Network theory and research have identified the powerful dynamic of homophily whereby individuals are more likely to connect with similar rather than dissimilar others. However, less is known about when individuals might connect with dissimilar others to enhance organizational diversity benefits and mitigate social exclusion. This study builds upon prior evidence linking homophily to both elite group closure and lower-status self-segregating social withdrawal to propose a new dynamic of middle-status ambivalence toward homophily, generating a greater propensity toward diversity. Indian society offers a unique context of a legally codified, stable, three-tiered status hierarchy to test the idea that middle-status individuals will develop the most diverse networks. Using a unique longitudinal friendship network dataset, we find empirical support for our prediction, a novel contribution extending classic homophily theory. In addition, due to potent status dynamics we theorize and find that lowest-status group members initiate more ties over time but also exhibit lower tie stability and greater losses of early ties than higher-status group members, resulting in the appearance of lowest-status group homophilous withdrawal. Overall, we advance homophily theory and offer practical suggestions for organizations seeking to enhance social ties across diversity and improve the effectiveness of social inclusion policies.

INTRODUCTION

Rohith Vemula, a lower-caste Ph.D. student, committed suicide at the University of Hyderabad in India on 17 January 2016. This was the eighth case of suicide in the prior decade by a lower-caste doctoral student at the University of Hyderabad alone. The event caused a national uproar and brought to fore the deep frustration, alienation, and discrimination faced by lower-caste students (see Kumar, 2016). The tragic suicide has provided the impetus to deepen the scholarly exploration of lower-caste perspectives and experiences in India (Chrispal et al., in press; DiTomaso, in press). This paper examines the culture of toxicity driven exclusion faced by lower-caste individuals in friendship networks and points to opportunities within the caste hierarchy for bridging these deep-seated caste divides.

Research on interpersonal networks has identified homophily, defined as the tendency of individuals to connect with similar others (Lawrence and Shah, 2020; Seidel, 2011), as a key basis for social relationships. Similarity appears to breed connections of all types between individuals, including marriage, friendships, and advice networks (McPherson et al., 2001). Informal networks within work organizations also tend to be homophilous (Ibarra, 1995; Reagans, 2011), with clear negative implications for organizational functioning and inclusion. In particular, homophilous networks limit and divide the social worlds of people and reduce their exposure to diversity (McPherson et al., 2001). Contemporary workplaces are becoming more diverse over time, but the tendency toward homophily limits the benefits of diversity for organizations as homophilous network ties limit the exchange of diverse ideas, resources, and knowledge (Ertug et al., in press).

Given the usefulness of diverse informal networks to organizational functioning, increasing the extent to which employees connect across diversity may add value to the organization. Achieving this goal requires better understanding of the factors driving homophily dynamics and their negative impact on network diversity (Rivera et al., 2010). We argue that status distinctions between identity groups are important sources of homophily creating barriers to network ties across diversity by inhibiting cross-group connections (Lawrence and Shah, 2020; Seidel, 2011). We develop the notion of middle-status ambivalence toward homophily to highlight the greater engagement of middle-status individuals in forming ties across diversity. We also highlight inclusion challenges for the lowest-status group by examining the substantial disadvantage that the lowest-status individuals experience in their attempts to build and maintain relationships. Overall, given the importance of diversity and inclusion in organizations, our research focuses on the development of friendship ties across status groups over time.

Our empirical work focuses on the development of friendship ties in a three-tiered status hierarchy. We choose to focus on friendships because these ties are associated with the exchange of many different types of resources, including information, emotional support, and even material help such as job opportunities. Friendship ties associated with interpersonal trust are valuable in the workplace because they enhance the sharing of scarce, tacit, and valuable information (Coleman, 1990; Krackhardt, 1992; Obstfeld, 2005). Prior research has highlighted the importance of friendship ties in both academic (Baldwin et al., 1997; Gibbons and Olk, 2003; Konrad et al., 2017) and organizational contexts (Krackhardt and Kilduff, 1990, 1999; Lincoln and Miller, 1979; Sasovova et al., 2010).

Two views inform our current understanding of hindrances to friendship ties across diversity in status group hierarchies. The lower-status social withdrawal perspective suggests that ties across diversity are inhibited by the networking behavior of members of lower-status groups as they self-segregate into homophilous networks (Mollica et al., 2003). An alternative elite closure (Konrad et al., 2017; Tomaskovic-Devey, 1993) view suggests that members of lower-status groups have little reason to self-segregate but instead have strong motivation to seek friendships with members of the resource-endowed higher-status groups (Ibarra, 1995). Hence, homophilous patterns result from higher-status individuals limiting their friendship ties across status diversity. These two alternative theoretical mechanisms point us in different directions regarding whether the actions of higher- or lower-status group members most hinder the development of diverse organizational networks.

One reason for our limited understanding is that prior findings emanate from research that typically has examined friendship ties using a two-tier status hierarchy where, for example, scholars in North America compare the networks of White MBA students to others such as their African American counterparts (Mollica et al., 2003) or use multiple identities (e.g., citizenship and gender) to rank status groups in a hierarchy (Konrad et al., 2017). By comparison, we examine the pattern of friendship ties in a more complex three-tier status hierarchy in order to observe more than two hierarchically-ranked status groups. As such, we can compare a middle group to both a higher-status and a lower-status group. Such comparisons enable us to test the arguments on both sides of prior debates regarding whether it is the higher- or the lower-status group members who choose to self-segregate. This approach also opens the possibility of confirming that both processes operate independently and simultaneously. We aim to disentangle these dynamics both theoretically and empirically.

Indian society offers a unique research context where a complex set of factors have led to a readily identifiable three-level status hierarchy based on caste identity (Cotterill et al., 2014; Davidson, 2018; Deshpande, 2011; Qureshi et al., 2018b; Srinivas, 1996; Vaid, 2014; Zacharias and Vakulabharanam, 2011). The three-tier status hierarchy in India includes a lowest-status group that has faced ostracism and degradation for thousands of years, an elite group that is legitimized by religion, and a middle-status group situated between the two. This context of a three-tier hierarchy is particularly useful in advancing theory developed primarily in Western contexts (e.g., Faems et al., 2018) to consider the stronger middle-status group inclination to build friendship ties across status diversity within the two mechanisms of elite closure and lower-status social withdrawal.

In sum, the purpose of this paper is to address the following research questions: do lower-middle- or elite- status group members show the most (least) homophily in their personal networks? and do the lowest-status group members more actively attempt to build network ties and have less success in maintaining network ties? Conceptually, we add a novel argument explaining that some individuals build more diversity into their personal networks than others, based upon their middle-status location in the identity group status hierarchy. Moreover, we propose that the lowest-status group members face particular disadvantage in their efforts to build munificent and long-lasting friendship networks. By collecting network data at two points in time, we are able to examine how the networks of different status groups develop and change over time and test predictions about group differences in initiating and maintaining network ties.

HOMOPHILY AND STATUS: THEORY DEVELOPMENT

Prior research has suggested the alternative mechanisms of lower-status group withdrawal and elite closure for explaining homophilous friendship ties among members of low- and high- status groups (Konrad et al., 2017; Mollica et al., 2003). In this section, we integrate these two mechanisms while making two assumptions. First, based upon prior research and theory, we assume that human identity groups exist in a status hierarchy (Sidanius and Pratto, 1999) and that, in a three-tiered hierarchy, the elite group enjoys the greatest access to material and non-material resources. Members of the lowest-status group have the lowest levels of resource endowments, while middle-status individuals show moderate resource endowment levels. Second, we also assume that individuals seek to maximize resources through their personal networks and that seeking resources is a powerful motivation for building friendship ties (Ingram and Roberts, 2000).

Given the strong motivation to maximize resources, we argue that the apparent tendency toward lowest-status group withdrawal emanates from mistreatment and toxicity in social interaction. Toxicity refers to the cluster of negative social experiences that lowest-status group members experience in organizational settings ranging from incivility, microaggressions, discrimination, and subtle prejudice to overt acts of hostility and harassment (Barling et al., 2009; Bowling and Beehr, 2006; Deitch et al., 2003). High levels of toxicity create barriers to forming ties across status diversity for lowest-status group members. We suggest that the greater the toxicity experienced by the lowest-status group members, the stronger would be their ultimate self-segregation into homophilous friendship networks. But we also acknowledge the possibility that some lowest-status group members may choose to distance themselves from their status group identity, at least in their visible actions, in an attempt to deflect such toxicity (Derks et al., 2011).

In contrast to the lowest-status groups, members of elite groups in most organizational settings enjoy expectations of high competence and are considered capable of adding value to the organization (Ridgeway, 2014). The tendency of deference towards members of higher-status groups means that individual members from the elite group expect and receive positive reactions when they approach members of lower-status groups (in this case, members of both the middle-status and lowest-status castes). Positive expectations of high capability minimize the likelihood of receiving toxicity, as perpetrators avoid targeting those who might be able to benefit them in the future. Thus, elite group members experience low levels of toxic social interaction in organizational settings; for example, meta-analytic research has shown that White men report fewer experiences of workplace mistreatment, harassment, and discrimination compared to racial minorities (McCord et al., 2018). Situated between the lowest and the elite groups, middle-status group members experience moderate levels of toxicity. Altogether, as group status increases, the toxicity-linked barriers to forming friendship ties across diversity also decline.

