending the conversation: a network analysis of academic associations in workplace spirituality

Examining academic networks provides insight into boundaries and boundary crossing as well as knowledge diffusion. We examine four academic networks focused on business, spirituality, and religion to identify network boundaries and boundary crossing. Scholars tend to align with networks consistent with the nature of their employing institution, both in its relative emphasis on research and its religious affiliation. Network and religious differences contribute to the relative isolation of research communities, despite shared topical interests and reliance on similar scholarly sources. The use of normative authorities may limit networking, but the relative absence of weak network links across networks may underestimate similarities. Increased boundary crossing may enhance innovation across networks.

Burgeoning interest in business, religion, and spirituality has catalyzed the creation of several research networks devoted to work and spirituality (cf. Oswick 2009, Neal 2013. While a general impression exists of the various niches these networks serve, their precise overlap and distinguishing characteristics have not been explored. Additionally, insights on the relationship of workplace spirituality and religion may be gleaned among these networks since the networks differ in their epistemological grounding (cf. Phipps and Benefiel 2013, Lips-Wiersma and Mills in press). Our purpose is to identify distinctive and shared elements of workplace spirituality networks to gain an understanding of the boundaries which separate them and the bridges which connect them, both of which impact knowledge diffusion. In our exploration, we will apply social network theory to understand workplace spirituality research communities.

Social network theory
Social network theory focuses on connections among nodeswhether these are individuals, organizations, nations, or webpagesand flows of resources from one node to another (cf. Marin and Wellman 2011). Summarizing Atkin (1974Atkin ( , 1977, Borgatti and Lopez-Kidwell (2011) describe the two basic subdivisions of a social network as being constituted by backcloth and flows. In their words: "The backcloth consists of an underlying infrastructure that enables and constrains the traffic, and the traffic consists of what flows through the network, such as information" (p. 44) (see Figure 1). The backcloth is divided into two elements: similarities among network members (e.g. shared attitudes, beliefs, employers, or demography) and social relations which may be role based (e.g. teacher of or colleague of) or cognitive/affective (e.g. knowing or liking). The traffic or relational events include the interactions themselves and the flows of information, contacts, citations, or other exchanges from one member to another. Relational events feedback to the backcloth, reinforcing or altering perceived similarities and social relations (Marin and Wellman 2011). Our focus is on the backcloth elements, which channel interactions and flows among workplace spirituality research networks. We begin by describing some backcloth of our ownthe nature of academic associationswhich provides a context for understanding how these particular networks function and interact.