The need to access social and material resources also motivates members of the lowest-status groups to seek ties with better-resourced higher-status groups. By comparison, elite-status individuals have limited motivation to build friendships across diversity to access resources. Situated between the elite and the lowest-status groups, the middle-status group members with moderate levels of resource endowment have moderate levels of motivation to seek friendship ties across diversity. Overall, as group status increases, motivation to seek resources by forming friendships across diversity declines.

In sum, we argue that the motivations driving homophily vary as a function of group status. Specifically, elites exhibit closure within homophilous networks due to their limited motivation to develop friendships across diversity for purposes of accessing resources. In contrast, lowest-status individuals, who are highly motivated to build friendships across diversity with higher-status others, may avoid doing so because they experience high toxicity in interactions across diversity (McCord et al., 2018). Our overall proposition, depicted in Figure 1, argues that for middle-status groups, pressures toward homophily are weaker in that both the closure dynamic driving homophily among elites and the toxicity avoidance driving homophily among the lowest-status group are less pronounced for middle-status individuals.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Conceptual relationship between social status and homophily in friendship ties

HOMOPHILY AMONG ELITE, MIDDLE, AND LOWEST-STATUS GROUPS

One of the research advantages of the Indian caste context is the fact that caste groups reflect three distinct status levels: the elite-castes, the middle-castes (so-called Other Backward Classes or OBCs), and the lowest-level Scheduled Caste (SC)/ Scheduled Tribe (ST) group. Legal enactments and the policy of affirmative action (AA) codify these status distinctions, formalizing the three levels of the Indian caste status hierarchy. Studying this rank-ordered, stable, three-tiered status hierarchy provides us with an opportunity to test whether homophily is weaker for the middle-status group as compared to the elite and the lowest-status groups.

The elite group is most well-endowed with cultural, economic, and social capital (Bapuji and Chrispal, 2020; Pandey and Pandey, 2018; Zacharias and Vakulabharanam, 2011). Elite status also confers non-material resources such as religion-based assessments of purity/pollution (Davidson, 2018; Vaid, 2014). Although elite group members experience low toxicity-linked barriers to ties across diversity, as highly resourced individuals, elites have little motivation to seek friendship ties across diversity to access resources. To the extent that elite group members desire to enhance their resource access through networking, they are most likely to connect with other elites because they are the most resource-endowed. As a result, the self-interested behavior of elite individuals generates high levels of homophilous closure within the elite group. Elite closure is also strengthened through their common interest in opposing AA quotas reserving educational and occupational positions for members of the middle- and lowest-caste groups (Kumar, 1992). The ease of forming friendship ties across similarity of food preferences, background similarly, shared habits (e.g., Bourdieu, 1986), and shared experiences of religious rituals further strengthen homophilous closure among elites.

Situated at the lowest-level of the three-tiered Indian status hierarchy are the Scheduled Caste (SC) and Scheduled Tribe (ST) group that includes the Untouchables^[¹^] (Dalits) and the marginalized indigenous communities (Adivasis) who were historically situated outside the human Varna hierarchy (Bapuji and Chrispal, 2020). The lowest-status SC/ST group has low resource endowments (Deshpande, 2011; Zacharias and Vakulabharanam, 2011) and members are motivated to seek ties across diversity, however, these lowest-status individuals face high levels of toxicity including prejudice (Cotterill et al., 2014) and unfriendly attitudes in social interaction (Pandey and Pandey, 2018). In academic institutions, this toxicity is blatant, overt, and strong such as students ‘not sitting on the same table or drinking water from the same tap as used by Dalits’ (Kumar, 2016, p. 12), as well as subtle including language nuances and body gestures (Kumar, 2016) reflecting a range of microaggressions. Consequently, lowest-caste group members face considerable barriers to forming friendship ties across diversity despite their strong motivation to develop such ties for resource access. Furthermore, the toxicity experienced by the lowest-status SC/ST group members pushes them towards stronger in-group identification, as rejection by other groups reinforces their separation and the non-permeability of group boundaries (e.g., Brown, 2010; Sidanius et al., 2008), pushing members of the lowest-status group into homophilous self-segregation (Bacharach et al., 2005; Mehra et al., 1998; Mollica et al., 2003).

Distinct from the elite group, middle-status group members are less endowed with resources than the elite status group and are still motivated to seek ties across diversity for resource access. The toxicity-driven self-segregation demonstrated by the lowest-status group is also weaker for the members of the middle-status group as they face fewer negative attitudes and less hostility than the lowest-status individuals do (Pandey and Pandey, 2018). Historically, the middle-status groups have faced a less toxic environment in terms of fewer restrictions on inter-dining and socio-economic exchange with elites compared to the members of the lowest-status group (Wiser, 1988), who have borne the brunt of caste-based toxicity. Consequently, the middle-status (OBC) group is likely to face fewer barriers in forming friendship ties with the members of the resource-endowed elite-status group.

In addition, the experiences of members of the middle-status group being lower in status compared to elites are likely to reduce their overall level of in-group bias (Bettencourt et al., 2001). Repeated experiences of being in the lower-status position reduce an individual’s sense of inherent superiority, thus making them more open to forming friendship ties with lowest-status individuals. The shared experience of toxicity, albeit more moderate for members of the middle-status group, makes them more empathetic towards the experiences of the lowest-status group and more open to forming friendships with them. Supporting this idea, past research in the context of resource sharing in rural communities shows that in contrast to elite leaders, middle-caste individuals holding community leadership positions create resource sharing arrangements that are more beneficial to the lowest-status castes, indicating greater empathy (Anderson, 2011). Consequently, we expect the middle-status OBC individuals to be more open to friendship ties across caste diversity. Generalizing from this illustration in the caste context to our broader theory, we expect:

Hypothesis 1: Compared to the elite and lowest-status group, homophily in friendship ties is weaker for the middle-status group.

LOWEST-STATUS GROUP NETWORKING BEHAVIOUR

We noted earlier that the lowest-status group faces a context where their group is low in resource endowments and subject to toxicity in social interactions across status diversity. This context shapes the distinctive networking behavior of the lowest-status group. More specifically, the context limits the agency (Gulati and Srivastava, 2014) of the lowest-status group (Chrispal et al., in press) to optimize their network ties in two important ways. First, because they are forced to seek ties across diversity for accessing resources, lowest-status members have to reach out to higher-status individuals (i.e., elite and middle-status) and this action involves substantial effort. By comparison, higher-status group members can simply reach out to homophilous ties to access resources (Ibarra, 1995). Second, due to potential toxicity in social interactions, many relationships may not come to fruition in terms that are desired by the lowest-status group. This additional challenge limits the choices of the lowest-status individuals, relative to higher-status groups, for selecting ties that will provide access to resources and persist over time. The resulting network patterns appear as homophilous withdrawal, even if the lowest-status individuals have made many efforts to build diverse ties. In order to understand why that pattern emerges, we develop hypotheses relating to the impact of constrained choices that lead to greater effort and lower likelihood of success on distinct patterns of tie seeking and tie maintenance for the lowest-status group.

Due to their limited resource endowments (Deshpande, 2011; Zacharias and Vakulabharanam, 2011) members of the lowest-status group experience substantial stress in organizations as the demands of their roles potentially exceed their access to resources for fulfilling the demands (Desmond, 2012). The disadvantage of low resource endowment creates an inherent motivation for lowest-status individuals to seek access to resources (Ibarra, 1992, 1995) and presses the lowest-status group to seek ties across diversity (Konrad et al., 2017). Ties across diversity may not, however, bring quick resource access as building trust across differences takes effort, time, and reciprocation. Facing these challenges, the lowest-status group may engage in relatively frequent initiation of new relationships, motivated by the expectation that among the many relationships sought, some will produce resource access (Desmond, 2012).

Furthermore, for at least some members of the lowest-status group (e.g., with weak in-group identification), the motivation to seek ties with higher-status groups may be driven by a desire to overcome toxicity arising from the salience of discrimination (Derks et al., 2015) or experienced discrimination (Derks et al., 2011). In addition, some lowest-status individuals may also be motivated to distance from their identity by seeking ties with middle- and elite-status individuals as a step toward social mobility and abandonment of an undesirable identity (Derks et al., 2011) or deflection of stigma (Pyke and Dang, 2003). Combining the mechanisms of resource seeking and stigma deflection, we expect lowest-status group members to actively seek ties across status diversity – yet have difficulty in building meaningful connections, with results that mimic a pattern of homophilous withdrawal in their ultimate realized networks.

Hypothesis 2: As compared to the higher-status group (elite and middle), members of the lowest-status group are more likely to initiate new friendships over time.

We propose that the low resource endowment of the lowest-status group members combined with the desire for stigma deflection drives these individuals to initiate a relatively large number of ties. However, even with that extra effort toward tie initiation, we predict that members of the lowest-status group will also have greater difficulty at maintaining relationships that will manifest in the greater loss of network connections over time. Ties across diversity may bring experiences of toxicity for the lowest-status group that create challenges in maintaining friendship ties and motivate exits from diverse relationships. Moreover, for members of the lowest-status group, accessing resources from ties across status diversity may be quite difficult, because it is hard for them to fulfill the reciprocity norm in munificent social exchanges with higher-status individuals (Shore et al., 2009). For this reason, members of higher-status group have limited motivation to maintain ties with resource-seeking members of the lowest-status group.

In contrast to the lowest-status group, the middle-status and elite groups will have less difficulty in maintaining relationships as manifested in fewer lost network connections over time. Higher-status group members are sought by lower-status groups (Ibarra, 1995; Konrad et al., 2017) for resource access and hence, higher-status individuals have a wide variety of choices in selecting network ties. Using the principle of minimal effort (Ibarra, 2016), the higher-status group can choose, amongst the many possibilities on offer to them, ties across diversity that they think will be most valuable and likely to last. Bringing the arguments together, we propose:

Hypothesis 3: As compared to higher-status group (elite and middle), members of the lowest-status group are less likely to maintain friendship ties over time.

METHODS

Data and Sample

We collected data from five section networks constituting an incoming cohort of 459 students pursuing an MBA equivalent program at a management school in India. A ‘section’ is a set of students assigned to the same classroom when they enter the program so that they take all of their foundation courses together as a group during their first year of study. Following prior studies (Konrad et al., 2017), we collected data on friendship networks within sections using the roster method (Wasserman and Faust, 1994), which helps to overcome the problem of recall purely based on memory (Marsden, 1990). The school had implemented AA quotas, thus ensuring representation of all the three status groups in each section network. The network size for each of the five networks was similar, ranging from 90 to 93. Participants were provided incentives including participation fees and dinner coupons (approximately US$20) for each round of data collection. Overall participation rates for both rounds of social network data collection were very high and comparable across the sections. Specifically, 422 out of 459 eligible participants provided usable information (92 per cent response rate). For the dyadic analysis testing hypothesis 1, we combined all the sections together treating ties across the sections as missing (illustrated in Figure 2). Specifically, as friendship data were collected for within section ties only, this resulted in about 80 per cent structural missing data, i.e., matrix cells treated as missing because we did not collect data on friendships formed between members of different sections.^[²^]

The data collection schedule was designed keeping in mind that caste is not immediately visible but revealed through social cues and interaction. To give sufficient time for status group identities to be revealed, we collected our first round of friendship network data eight weeks (Time T1) after the beginning of the program. The second round of network data collection took place approximately four months later (Time T2). Consistent with the pattern of distribution in most publically funded business schools and academic institutions in India, both females (Financial Times Global MBA Rankings, 2017,^[³^] network percentage 10.4, range 8.9 to 11.5 per cent) and Muslims (nationally 14.23 per cent of the population, network percentage 2.2 per cent, range 0 to 5.6 per cent) were highly underrepresented in our sample. While we controlled for gender, given the relative rarity of Muslims, we did not include religion in the analyses.

Measures

Strength of friendship tie

For friendship ties, ‘0’ indicates don’t know, and, ‘1’, ‘2’ and ‘3’ indicate increasing levels of friendship, with ‘3’ being a close friendship. In our Multiple Regression Quadratic Assignment Procedures (MRQAP) and Stochastic Actor-Oriented Model (SAOM) analyses described in greater detail below, we used the directional strength of friendship ties. For the MRQAP analysis we used the tie values (1 through 3) as the dependent variable, whereas for the SAOM analysis that uses the presence or absence of ties, we used ‘3’ for dichotomizing network ties, focusing our analysis on close friendships.

Caste group similarity

Participants were asked to provide data on the caste group to which they belonged (i.e., Scheduled Caste, Scheduled Tribe, OBC, General Category, and Other). We used this participant provided data to assess their membership in the elite, middle (OBC), and the lowest-status (SC/ST) group. We created an elite similarity matrix by coding a cell as ‘1’ if both individuals were elite, and ‘0’ otherwise. Similar to the treatment for the friendship data, we treated the structural missing data as missing. Similarly, we created a middle similarity matrix by coding a cell as ‘1’ if both were middle-status, and ‘0’ otherwise. We followed the same process for creating the lowest-status similarity matrix. In total, 61 per cent of the participants belonged to the elite-status group, 22 per cent belonged to the middle-status (OBC), and 17 per cent belonged to the lowest-status (SC/ST).

Control variables

We controlled for the similarity of work experience and similarity of undergraduate academic discipline at the dyadic level. The work experience data were highly positively skewed, and before creating the similarity matrix we normalized the data using a log transformation [log (1 + X)], where X is months of work experience. In addition, we also controlled for gender similarity. In both these matrices, we again treated the structural missing data as missing.

Dummy matrices for the sections (networks)

We also created four dummy matrices to control for possible section effects. Network 1 dummy matrix was created by replacing all the cells in Network 1 in the combined matrix with ‘1’ and all the cells in Network 2, 3, 4, and 5 with ‘0’. The structural missing remained as missing. Similarly, Network 2 dummy matrix was created by replacing all the cells in Network 2 in the combined matrix with ‘1’ and all the cells in Network 1, 3, 4, and 5 with ‘0’. We repeated this process for Network 3 dummy and Network 4 dummy matrices. Network 5 is our reference category for these dummy matrices.

ANALYSIS AND RESULTS

To better illuminate the underlying data, we present the ties within and between the status groups in Tables I and II respectively for T1 and T2. These tables depict combined data from all five sections. Tie strength in these Tables refers to the closeness of friendships where higher numbers indicates stronger ties (i.e., 3 for close friendship), and lower numbers indicate weaker ties (i.e., 2 indicates casual friend, 1 indicates acquaintance). In each cell, the first number represents the raw score of ties between the respective status groups, whereas the second number in parentheses indicates the fractional percentage of maximum possible ties in the respective status group.

Table I. Friendship ties distribution within and across the status groups at T1

	Elite	Middle (OBC)	Lowest (SC/ST)
Tie strength = 3 (close friend)
Elite	899 (6.99%)	297 (6.31%)	234 (6.19%)
Middle (OBC)	289 (6.14%)	106 (6.55%)	63 (4.62%)
Lowest (SC/ST)	253 (6.70%)	105 (7.70%)	95 (9.05%)
Tie strength = 2 (casual friend)
Elite	1877 (14.59%)	605 (12.85%)	476 (12.60%)
Middle (OBC)	734 (15.59%)	250 (15.45%)	196 (14.37%)
Lowest (SC/ST)	581 (15.38%)	184 (13.49%)	178 (16.95%)
Tie strength = 1 (acquaintance)
Elite	2886 (22.43%)	917 (19.47%)	735 (19.45%)
Middle (OBC)	1052 (22.34%)	357 (22.06%)	301 (22.07%)
Lowest (SC/ST)	749 (19.83%)	271 (19.87%)	193 (18.38%)

Note: For tie strength, a higher number from 1 to 3 indicates a stronger tie. Tie strength = 3 represents close friendship ties
In each cell, the first number represents the raw number of ties between the respective status groups, whereas the second number, (i.e., the number in parentheses) indicates the fractional percentage of maximum possible ties in the respective status groups.
For each of the three types of ties, the numbers on the diagonal represent within group ties. The diagonal numbers can be compared with the other two numbers in the same row for outside group ties. For example, for close friendship ties (Tie strength = 3) elite group has a fractional percentage of within status ties (6.99%) that is a greater than their fractional percentage ties with the middle-status (6.31%) and the lowest-status group (6.19%).

Table II. Friendship ties distribution within and across the status groups at T2

	Elite	Middle (OBC)	Lowest (SC/ST)
Tie strength = 3 (close friend)
Elite	1195 (9.29%)	332 (7.05%)	272 (7.20%)
Middle (OBC)	354 (7.52%)	105 (6.49%)	104 (7.62%)
Lowest (SC/ST)	291 (7.70%)	115 (8.43%)	97 (9.24%)
Tie strength = 2 (casual friend)
Elite	2451 (19.05%)	830 (17.63%)	630 (16.68%)
Middle (OBC)	1020 (21.66%)	265 (16.38%)	275 (20.16%)
Lowest (SC/ST)	670 (17.73%)	230 (16.86%)	234 (22.29%)
Tie strength = 1 (acquaintance)
Elite	4075 (31.68%)	1524 (32.36%)	1122 (29.70%)
Middle (OBC)	1560 (33.13%)	464 (28.68%)	448 (32.84%)
Lowest (SC/ST)	1101 (29.14%)	427 (31.31%)	294 (28%)

Note: For tie strength, a higher number from 1 to 3 indicates a stronger tie. Tie strength = 3 represents close friendship ties
In each cell, the first number represents the raw number of ties between the respective status groups, whereas the second number, (i.e., the number in parentheses) indicates the fractional percentage of maximum possible ties in the respective status groups.
For each of the three types of ties, the numbers on the diagonal represent within group ties. The diagonal numbers can be compared with the other two numbers in the same row for outside group ties. For example, for close friendship ties (Tie strength = 3) middle-status has a fractional percentage of within status ties (6.49%) that is smaller than their fractional percentage ties with the elite (7.52%) and the lowest-status group (7.62%).

The data in Tables I and II indicates initial support for our basic proposition that the middle-status group will show less homophily in their networks than either the elite or lowest-status groups. Examining close friendships (Tie strength = 3) at T1, we find evidence of homophily for all three status groups. For example, the middle-status group shows the highest fractional percentage of strong ties with other members of the middle-status group (6.55 per cent) and a lower fractional percentage of ties with the elite (6.14 per cent) and lowest-status groups (4.62 per cent). The elite and the lowest-status groups also show homophily among close friendships. Interestingly, the middle-status group is not homophilous for either casual friendship ties or acquaintances. The lowest-status SC/ST group is homophilous in their casual friendship ties but not in their acquaintances. Consistent with our theorizing that resource access is an important motivator of networking behavior, elites were the only group to show homophily across all levels of tie strength.