Academic networks
Formal academic associations and regularly occurring conferences are networks, which establish and reinforce norms regarding theory, research, and education (Greenwood et al. 2002). Participation in associations and conferences creates social network similarities, which link members to each other (Marin and Wellman 2011). Networks can promote knowledge convergence (Cooke andLang 2009, Wilson andMcKiernan 2011) while simultaneously providing a forum for divergent views (Perry-Smith andShalley 2003, Hülsheger et al. 2009). Associations influence research by reinforcing professional norms, catalyzing research themes, and providing research outlets and forums, which influence research, teaching, and practice through conferences and journals (cf. Greenwood et al. 2002, Baumgartner and Pieters 2003, March 2004, Daft and Lewin 2008, Parada et al. 2010, Xiao 2010. As Daft and Lewin (2008, p. 178) state, "Journal subcommunities evolve toward local convergence in knowledge, beliefs, and paradigms as reflected in a journal's published articles. Publications in a specific journal typically share a common world view …" Academic networks often shape members' perceptions of themselves and of others outside the profession (Hewstone et al. 2002, Cooper and Thatcher 2010, Gioia et al. 2010, Jones and Volpe 2011. Associations serve niches of scholars, often developing unique subcultures in the process (Glynn andAbzug 2002, Mataira andVan Peursem 2010).
Academic research networks often focus on market segments such as subdisciplines or institutional types, thereby avoiding competition unless resources become scarce (Haider-Markel 1997). As "mutual nonprofits", these organizations lean toward serving members rather than the public (Quarter et al. 2001). This inward orientation often weakens the perceived need to compete or grow beyond current and/or potential members as long as resources (e.g. members, manuscripts) are adequate. Despite their specialty focus, research networks may seek legitimacy by mimicking the practices of high status peer organizations (Suddaby 2010). As organizational life cycles progress, organizations often calcify and their democratic governance is not always amenable to innovation. Quarter et al. (2001) suspected as much in their study of mutual nonprofits (p. 371).
Among mutual nonprofits, academic research networks are unique in some ways. The public nature of scholarly knowledge and scholarly search tools allows knowledge to flow from one researcher to another, even if they are not members of the same association; membership in one organization does not preclude linking with another's nodes (e.g. members or articles). Incentives exist for scholars to network broadly since diverse linkages yield knowledge and innovation often not found in closed networks (Brass et al. 2004, Zou andIngram 2013).
One last element to introduce in considering a field of networks is linking from one network to another via boundary crossing. As mentioned, academic networks have porous boundaries because knowledge and scholars can be accessed from outside the network. While direct participation with others may create relatively strong ties and reinforce flows of research knowledge, methods, and partnerships (cf. LeRoux et al. 2010), spanning multiple networks can produce weak ties which lead to innovation and learning (Akkerman and Bakker 2011, Marsden and Campbell 2012, Zou and Ingram 2013. Thus, while association subcommunities may normatively form their members, interactions across borders may enhance scholarly breadth and insight. Strong academic ties are often characterized by frequent or salient interactions, which reinforce particular viewpoints.
So where does this discussion take us? First, associations and conferences are important networks among academic researchers. They catalyze the formation of relatively strong links which open flows of knowledge, interpersonal connections, and other resources. Academic networks often coexist, focusing on niches of members and their interests. They also tend to be aware of and mimic one another in select ways along a perceived social hierarchy. Although strong linkages within networks may buttress world views, cross-boundary linkages may stimulate innovation. These general understandings provide a context for the present study and invite a closer examination of workplace spirituality research networks.