At T2, the middle-status group is not homophilous even for close friendship ties as the fractional percentage for within status ties for the middle-status group (6.49 per cent) is lower than fractional percentage for ties with elite (7.52 per cent) and lowest-status group (7.62 per cent). The middle-status group is again not homophilous for casual friendship ties (tie strength = 2) while the elites and the lowest-status groups again show homophily in their close friendship ties and casual friendships. Interestingly, none of the three status groups are homophilous for acquaintance ties at T2. In general, at T1, the middle-status group shows less homophily compared to the elite and the lowest- status groups. This effect continued over time and appears to be more pronounced at T2.

We used the multiple regression quadratic assignment procedures (MRQAP) implemented in UCINET for testing Hypothesis 1. MRQAP is an analytical technique equivalent to multiple regression analysis that overcomes violations of data independence and allows the use of valued networks, enabling us to use the strength of friendship ties as the dependent variable.^[⁴^] Table III provides the correlation matrix. As expected, gender similarity was significantly and positively correlated to the strength of friendship ties at both T1 (r = 0.05) and T2 (r = 0.04). Consistent with our theorizing, shared middle-status similarity was significantly and negatively correlated to the strength of friendship ties at both T1 and T2.

Table III. QAP correlations (combined network)

		1	2	3	4	5	6	7	8	9	10	11
1	Work Experience similarity
2	Discipline similarity	0.00**
3	Gender similarity	0.00 *	0.14**
4	Status (Elite) similarity	0.00	−0.03	−0.05
5	Status (Middle) similarity	0.00	0.06*	0.04^†	−0.17***
6	Status (Lowest) similarity	0.00**	−0.04	−0.01	−0.13***	−0.04**
7	Friendship1	0.00	0.01	0.05***	0.05***	−0.02^†	0.01
8	Friendship2	0.02^†	−0.01	0.04**	0.08***	−0.04**	0.01	0.53***
9	Network 1	0.00^†	0.07	−0.01	−0.04	0.02	0.01	0.00	0.02
10	Network 2	0.00	−0.07	−0.00	−0.02	0.01	0.00	0.00	0.01	−0.27***
11	Network 3	0.00	−0.05	0.03	0.02	−0.03	0.02	0.01	−0.02	−0.24***	−0.26***
12	Network 4	0.00	0.03	−0.04	−0.02	0.01	0.01	−0.02^†	−0.05**	−0.25***	−0.27***	−0.24***

Note: Number of Permutations = 5000.
^† p < 0.1;
* p < 0.05;
** p < 0.01;
*** p < 0.001.

H1 suggested that compared to the elite and the lowest-status group, homophily in friendship ties is weaker for the middle-status group. To test this hypothesis, we created similarity matrices for each of the three groups. The results are presented in Table IV. Model T1-1 shows that among the control variables, only gender similarity was significantly related to friendship at T1 (β = 0.04, p < 0.001). Model T1-2 adds status similarity to the T1 equation and indicates that consistent with our reasoning, homophily effects were significant and positive for the elite (β = 0.05, p < 0.001) and lowest-status (β = 0.02, p < 0.01) groups. In line with our rationale, status similarity was a significant and negative predictor of friendship for the middle-status group at T1 (β = −0.01, p < 0.05) indicating middle-status group heterophily. We repeated the analysis at T2 and the results, shown in Model T2-2, were again consistent with our reasoning, showing positive and significant status similarity effects indicating homophily for the elite (β = 0.08, p < 0.001) and the lowest-status (β = 0.02, p < 0.001) groups, with a negative and significant status similarity effect for the middle-status group (β = −0.03, p < 0.001) indicating middle-status group heterophily.^[⁵^]

Table IV. (Combined Networks): Effect of status group similarity on friendship –MRQAP (Unit of Analysis –Dyad)

	Model T1-1 Control	Model T1-2 Hypothesis 1	Model T2-1 Control	Model T2-2 Hypothesis 1
	DV – Friendship T1		DV Friendship T2
Work Experience similarity	0.00	0.00	0.02*	0.02*
Discipline similarity	0.01	0.01	−0.01	−0.01
Gender similarity	0.04***	0.05***	0.04***	0.04***
Network 1	−0.01	−0.00	−0.02	−0.02
Network 2	−0.01	−0.01	−0.03^*	−0.03^*
Network 3	−0.00	−0.00	−0.06**	−0.06***
Network 4	−0.02	−0.02	−0.07***	−0.07***
Status (Elite) similarity		0.05***		0.08***
Status (Middle) similarity		−0.01*		−0.03***
Status (Lowest) similarity		0.02**		0.02***
R²	0.002***	0.006***	0.006***	0.014***

Note: Number of Permutations = 5000. N(T1) = 34432 (Dyads); N(T2) = 34461 (Dyads).
Network 1 to Network 4 are the dummy matrices for the sections 1 to 4. Network 5 (section 5) is the reference matrix.
^† p < 0.1;
* p < 0.05;
** p < 0.01;
*** p < 0.001.

NETWORK DYNAMICS ANALYSIS AND RESULTS

To test the hypotheses that the lowest-status group will initiate new friendship ties (H2) and have a lower likelihood of maintaining friendship ties over time (H3), we used temporal analysis of the Stochastic Actor-Oriented Model (SAOM) implemented in Simulation Investigation for Empirical Network Analysis (SIENA) software and its R version 1.1-288 R-SIENA (Ripley et al., 2018; Snijders, 1996, 2001).^[⁶^] Because the network and actor attribute effects on this network objective function are simultaneously estimated, it is possible to make strong statistical inferences regarding the effects of attributes (status group in our case) on network evolution (friendship networks in our case) controlling for structural network effects (Kalish et al., 2015; Steglich et al., 2010). The SAOM analysis can only include a two-category attribute. So, consistent with our motivation to contrast the lowest-status group with the other groups in the status hierarchy; we combined elite and middle status together to enable us to compare the lowest-status group with this combined higher-status group.

We conducted our analysis using R-SIENA software version 1.1-290 (Ripley et al., 2016). We were unable to run R-SIENA on the combined matrix for the five sections, with a high level of missing structural data, as R-SIENA results do not converge if the proportion of missing data is high. Hence, we report results for the R-SIENA analysis for each section separately.^[⁷^]

We tested the five evolving networks labelled as Net-1-5. Table V provides the abridged results of SAOM analysis with the network and attributes/behaviour controls (for complete results see supplementary Table SI), presenting only the parameters relevant to H2 and H3. H2 states that as compared to the higher-status group (elite and middle), members of the lowest-status group are more likely to initiate new friendships over time. Because the higher-status group is coded as ‘1’ with lowest-status as ‘0’, the significant negative coefficients in the top row of Table V support this prediction. Specifically, H2 is supported for Net-1 (parameter = −18.88, p < 0.001), Net-2 (parameter = −15.51, p = 0.002), Net-4^[⁸^] (parameter = −20.62, p = 0.005), and marginally for Net-3 (parameter = −11.60, p = 0.09); but not for Net-5^[⁹^] (parameter = −9.85, p = 0.22). In sum, H2 was supported in four of the five network analyses.

Table V. (Abridged): SIENA model results-effect of status group on friendship network dynamics (Elite/OBC = 1, SC/ST = 0)

Effect Name	Positive Effect meaning	Net-1	Net-2	Net-3	Net-4^a	Net-5^b
Status Group Effects of Interest (H2 & H3): Higher-Status Group (*Elite + Middle* )
Status-ego creation (H2)	‘Higher’ status ego are more likely to create a friendship tie with a new alter	−18.88 (2.75)	−15.51 (5.03)	−11.60 (6.77)	−20.62 (7.41)	−9.85 (8.03)
Status-ego creation (H2)		p < 0.001	p = 0.002	p = 0.09	p = 0.005	p = 0.22
Status-ego endowment (H3)	‘Higher’ status ego are more likely to maintain a friendship tie with an existing alter	19.34 (2.75)	14.49 (4.34)	11.49 (6.91)	20.50 (7.39)	9.85 (8.04)
Status-ego endowment (H3)		p < 0.001	p < 0.001	p = 0.096	p = 0.006	p = 0.22

Note: Node filled with dots () represents any individual, irrespective of his/her status group). White node (○) represents lower-status member; black node (●) represents higher-status member. Black arrow represents either existing ties or ties that were maintained between T1 and T2. Red dashed arrow indicates the ties that were created between T1 and T2.
The following sets of controls were used for the SIENA Model:
i) Network Effects: Rate function, Outdegree, Reciprocity, Transitive triplets, 3-Cycles, In-degree popularity, Out-degree popularity, and Structural equivalence.
ii) Gender Homophily Effect.
iii) Status Group Effects: Status-alter creation, Status-alter-endowment, Same-status group creation, and Same-status group endowment.
Unabridged Table V results are available in Table S-I in the supplemental material.
^a Net-4 results should be read with caution. See Table SI in the supplemental material.
^b Net-5 results should also be read with caution. See Table SI in the supplemental material.