The study
Although publications and presentations exist beyond them (Tracey 2012), four networks specialize in workplace spirituality research and represent a majority of academic presentations and publications on the topic. These networks are as follows: The Christian Business Faculty Association (CBFA), Colleagues in Jesuit Business Education (CJBE), the International Symposium on Catholic Social Thought and Management Education sponsored by the John A. Ryan Institute at the University of St. Thomas (CSTME), and the Management, Spirituality, and Religion interest group of the Academy of Management (MSR) (see Table 1). 1 The degree of distinctive and shared elements of these networks will aid in understanding their boundaries and their existing and potential cross-boundary connections.
Do networks vary in the scholars they attract? Colleges and universities often group together in networks of similar institutions (e.g. by funding, geography, religious affiliation, mission, competitive sets, aspiration), so it might be reasonable to expect that faculty at these institutions anticipate more commonalities with members of similar institutions than they would from gatherings of more diverse membership (Chen and Kenrick 2002) or that their institutions tend to endorse particular affiliations formally or informally. Thus, two of the networking similarities which might direct academics to research networks are the mission and academic emphases of their institution. Hence, we reason that: H1: Scholars tend to present research at business and spirituality associations whose religious orientation corresponds with the religious affiliation of their home institution.
H2: Scholars tend to present research at business and spirituality associations whose research orientation corresponds with the research orientation of their home institution.
Do networks vary in the scholarship they sponsor? Given that institutions vary by religious affiliation and research orientation, the type of scholarship they present at conferences and publish in journals is likely to reflect these differences. Some institutions are more likely to encourage research, which link business with particular religions while others focus on non-sectarian expressions of spirituality. Some may encourage basic empirical research while others encourage pedagogical or applied scholarship. If these distinctions exist, they suggest that the associations may differ in the niche they serve and/or type of knowledge flows they contribute to workplace spirituality research or that they overlap in one or both. Given these observations, we hypothesize that: H3: Teaching-oriented associations attract a higher proportion of theological and pedagogical genre papers than do research-oriented associations. The study of the relationship and relevance of spirituality and religion in management and organizations. Major topics include the following: theoretical advances or empirical evidence about the effectiveness of spiritual or religious principles and practices in management, from approaches represented in the literature including religious ethics, spirituality and work, and spiritual leadership, as well as applications of particular religions, and secular spiritualities to work, management/ leadership, organization, and the business system; and evaluation studies of the effectiveness of management approaches that nurture the human spirit in private, non-public, or public institutions H4: Research-oriented associations attract a higher proportion of theoretical and empirical genre papers than do teaching-oriented associations. Lynn et al. (2011) observed that denominational affiliation does not appear to alter the relative emphasis of various religious concepts. Similarly, we suspect that business scholars, regardless of the association, will explore similar concepts and issues. If this is true, it would follow that workplace spirituality networks overlap in their flows of knowledge. Whether from religious or non-religious institution, we hypothesize that: H5: Associations focusing on business and spirituality have more topical content in common than not.
An association's epistemological assumptions are reflected in each association's peer-review process, resulting in commonalities and differences across networks. Especially, in normative-based scholarship, this substructure may be particularly pronounced (cf. Lips-Wiersma and Mills in press). Papers presented at associations which address business-religious matters may appeal to religious authority consistent with the association. Papers presented at Catholic-oriented associations, for example, can be expected to more frequently reference Catholic documents than scholars at non-Catholic-oriented associations. Biblical references may be more common at religious than in secular associations. The authorities relied upon potentially differentiate knowledge flows in workplace spirituality. Thus, we hypothesize that: H6: The basis of religious authority used in papers presented at associations focusing on business and spirituality differ by the nature of the association.
Finally, the degree to which business practice is viewed as acceptable by business faculty members is dependent on the criteria used to make the evaluation. Since religiosity and spirituality have been shown to affect individuals' assessments of the acceptability of business practices (Keller et al. 2007, Kum-Lung andTeck-Chai 2010), it is logical to expect that on average, critical views will share more in common within networks that across networks. Hence, we hypothesize that: H7: The degree of business critique expressed in papers presented at associations focusing on business and spirituality varies with the association's religious affiliation.

Associations
The CBFA is Protestant in nature. CJBE and the International Symposium on Catholic Social Thought and Management Education (CSTME) are Catholic. The Management, Spirituality, and Religion interest group of the Academy of Management (MSR) exists as a subgroup within a large secular association.
CBFA and CJBE are associations, MSR is an interest group within a larger organization, and CSTME is a series of conferences tied to an academic center with no formal membership. Each network issues calls for papers and utilizes a peer-review process for selecting papers for conference presentation. CBFA, CJBE, and MSR publish journals in addition to holding annual conferences. We relied on authors and citations within journals and proceedings to gage network linkages and focus on network participants and the scholarship they produce as indicators of member similarities and social ties.