H3 states that as compared to the higher-status group (elite and middle), members of the lowest-status group are less likely to maintain friendship ties over time. The significant positive parameters in the bottom row of Table V support H3. Recall that the higher-status group is coded as ‘1’ with lowest-status as ‘0’, so the significant positive coefficients support the prediction that the higher-status group (elite and middle) is more likely to maintain friendship ties over time (or as stated in H3, the lowest-status group is less likely to maintain friendship ties over time). H3 is supported for Net-1 (parameter = 19.34, p < 0.001), Net-2 (parameter = 14.49, p < 0.001), Net-4^[¹⁰^] (parameter = 20.50, p = 0.006), and marginally Net-3 (parameter = 11.49, p = 0.096); but not for Net-5^[¹¹^] (parameter = 9.85, p = 0.22).^[¹²^]

GENERAL DISCUSSION

Globally, organizations are becoming more diverse through increased inclusion of historically marginalized groups in all types of positions. However, within organizations, informal networks among members, which are important to organizational functioning, show a clear tendency toward homophily and away from building connections across diversity. These homophily dynamics limit information and resource exchanges across the organization because many individuals are disconnected due to gaps/lack of connections across diversity. We theorize that status distinctions between identity groups constitute strong barriers to network diversity which are particularly damaging to the efforts of lowest-status groups to build and maintain network ties, ultimately resulting in homophilous withdrawal network patterns. We argue further that middle-status groups are likely to show the greatest heterophily.

Past research examining the linkages between status and homophilous friendship ties have produced mixed findings of either homophilous elite closure or lowest-status group self-segregation (Konrad et al., 2017; Mollica et al., 2003). The non-Western context of India with a well-recognized, stable, and legally codified three-tiered status group hierarchy provided us with the opportunity to compare a middle-status group with both the elite and the lowest-status groups. Findings from our research suggest that middle-status group members are the most likely to build ties across status diversity. The key insight we offer of middle-status ambivalence towards homophily in friendship ties is an important theoretical advancement to our understanding of the relationship between status and homophily in networks, and one that can be built upon to help mitigate the negative organizational and societal impacts of homophily.

While we see evidence of greater homophily among both the elite and the lowest-status groups, we find that the strongest homophily is amongst the elite group, supporting arguments of elite closure (Konrad et al., 2017; Tomaskovic-Devey, 1993). This pattern of findings held at both T1 and T2 with elite closure effects strengthening over time. While beyond our data’s testing capabilities, our conceptual reasoning suggests that the greater the resource endowment advantage of the elite, the stronger is the expected elite closure into homophilous ties. This dynamic warrants further research on the role of organizational and public policies that support the wider distribution of resource endowments (Riaz, 2015). Elite closure represents one manifestation of high status in-group behavior that bolsters the existing system within the system justification perspective (Jost et al., 2004), and future research could empirically examine the interplay of individual attitudes and networking behaviour across status groups for its contribution to system justification (Jost and Banaji, 1994) and status-hierarchy maintenance (Sidanius and Pratto, 1999).

It is noteworthy that within the three types of ties (close friendships, casual friendships, and acquaintances) the general pattern (Tables I and II) suggests that the greatest heterophily in close friendship ties emerges at T2 where the middle-status group builds connections to both the elite and the lowest-status groups. For instance, for close friendship ties at T2, middle-lowest (7.62 per cent) and middle-elite (7.52 per cent) ties are both greater than middle-middle ties (6.49 per cent). This pattern also holds for their T2 casual friendships and acquaintances. It is also interesting that for acquaintance ties, all three groups exhibit heterophilous tendencies at T2. This is potentially promising if such acquaintance ties can be strengthened through structural interventions. It appears time has a complex influence on patterns of heterophily, and future research can examine the evolution of network ties for a longer time period to unravel further differences in network development across status groups and types of relationships. Along these lines, policymakers should also consider the important role of societal institutions (e.g., universities) as facilitators of social interaction for the prolonged periods of time required for the emergence of friendship ties across diversity, and perhaps facilitated by middle-status group members.

Our findings of homophily within the elite group in a context where status group identities may be initially visibly indistinguishable point to deeply embedded social mechanisms that drive elite homophily. But the finding of middle-status heterophily also suggests porous status group boundaries that were previously missed in simplified two-tier status models. Organization leaders seeking to enrich the diversity of the informal networks in their organizations should take note of our novel middle-status heterophily findings in the Indian context of near universal homophilous tie formation (McPherson et al., 2001).

Our findings of heterophily among members of the middle-status group point to the important potential role of the middle-status group in bridging status group divides, which is likely to enhance collaboration, social inclusion, and organizational functioning. Our conceptual development leveraging the two mechanisms of resource endowments and exposure to toxicity across status groups suggests that findings of middle-status heterophily are not a result of contextual specificity, and are a more generalized and broadly applicable theoretical dynamic. Nevertheless, testing the generalizability of our findings in other contexts is important, perhaps with more direct measures of resource endowments and toxicity.

An interesting direction for future research is to understand the extent to which individuals within the different status groups identify with their specific identity category as well as whether they strengthen or weaken their level of in-group identification as a result of building friendships across status diversity. Prior authors have identified what they term a ‘queen bee effect’ among lower-status identity groups whereby individual members may work to distance themselves from the group in order to shed stigma and disadvantage (Derks et al., 2011). For example, prior research among Surinamese Hindustani employees in Netherlands, with roots in a history of indentured servitude, exhibit weaker identification with their own group when primed about ethnic discrimination (Derks et al., 2015). Our theoretical logic underlying middle-status heterophily raises interesting questions about whether, and the extent to which, middle-status individuals strengthen or weaken their personal sense of identification with their middle-status group as well as the implications of that dynamic for organizational functioning and collective societal change.

Although we see the evidence of homophily in both the elite and lowest-status groups, the two groups are very distinct. Specifically, elites show evidence of stronger self-segregation (at both T1 and T2), while lowest-status individuals show evidence of weaker self-segregation. Further, lowest-status group individuals demonstrate greater efforts to build friendships with higher-status individuals but a lower likelihood of maintaining friendship ties over time. Merely observing the segregated pattern of lowest-status group friendships fails to show the efforts of the lowest-status group. Our temporal analysis demonstrates this dynamic as we find that the lowest-status group makes more attempts to build ties (H2) and has a lower likelihood of maintaining friendship ties over time (H3) which ultimately results in homophilous withdrawal network patterns. Such a dynamic helps to integrate the two prevalent perspectives of elite closure and homophilous withdrawal, and link them jointly to a culture of toxicity.

A key feature of the friendship ties of the lowest-status individuals appears to be instability as manifested in the lower likelihood of maintaining friendship ties over time. Our findings reveal that the lowest-status group, compared to the higher-status group, lost more connections to people whom they believed were friends at an earlier point in time. Challenges to maintenance of friendship ties undermines the benefits of networks (Wolf and Moser, 2009) and greatly diminishes the value of networking activities as the hoped-for exchange of resources fails to materialize for members of the lowest-status group. Given lower resource endowments, the lowest-status group are also disadvantaged with constrained agency as they expend greater efforts in tie seeking and accessing resources that may be available to higher-status group members with much lower effort. These findings point to the additional challenges and the lack of a level playing field in the development of social networks for members of the lowest-status group, as their choices or network agency are exercised in much more limited and constrained ways.

The lowest-status group members in the Indian caste group hierarchy have faced ostracism and degradation for more than two thousand years, and the lowest-caste group in India is quite distant from elites in terms of power, influence, resources, and treatment. Future research may explore lowest-status group tie seeking behaviour in societies with less extreme group status differences to further establish the generalizability of our findings. An interesting line of future research could focus on understanding the multiple motivations of the lowest-status group for seeking ties across diversity, specifically, resource access, toxicity avoidance, and their potential interplay with stigma deflection (Pyke and Dang, 2003). Further, future research should simultaneously examine the challenges to tie maintenance among lowest-status group members across both visible (e.g., race, gender) and other non-visible distinctions or distinctions that do not need to be disclosed (e.g., sexual orientation, religious identity).

While we did not directly measure the mechanisms of toxicity and resource endowment, survey evidence from an academic institution in India supports our logic. Specifically, the survey revealed that the lowest-status SC/ST group students report the highest levels of negative attitudes from fellow students while the elites report the lowest levels of negative attitudes with the middle-status BC group placed in the middle (Pandey and Pandey, 2018). Further and consistent with Ridgeway (1993), the survey also revealed the relatively positive assessment of the abilities of elite status group members as well as their higher levels of resource endowments (Pandey and Pandey, 2018). Yet follow-on studies of middle-status networking behaviors could certainly test these dynamics in more detail to further refine the theoretical model.

Our observations also suggest that middle-status individuals often cross status group boundaries to form friendship ties. A recent comprehensive and nationally representative survey in India confirms that compared to the elite and lowest-status groups, middle-status individuals report more heterophily in close friendship ties (Pew Research Centre, 2021) greatly adding to the generalizability of our findings to other contexts in India. The reasons for these patterns may not be revealed in these data and observations, but we speculate that the middle-status group’s greater empathy from shared experiences (e.g., toxic exposure) and their relatively higher social status (than SC/ST group members) may ease middle-status actor tie formation across diversity in both directions.

Practical Implications

The relative position of the lowest-status group members in India brings to the fore the urgency of policy initiatives in a wide variety of contexts to alleviate this dire situation. Social psychological research suggests that biases and prejudices pose barriers that diminish the quality of interpersonal interaction and reduce the likelihood of building a friendship (Esses and Dovidio, 2002). Hence, organizations could act by implementing initiatives designed to reduce these interpersonal biases. Research shows that prejudice reduction initiatives based upon habit-breaking can be effective for reducing interpersonal bias (Devine et al., 2012). Organizations could undertake activities such as perspective taking, whereby organizational members think about and discuss how the organizational experience might be different if they belonged to a different status group. This type of lesson in empathy has been linked to prejudice reduction (Levy and Green, 2009), which results in improved interaction across diversity and increased likelihood of developing lasting positive relationships, including friendships (Ragins and Ehrhardt, in press). Furthermore, serious attention is required for creating opportunities for the mixing of people across status divides. Intergroup contact in contexts of equality and interdependence creates opportunities for friendship formation on the basis of knowledge about individuals’ values and capabilities and leads to reduction in intergroup prejudice (Pettigrew and Tropp, 2006).

Given the importance of diversity in friendship ties within organizations, middle-status individuals appear to provide a necessary path to build connections between those on the status extremes. Connecting disparate aspects of the organization creates organizational value beyond what is possible if these networks remained completely disconnected from one another. Middle-status group members with ties across diversity that provide a uniquely broad and varied source of information and knowledge (e.g., Qureshi et al., 2018a) and who serve as connectors linking otherwise separate networks also may become role models for other actors in a broad variety of societal contexts.

While the caste-based status hierarchy of India is frequently cited, similar systems persist in many countries in South Asia (e.g., Zulfiqar and Prasad, 2021) and globally. Our findings have implications for many societies outside of India as well. Many societies carry the historical legacy of institutions that generate a three-tiered hierarchy among identity groups. For example, Australia, Canada and the USA share a legacy of British colonialism. In all three, the indigenous populations were so decimated by colonialism that their contemporary descendants constitute the poorest (economically) and lowest-status identity groups, similar to Adivasis or scheduled tribes (ST) in India. Domestic Black populations stigmatized by the history of slavery also have uniquely negative experiences (Wilkerson, 2020), creating another lowest-status group, similar to the Dalits or scheduled castes (SC). Immigrants of color from East and South Asia constitute substantial minority populations in these societies, who generally find themselves in the middle (Bell et al., 2014) such that they are devalued compared to the predominant White population but considered more positively than either the Black or indigenous groups.

Caste-based systems also need to be examined in places with high proportions of people of Indian origin such as Mauritius, Guyana, Fiji, Trinidad and Tobago, and Qatar. Caste systems traverse national boundaries as immigrants embedded within the system migrate to new locations (e.g., South Asians immigrants to Europe and North America) and carry the practices of caste discrimination to new contexts (Rai, 2021). Understanding the network mechanisms facilitating friendship ties across diversity offers the potential for knowledge sharing of processes, public policies, and organizational practices across contexts.

Another important factor to consider for practical initiatives resulting from this research is the attitudes of the organizational leaders who are selecting and guiding organizational interactions. For instance, returning to our research context, although nationally the SC/ST group comprises about 25.2 per cent of the population (India Census, 2011), less that 1 per cent of the faculty at a group of publically-funded Indian management institutions belong to the lowest SC/ST group (Joshi and Malghan, 2017). Most faculty members now are members of the elite castes and many have negative attitudes toward AA policies in India, based in part on self-interested worry about their own children’s outcomes. But if organizational leaders have negative views of AA beneficiaries, then they might (unintentionally or intentionally) treat people from the middle and lowest-status in ways that diminish their performance, as in the golem effect (Babad et al., 1982). With fewer opportunities to excel, lowest and middle-status people become less attractive as potential friends. Therefore, it is imperative for organizations to develop initiatives supporting empowerment across the entire status spectrum. Furthermore, hiring organizational leaders who represent the entire societal status spectrum is a longer-term solution for providing all organizational members with role models and opportunities to excel, thereby enhancing the benefits of network diversity for all.

Active engagement of organizational leaders in reducing toxicity is another way to advance the development of friendship ties across diversity. The middle-status groups provide a good avenue for the start of such efforts. Identity formation and categorization processes across shared interests and common goals may offer additional opportunities for leaders to encourage ties across new bases of identity. Efforts to encourage and highlight common forms of identity around the use of reason and potential for shared ideological positions (e.g., the role of capitalism and free markets) may provide new and common value-based grounds for friendships across status group divisions (Harrison et al., 1998, 2002; Konrad et al., 2017).

Limitations

We note some limitations of our research. First, our analysis was limited to the three-tiered status hierarchy in India, and we do not have data on the 4000 jatis or castes that are the underlying basis for the status hierarchy. Other scholars (Davidson, 2018) have examined status group homophily in the context of villages where finer grained caste data can be more plausibly gathered due to the more limited number of jatis in a particular localized context, but even in such limited settings it would be difficult to incorporate the fine-grained caste data in network analysis. Considering that there are several thousand castes at the national level, rank ordering them and examining caste homophily in a network consisting of 90 people is not feasible. Further, our research may be biased due to omitted variables such as geographic region, first language, and other similar mechanisms (Wimmer and Lewis, 2010) which are possible mediators of the observed homophily. But we would argue that such clustering bases for homophily are also status based.

CONCLUSION

Diversity among close friendships is important for the sharing of scarce, tacit, and valuable information crucial to the functioning of organizations and society. Diverse networks increase the variety and flow of information, knowledge, and perspectives that bring benefits to both individuals and organizations and enhance creativity in problem solving. Hence, successful efforts aimed at integrating and diversifying social networks are beneficial to organizations, the business sector, and lead to a more inclusive overall society. Our research can contribute to a theoretically informed identification and reduction of the barriers to inclusion for the lowest-status group and to understanding the role of middle-status group member networking behaviours as the key to mitigating status divides.

ACKNOWLEDGEMENTS

We would like to thank the consulting editor Li-Qun Wei, the three anonymous reviewers, and the anonymous second editor for a highly developmental review process. This paper draws on research that is supported by funding from the Social Sciences and Humanities Research Council (SSHRC) of Canada, SSHRC Grant No. 410-2011-0842. The fifth author is grateful for support from the Corus Entertainment Chair in Women in Management at the Ivey Business School. We would like to thank Priyanka Pandey (World Bank) and Magsaysay Awardee Sandeep Pandey for sharing their valuable insights on the Indian caste hierarchy. An earlier version of the paper was presented at the 2016 meeting of the Academy of Management in Anaheim, California.

Notes

[1] Untouchability was formally abolished when the Indian Constitution was adopted on 26 November 1949 and came into effect on 26 January 1950.

[2] The structural missing data resulted in very long computation time for each analysis. For stochastic actor-oriented models (SAOM) to test H2 and H3, we analyzed the data for each of these five networks separately, because SAOM models cannot handle high levels of structural missing data.

[3] For example, the gender percentage in MBA equivalent program at the three publically-funded schools that ranked in the top 100 MBA schools in the Financial Times (2017) rankings was 14%, 22%, and 15% respectively.

[4] In our analysis, we specifically used the Double-Dekker-Semi-Partialing MRQAP method as it avoids any multi-collinearity issues (Dekker et al., 2007). See additional note following supplemental material Table SIII.

[5] See additional ANOVA results following supplemental material Table SIII.

[6] SAOM is capable of incorporating network matrices (dyadic data) at more than one time point to test creation of new ties or maintenance of existing ties. See supplemental material for additional details.

[7] See supplementary Table SI.

[8] Net-4 results should be read with caution, see supplementary Tables SI and SII.

[9] Net-5 results should also be read with caution, see supplementary Tables SI and SII.

[10] Net-4 results should be read with caution, see supplementary Tables SI and SII

[11] Net-5 results should also be read with caution, see supplementary Tables SI and SII.

[12] As a robustness check we tested an alternative specification that combined the lowest-status group with the middle-status group and compared this combined group to the elite-status group and this analysis provided a similar pattern of results. The results for this alternate analysis are available in the supplementary files, Table SII. The disaggregated data of ties that were maintained, dissolved, and created between T1 and T2 is available in supplementary Table SIII.