Sampling, data, and analyses
We tested our hypotheses using proceedings and journal articles from each of the four research networks from which we examine article content, author institutions, and reference citations. We utilized a mixed-methods approach (following Heyvaert et al. 2013), combining qualitative and quantitative measures on the grounds that qualitative methods were best suited to examining scholarship and quantitative methods were best for examining scholars. We began by gathering all publically available peer-reviewed proceedings from the six most recent conferences for each association. We included entire papers and extended abstracts of five pages or more, excluding plenary presentations and abstracts shorter than five pages on the basis that the former often are unique in genre and the latter did not offer comparable content to a full paper. Because we are interested in issues of business and religious content, we deleted papers from the population which contained no discernable religious or spiritual reference. From this population of proceedings papers, we took a random sample of twenty papers from each association, stratified by year of presentation. This yielded a total of eighty peer-reviewed conference papers for examination.
Some variables in the study did not require interpretative coding, but those that did were as follows: The degree to which various authority sources (e.g. scripture, experience) were cited; the academic discipline of the paper (e.g. accounting, economics); the degree of faith integration represented in the paper; the spiritual/religious focus (e.g. overtly Christian, spiritual but not religious); and the academic genre (e.g. empirical, pedagogical) (see Table 2 for measurement definitions).
Because proceedings sometimes omit references, a second source of data was needed to examine the citations to academic serials. Thus, two years of journals were gathered from the associations and citations from all articles were examined. The journals utilized were as follows: Journal of Biblical Integration in Business (JBIB) and Christian Business Academy Review (CBAR) published by CBFA; Journal of Jesuit Business Education (JJBE) published by CJBE; and Journal of Management, Spirituality, and Religion (JMSR) published by MSR. CSTME does not sponsor a journal. In addition to the proceedings and journal content data, we gathered archival data from a variety of sources to identify the characteristics of author institutions. These included the college or university's institutional network membership and business accreditation (see Table 2). Two types of coding were employed in the study. Automated coding for vocabulary word counts and religious references was performed with NVivo 9.0. Interpretive coding for other variables required a coding scheme (cf. MacQueen et al. 1998). A researcher and an assistant coded a sample of papers and compared their ratings. When differences occurred, coders discussed their ratings. With inter-rater reliability established, the researcher and assistant divided the remainder of the papers and coded them separately. Both coders were unaware of the research hypotheses until they completed their tasks.
In terms of analysis, in addition to using NVivo 9.0 to generate word frequency counts qualitatively, BCFinder was used for citation analysis (Lehmann et al. 2008) and SPSS (PASW Statistics 18) was utilized for author and institution analyses. For analyses dealing with scholars, the individual authors were the unit of analysis. When scholarship was examined, the paper was the unit of analysis.