Supporting Information

REFERENCES

Anderson, S. (2011). ‘Caste as an impediment to trade’. American Economic Journal: Applied Economics, 3, 239–63.
10.1257/app.3.1.239
Web of Science® Google Scholar
Babad, E. Y., Inbar, J. and Rosenthal, R. (1982). ‘Pygmalion, Galatea, and the Golem: Investigations of biased and unbiased teachers’. Journal of Educational Psychology, 74, 459–74.
10.1037/0022-0663.74.4.459
Web of Science® Google Scholar
Bacharach, S. B., Bamberger, P. A. and Vashdi, D. (2005). ‘Diversity and homophily at work: Supportive relations among White and African-American peers’. Academy of Management Journal, 48, 619–44.
10.5465/AMJ.2005.17843942
Web of Science® Google Scholar
Baldwin, T. T., Bedell, M. D. and Johnson, J. L. (1997). ‘The social fabric of a team-based M.B.A. program: Network effects on student satisfaction and performance’. Academy of Management Journal, 40, 1369–97.
10.2307/257037
Web of Science® Google Scholar
Bapuji, H. and Chrispal, S. (2020). ‘Understanding economic inequality through the lens of caste’. Journal of Business Ethics, 162, 533–51.
10.1007/s10551-018-3998-8
Web of Science® Google Scholar
Barling, J., Dupré, K. E. and Kelloway, E. K. (2009). ‘Predicting workplace aggression and violence’. Annual Review of Psychology, 60, 671–92.
10.1146/annurev.psych.60.110707.163629
PubMed Web of Science® Google Scholar
Bell, M. P., Marquardt, D. and Berry, D. P. (2014). ‘“Diversity”, immigration, and the new American multi-racial hierarchy’. Journal of Managerial Psychology, 29, 285–303.
10.1108/JMP-08-2012-0242
Web of Science® Google Scholar
Bettencourt, B. A., Dorr, N., Charlton, K. and Hume, D. L. (2001). ‘Status differences and in-group bias: A meta-analytic examination of the effects of status stability, status legitimacy, and group permeability’. Psychological Bulletin, 127, 520–42.
10.1037/0033-2909.127.4.520
CAS PubMed Web of Science® Google Scholar
Bourdieu, P. (1986). ‘ The forms of capital’. In J. G. Richardson (Ed.), Handbook of Theory and Research for the Sociology of Education. New York: Greenwood, 241–58.
Google Scholar
Bowling, N. A. and Beehr, T. A. (2006). ‘Workplace harassment from the victim's perspective: A theoretical model and meta-analysis’. Journal of Applied Psychology, 91, 998–1012.
10.1037/0021-9010.91.5.998
PubMed Web of Science® Google Scholar
Brown, R. (2010). Prejudice: Its Social Psychology. Oxford: Wiley-Blackwell.
Google Scholar
Chrispal, S., Bapuji, H. and Zietsma, C. (in press). ‘Caste and organization studies: Our silence makes us complicit’. Organization Studies, in press.
Google Scholar
Coleman, J. S. (1990). Foundations of Social Theory. Cambridge, MA: Belknap Press of Harvard University Press.
Web of Science® Google Scholar
Cotterill, S., Sidanius, J. S., Bhardwaj, A. and Kumar, V. (2014). ‘Ideological support for the Indian caste system: Social dominance orientation, right-wing authoritarianism and karma’. Journal of Social and Political Psychology, 2, 98–116.
10.5964/jspp.v2i1.171
Google Scholar
Davidson, T. (2018). ‘ Variation in caste homophily across villages and contexts in rural India’. doi: 10.31235/osf.io/j367u. (accessed 6 October 2019)
Google Scholar
Deitch, E. A., Barsky, A., Butz, R. M., Brief, A. P., Chan, S. and Bradley, J. C. (2003). ‘Subtle yet significant: The existence and impact of everyday racial discrimination in the workplace’. Human Relations, 56, 1299–324.
10.1177/00187267035611002
Web of Science® Google Scholar
Dekker, D., Krackhardt, D. and Snijders, T. A. B. (2007). ‘Sensitivity of MRQAP tests to collinearity and autocorrelation conditions’. Psychometrika, 72, 563–81.
10.1007/s11336-007-9016-1
PubMed Web of Science® Google Scholar
Derks, B., Ellemers, N., Van Laar, C. and de Groot, K. (2011). ‘Do sexist organizational cultures create the queen bee?’. British Journal of Social Psychology, 50, 519–35.
10.1348/014466610X525280
PubMed Web of Science® Google Scholar
Derks, B., van Laar, C., Ellemers, N. and Raghoe, G. (2015). ‘Extending the queen bee effect: How Hindustani workers cope with disadvantage by distancing the self from the group’. Journal of Social Issues, 71, 476–96.
10.1111/josi.12124
Web of Science® Google Scholar
Deshpande, A. (2011). The Grammar of Caste: Economic Discrimination in Contemporary India. New Delhi: Oxford University Press.
10.1093/acprof:oso/9780198072034.001.0001
Google Scholar
Desmond, M. (2012). ‘Disposable ties and the urban poor’. American Journal of Sociology, 117, 1295–335.
10.1086/663574
Web of Science® Google Scholar
Devine, P. G., Forscher, P. S., Austin, A. J. and Cox, W. T. L. (2012). ‘Long-term reduction in implicit race bias: A prejudice habit-breaking intervention’. Journal of Experimental Social Psychology, 48, 1267–78.
10.1016/j.jesp.2012.06.003
PubMed Web of Science® Google Scholar
DiTomaso, N. (in press). ‘Why difference makes a difference: Diversity, inequality, and institutionalization’. Journal of Management Studies.
Google Scholar
Ertug, G., Brennecke, J., Kovacs, B. and Zou, T. (in press). ‘What does homophily do? A review of the consequences of homophily’. Academy of Management Annals, in press.
Google Scholar
Esses, V. M. and Dovidio, J. F. (2002). ‘The role of emotions in determining willingness to engage in intergroup contact’. Personality & Social Psychology Bulletin, 28, 1202–14.
10.1177/01461672022812006
Web of Science® Google Scholar
Faems, D., Muzio, D., Post, C., Prencipe, A., Sarala, R. and Wei, L.-Q. (2018). Diversity Perspective on Management: Towards More Complex Conceptualizations of Diversity in Management Studies. Durham, United Kingdom: Journal of Management Studies. Available at http://www.socadms.org.uk/wp-content/uploads/SI-Diversity-Perspectives-on-Management-Towards-more-complex-conceptualizations-of-diversity-in-management-studies.pdf (accessed 26 November 2018).
Google Scholar
Financial Times Global MBA Rankings (2017). Global MBA Ranking 2017. London, United Kingdom: The Financial Times. Available at http://rankings.ft.com/businessschoolrankings/global-mba-ranking-2017 (accessed 26 November 2018).
Google Scholar
Gibbons, D. E. and Olk, P. M. (2003). ‘Individual and structural origins of friendship and social position among professionals’. Journal of Personality and Social Psychology, 84, 340–51.
10.1037/0022-3514.84.2.340
PubMed Web of Science® Google Scholar
Gulati, R. and Srivastava, S. B. (2014). ‘ Bringing agency back into network research: Constrained agency and network action’. In D. J. Brass, G. Labianca, A. Mehra, D. S. Halgin and S. P. Borgatti (Eds), Contemporary Perspectives on Organizational Social Networks (Research in the Sociology of Organizations). Bingley: Emerald, 40, 77–93.
10.1108/S0733-558X(2014)0000040004
Google Scholar
Harrison, D. A., Price, K. H. and Bell, M. P. (1998). ‘Beyond relational demography: Time and the effects of surface- and deep-level diversity on work group cohesion’. Academy of Management Journal, 41, 96–107.
10.2307/256901
Web of Science® Google Scholar
Harrison, D. A., Price, K. H., Gavin, J. H. and Florey, A. T. (2002). ‘Time, teams, and task performance: Changing effects of surface- and deep-level diversity on group functioning’. Academy of Management Journal, 45, 1029–45.
10.2307/3069328
Web of Science® Google Scholar
Ibarra, H. (1992). ‘Homophily and differential returns: Sex differences in network structure and access in an advertising firm’. Administrative Science Quarterly, 37, 422–47.
10.2307/2393451
Web of Science® Google Scholar
Ibarra, H. (1995). ‘Race, opportunity, and diversity of social circles in managerial networks’. Academy of Management Journal, 38, 673–703.
10.2307/256742
Web of Science® Google Scholar
Ibarra, H. (2016). ‘5 misconceptions about networking’. Harvard Business Review, Available at https://hbr.org/2016/04/5-misconceptions-about-networking (accessed 7 July 2021). Digital Article Dated April 18, 2016.
Google Scholar
India Census. (2011). A-5 Union Primary Census Abstract. New Delhi, India: Registrar General and Census Commissioner. Available at http://www.censusindia.gov.in/2011census/hlo/pca/pca_pdf/PCA-CRC-0000.pdf (accessed 26 November 2018).
Google Scholar
Ingram, P. and Roberts, P. W. (2000). ‘Friendships among competitors in the Sydney hotel industry’. American Journal of Sociology, 106, 387–423.
10.1086/316965
Web of Science® Google Scholar
Joshi, S. and Malghan, D. (2017). ‘Faculty diversity at the Indian Institutes of Management: A preliminary snapshot’. Economic and Political Weekly, 52, 12–14.
Google Scholar
Jost, J. T. and Banaji, M. R. (1994). ‘The role of stereotyping in system-justification and the production of false consciousness’. British Journal of Social Psychology, 33, 1–27.
10.1111/j.2044-8309.1994.tb01008.x
Web of Science® Google Scholar
Jost, J. T., Banaji, M. R. and Nosek, B. A. (2004). ‘A decade of system justification theory: Accumulated evidence of conscious and unconscious bolstering of the status quo’. Political Psychology, 25, 881–919.