Results
Hypothesis 1 suggested that associations differ in the types of scholars they attract. Specifically, H1 predicted that associations attract scholars from networks of corresponding religious orientation. Using network membership as an indicator of institutional affiliation and association mission statements to identify the affiliation of the network, this hypothesis was strongly supported ( χ 2 = 181.10 (6, N = 130), p < .001, V = 0.84, likelihood ratio = 181.10): Scholars from institutions aligned with the Association of Catholic Colleges and Universities (ACCU) constituted the majority of sampled attendees at CSTME and CJBE; scholars from institutions affiliated with the Protestant, Council for Christian Colleges and Universities (CCCU) were in the majority in the CBFA sample; and scholars from institutions affiliated with neither of these networks formed the largest group in the MSR sample (see Table 3). CSTME was the most mixed in institutions represented. Networks were not exclusive in their membershipsdiversity existedbut network-association alignments were apparent. Hypothesis 2 predicted that scholars would gather together according to the research orientation of their institution. This hypothesis also was supported ( χ 2 = 74.99 (9, N = 116), p > .001, V = 0.46, likelihood ratio = 70.43). The majority of institutions represented among the CBFA sample were affiliated with a business accreditation with little or no research emphasis; the other accreditations generally attracted institutions with a strong or moderate research emphasis accreditation. 2 Although both H1 and H2 were highly significant, the Cramer's V statistic (a standardized effect size estimate) indicates that network affiliation is the stronger of the two variables in predicting association membership.
Hypotheses 3a and 3b addressed types of scholarship expected at the associations. H3a predicted that networks with a strong teaching-oriented culture would attract a higher proportion of theological and pedagogical genre papers. H3b predicted that networks with strong research-oriented cultures would attract a higher proportion of theoretical and empirical papers. Teaching and research orientation was determined by considering the research and teaching emphases of each association's mission statement (Table 1) and two proxy variables for teaching/research orientation: the percentage of author institutions with a Carnegie Classification of Masters/Large or higher and the percentage of author institutions with a business accreditation that emphasized research strongly or moderately (i.e. AACSB, EQUIS, and AMBA). CBFA was determined to lean toward a teaching-oriented culture while the other networks favored a research orientation. In the analysis, CBFA had a nearly identical proportion of theoretical/empirical and pedagogical papers as did the research-oriented associations ( χ 2 = 0.01 (1, N = 64), p = .92). H3 was not supported.
Among the twenty CBFA papers, 50% were coded as theoretical/empirical in genre compared with 45% of the papers in the research associations. CBFA's pedagogical papers accounted for 35% of the total, whereas the research associations accounted for 33% of the total. Table 4 provides more detail on the papers presented at the individual associations. Noteworthy observations include the following: the large percentage of pedagogical papers across the associations except for MSR; the relatively high empirical contributions at MSR and CBFA; the pedagogical focus of CJBE; and the similarities in genre between CJBE and CSTME. Hypothesis 4 predicted that associations have more topical content in common than not. To test this hypothesis, NVivo was utilized to compare the 30 most frequently used words in the papers of each association. While no statistical test of word similarity is available and the topics of papers vary considerably, several words common across groupings of papers were identified (Table 5). Seven (23%) of the thirty most frequently occurring words in each network appeared in all four networks. An additional seven words (47%) appeared in the top thirty words across three networks, and an additional 13 words (90%) appeared among the most common thirty words of two associations. Despite the absence of a single quantitative test, these percentages suggest cognate topics are being discussed across networks. Hypothesis 5 predicted that the authorities used in papers would correspond to the nature of the association. For this analysis, MSR was coded as secular and the other associations, as religious. This hypothesis was supported (F = 7.56 (N = 79), p = .007). MSR papers averaged 1.73 when coded on a Likert scale where 1 = none or slight, 2 = occasional, 3 = moderately, and 4 = strongly appeal to the authority of scripture and/or religious tradition. Religious associations averaged 2.32 when coded, scoring between occasional and moderate. A secondary analysis of biblical and Catholic social teaching references provided additional insight in religious authority (see Table 6). CBFA authors relied heavily on scripture as an authoritative source while CSTME authors relied heavily on sources of Catholic social doctrine. 3 To compare the use of scholarly authority sources, we tabulated the most frequently cited serials in each journal (Table 7). JMSR has a longer list and more citations in part because it published four issues a year rather than the single annual issue produced by each of the other journals. Figure 2 shows the most frequently cited top third of serials and the journals which cite them. Although no statistical test exists for this analysis, Table 7 and Figure 2 indicate that authors publishing in each network share several of the same scholarly serials as their most frequently cited sources. They do not, however, cite each other's publications, despite the common interests suggested by their vocabulary (H4). Finally, hypothesis 6 predicted that the degree of business critique in papers presented at associations varies with the association's religious affiliation. The hypothesis was supported (F = 5.33 (N = 67), p = .002). Table 8 shows the distribution of papers across the business critique spectrum for each association. MSR papers generally suggested augmenting business, CSTME suggested adjusting business practice, and the other two associations fell between these points. Thus, the secular network suggests the least adjustment in business and Catholic networks suggest the most. A second analysis examined the religious or spiritual focus of papers, from "Spiritual but not Religious" to "Overtly Christian" (see Table 2 for details). 4 Again, there were significant differences among the associations (F = 36.95 (N = 76), p < .001) with MSR toward the spiritual end of the spectrum, CBFA and CSTME being overtly Christian, and CJBE falling between (Table 9).

Discussion
The aim of this exploratory study was to compare the scholars and scholarship of four professional associations researching business and spirituality to identify distinctive and shared elements of workplace spirituality networks, gaining an understanding of the boundaries which separate them and the bridges which connect them, and enhancing workplace spirituality research within and across research networks. The existence of four academic networks servicing a relatively specialized topic may appear redundant. Indeed, network authors cite similar academic literature, discuss similar topics, and produce similar genres of scholarship. Networks, however, serve different groups of higher education institutions and differ in their epistemological authorities, their focus on spiritual or religious workplace emphases, and their views of business. These distinctions and similarities explain in part the reason for separate networks and suggest possibilities for boundary crossing. Normative forces among the academic networks appear to buttress association boundaries. The secular or religious identification of the member institutions, the value placed on research, and the use of religious authorities distinguish religious from secular workplace spirituality networks, despite their common interests and scholarly sources. Reimer (2011) suggested that more conservative Christian denominations demonstrate narrower doctrinal allowances among congregations associated with a particular denomination. Conservative religious colleges serve in part as moral communities, assuming religion to be rational and vital to the institution's mission, and accepting scriptural and theological authority (cf. Hill 2011). Within secular institutions, positivist approaches are valued in social science research and religious sources are uncommon. Although the associations overlap in interests and scholarly sources, the common use of religious scripture or church tradition differentiates religious from secular networks. Thus, despite appeals for scholars from across the interest area to collaborate (Bostwick and Lowhorn 2012), sources of authority may provide an obstacle for some scholars and/or institutions. This divide not only suggests different approaches to scholarship but complicates the mutual citing of research. Differences in normative authorities may explain in part the paucity of Protestant and Catholic network scholars citing the other's work, although a lack of awareness due to network and institutional alignment may also contribute to the relatively parallel scholarship.
Boundary crossing: opportunities for conversation Despite common interests, boundaries appear to divide workplace spirituality networks and scholars. But this does not mean that boundary crossing is impossible or inappropriate. Indeed, some boundary crossing already occurs as evidenced by a small number of scholars who contribute scholarship to networks beyond their institutional affiliation or interests. Research suggests that boundary crossing yields at least two benefits. First, Chen and Kenrick (2002) show that while individuals may be attracted to groups of other like-minded scholars, they also can discover that they have more in common with outgroups than they originally perceived. In other words, scholars may find more similarities in scholarly conversation in other networks than they presumed.
Second, innovative and critical thinking through cross-network communication and an awareness of diverse perspectives may enhance the quality and insight of research among scholars (Perry-Smith andShalley 2003, Hülsheger et al. 2009). Although networks may retain their unique character (Van Hise and Porco 2010), research may become richer through connections across networks.
Because rhetoric and ritual influence institutionalization (Sillince and Barker 2012), it is possible that network leaders could encourage greater boundary spanning through familiarization with other scholars and scholarship attracted to other associations. Cross-network conversation and collaboration are likely to benefit from trust, complementary competencies, and incentives for collaboration (Stephenson and Schnitzer 2006, Tsasis 2009, Bunger 2012.

Limitations and future research
Several research limitations should be noted. Although twenty papers from each association spread over five conferences are a large sample for a qualitative study, the sampling may unevenly represent the four associations. Papers vary substantially in their word usage, scripture, or Catholic social thought references, for example, and can distort means for each association. Multiple coders were used during the calibration stage, but a single individual coded most papers. Even with a coding rubric, coding errors could be introduced. Utilizing multiple coders for each document is a safeguard against some coding errors. Finally, minimal institutional research exists on network fields, and qualitative data lend themselves to numerous angles and analyses. Institutional theory itself is subject to social construction (Mizruchi and Fein 1999). An attempt was made to be conservative in the tests used and interpretations drawn, but the findings and interpretation are subject to interpretative bias. Expanding the data set could increase confidence in findings beyond the present study. Despite this wariness, the findings are strong and offer insight into the boundaries of scholarly networks.

Notes on contributors
Monty L. Lynn is WW Caruth Chair of Owner-Managed Business and Associate Dean at Abilene Christian University. His research interests include international poverty and development and workplace spirituality.
David J. Burns is professor of Marketing and Director of Faculty Programs at Xavier University. He has co-authored several books, published over 95 journal articles and book chapters and presented over 200 papers. His research interests include mission integration, retail location and atmospherics, ethics, and consumer culture. Notes 1. Omitted from the list of networks are ones which focus exclusively on business ethics (e.g. the International Vincentian Business Ethics Conference) and those which do not regularly host national or international scholarly conferences with several paper presentations (e.g. the Association of Christian Economists). 2. Research orientation aligns with the institution's total student enrollment. Mean enrollment of the authors' institutions by association were as follows: MSR = 16,420; CSTME = 11,883; CJBE = 9793; and CBFA = 3412. This difference was significant (F = 17.05 (N = 127), p < .001).