10.1111/j.1467-9221.2004.00402.x
Web of Science® Google Scholar
Kalish, Y., Luria, G., Toker, S. and Westman, M. (2015). ‘Till stress do us part: On the interplay between perceived stress and communication network dynamics’. Journal of Applied Psychology, 100, 1737–51.
10.1037/apl0000023
PubMed Web of Science® Google Scholar
Konrad, A. M., Seidel, M.-D.- L., Lo, E., Bhardwaj, A. and Qureshi, I. (2017). ‘Variety, dissimilarity, and status centrality in MBA networks: Is the minority or the majority more likely to network across diversity?’. Academy of Management Learning and Education, 16, 349–72.
10.5465/amle.2015.0256
Web of Science® Google Scholar
Krackhardt, D. (1992). ‘ The strength of strong ties: The importance of philos in organizations’. In N. Nohria and R. Eccles (Eds), Networks and Organizations: Structure, Form, and Action. Boston, MA: Harvard Business School Press, 216–39.
Web of Science® Google Scholar
Krackhardt, D. and Kilduff, M. (1990). ‘Friendship patterns and culture: The control of organizational diversity’. American Anthropologist, 92, 142–54.
10.1525/aa.1990.92.1.02a00100
Web of Science® Google Scholar
Krackhardt, D. and Kilduff, M. (1999). ‘Whether close or far: Social distance effects on perceived balance in friendship networks’. Journal of Personality and Social Psychology, 76, 770–82.
10.1037/0022-3514.76.5.770
Web of Science® Google Scholar
Kumar, D. (1992). ‘The affirmative action debate in India’. Asian Survey, 32, 290–302.
10.2307/2644940
Web of Science® Google Scholar
Kumar, V. (2016). ‘Discrimination on campuses of higher learning’. Economic and Political Weekly, 51, 12–15.
Google Scholar
Lawrence, B. and Shah, N. (2020). ‘Homophily: Measures and meaning’. Academy of Management Annals, 14, 513–97.
10.5465/annals.2018.0147
Web of Science® Google Scholar
Levy, P. E. and Green, D. P. (2009). ‘Prejudice reduction: What works? A review and assessment of research and practice’. Annual Review of Psychology, 60, 339–67.
10.1146/annurev.psych.60.110707.163607
PubMed Web of Science® Google Scholar
Lincoln, J. R. and Miller, J. (1979). ‘Work and friendship ties in organizations: A comparative analysis of relational networks’. Administrative Science Quarterly, 24, 181–99.
10.2307/2392493
Web of Science® Google Scholar
Marsden, P. (1990). ‘Network data and measurement’. Annual Review of Sociology, 16, 435–63.
10.1146/annurev.so.16.080190.002251
Web of Science® Google Scholar
McCord, M. A., Joseph, D. L., Dhanani, L. Y. and Beus, J. M. (2018). ‘A meta-analysis of sex and race differences in perceived workplace mistreatment’. Journal of Applied Psychology, 103, 137–63.
10.1037/apl0000250
PubMed Web of Science® Google Scholar
McPherson, M., Smith-Lovin, L. and Cook, J. M. (2001). ‘Birds of a feather: Homophily in social networks’. Annual Review of Sociology, 27, 415–44.
10.1146/annurev.soc.27.1.415
Web of Science® Google Scholar
Mehra, A., Kilduff, M. and Brass, D. J. (1998). ‘At the margins: A distinctiveness approach to the social identity and social networks of underrepresented groups’. Academy of Management Journal, 41, 441–52.
10.2307/257083
Web of Science® Google Scholar
Mollica, K. A., Gray, B. and Treviño, L. K. (2003). ‘Racial homophily and its persistence in newcomers' social networks’. Organization Science, 14, 123–36.
10.1287/orsc.14.2.123.14994
Web of Science® Google Scholar
Obstfeld, D. (2005). ‘Social networks, the tertius iungens orientation, and involvement in innovation’. Administrative Science Quarterly, 50, 100–30.
10.2189/asqu.2005.50.1.100
Web of Science® Google Scholar
Pandey, P. and Pandey, S. (2018). ‘Survey at an IIT campus shows how caste affects students' perceptions’. Economic & Political Weekly, 53. Available at https://www.epw.in/engage/article/Survey-at-an-IIT-Campus-Shows-How-Caste-Affects-Students-Perceptions (accessed 24 February 2020).
PubMed Google Scholar
Pettigrew, T. F. and Tropp, L. R. (2006). ‘A meta-analytic test of intergroup contact theory’. Journal of Personality & Social Psychology, 90, 751–83.
10.1037/0022-3514.90.5.751
PubMed Web of Science® Google Scholar
Pew Research Center (2021). Religion in India: Tolerance and Segregation. Available at https://www.pewforum.org/2021/06/29/religion-in-india-tolerance-and-segregation/ (accessed 14 July 2021).
Google Scholar
Pyke, K. and Dang, T. (2003). ‘“FOB” and “whitewashed”: Identity and internalized racism among second generation Asian Americans’. Qualitative Sociology, 26, 147–72.
10.1023/A:1022957011866
Google Scholar
Qureshi, I., Fang, Y., Haggerty, N., Compeau, D. R. and Zhang, X. (2018a). ‘IT mediated social interactions and knowledge sharing: Role of competence based trust and background heterogeneity’. Information Systems Journal, 28, 929–55.
10.1111/isj.12181
Web of Science® Google Scholar
Qureshi, I., Sutter, C. and Bhatt, B. (2018b). ‘The transformative power of knowledge sharing in settings of poverty and social inequality’. Organization Studies, 39, 1575–99.
10.1177/0170840617727777
Web of Science® Google Scholar
Ragins, B. R. and Ehrhardt, K. (in press). ‘Gaining perspective: The impact of close cross-race friendships on diversity training and education’. Journal of Applied Psychology, in press.
Google Scholar
Rai, S. (2021). How Big Tech is Importing India’s Caste Legacy to Silicon Valley. Bloomberg.com. Available at https://www.bloomberg.com/news/features/2021-03-11/how-big-tech-is-importing-india-s-caste-legacy-to-silicon-valley (accessed 14 July 2021).
Google Scholar
Reagans, R. (2011). ‘Close encounters: Analyzing how social similarity and propinquity contribute to strong network connections’. Organization Science, 22, 835–49.
10.1287/orsc.1100.0587
Web of Science® Google Scholar
Riaz, S. (2015). ‘Bringing inequality back in: The economic inequality footprint of management and organizational practices’. Human Relations, 68, 1085–97.
10.1177/0018726715584803
Web of Science® Google Scholar
Ridgeway, C. L. (1993). ‘ Gender, status and the social psychology of expectations’. In P. England (Ed.), Theory on Gender: Feminism on Theory. New York: de Gruyter, 175–97.
Google Scholar
Ridgeway, C. L. (2014). ‘Why status matters for inequality’. American Sociological Review, 79, 1–16.
10.1177/0003122413515997
Web of Science® Google Scholar
Ripley, R., Boitmanis, K., Snijders, T. A. B. and Schoenenberger, F. (2018). RSiena - R package version 1.1-288/r288. Available at http://R-Forge.R-project.org/projects/rsiena (accessed 12 March 2020).
Google Scholar
Ripley, R. M., Snijders, T. A. B., Boda, Z., Vӧrӧs, A. and Preciado, P. (2016). Manual for SIENA Version 4.0 (version 25 February 2016). Oxford: University of Oxford.
Google Scholar
Rivera, M., Soderstrom, S. and Uzzi, B. (2010). ‘Dynamics of dyads in social networks: Assortative, relational, and proximity mechanisms’. Annual Review of Sociology, 36, 91–115.
10.1146/annurev.soc.34.040507.134743
Web of Science® Google Scholar
Sasovova, Z., Mehra, A., Borgatti, S. P. and Schippers, M. C. (2010). ‘Network churn: The effects of self-monitoring personality on brokerage dynamics’. Administrative Science Quarterly, 55, 639–70.
10.2189/asqu.2010.55.4.639
Web of Science® Google Scholar
Seidel, M.-D.- L. (2011). ‘ Homophily’. In G. Barnett (Ed.), Encyclopedia of Social Networks. Newbury Park, CA: Sage, 383–84.
10.4135/9781412994170.n150
Google Scholar
Shore, L. M., Bommer, W. H., Rao, A. N. and Seo, J. (2009). ‘Social and economic exchange in the employee-organization relationship: The moderating role of reciprocation wariness’. Journal of Managerial Psychology, 24, 701–21.
10.1108/02683940910996752
Web of Science® Google Scholar
Sidanius, J. S., Levin, S., van Laar, C. and Sears, D. O. (2008). The Diversity Challenge: Social Identity and Intergroup Relations on the College Campus. New York: The Russell Sage Foundation.
Google Scholar
Sidanius, J. and Pratto, F. (1999). Social Dominance: An Intergroup Theory of Social Hierarchy and Oppression. New York: Cambridge University Press.
10.1017/CBO9781139175043
Google Scholar
Snijders, T. A. (1996). ‘Stochastic actor-oriented models for network change’. Journal of Mathematical Sociology, 21, 149–72.
10.1080/0022250X.1996.9990178
Web of Science® Google Scholar
Snijders, T. A. B. (2001). ‘The statistical evaluation of social network dynamics’. Sociological Methodology, 31, 361–95.
10.1111/0081-1750.00099
Web of Science® Google Scholar
Srinivas, M. N. (1996). ‘ Introduction’. In M. N. Srinivas (Ed.), Caste: Its Twentieth Century Avatar. New Delhi: Penguin, IX–XXXVIII.
Google Scholar
Steglich, C., Snijders, T. A. and Pearson, M. (2010). ‘Dynamic networks and behavior: Separating selection from influence’. Sociological Methodology, 40, 329–93.
10.1111/j.1467-9531.2010.01225.x
Web of Science® Google Scholar
Tomaskovic-Devey, D. (1993). Gender and Racial Inequality at Work: The Sources and Consequences of Job Segregation. Ithaca, NY: ILR Press.
10.7591/9781501717505
Google Scholar
Vaid, D. (2014). ‘Caste in contemporary India: Flexibility and persistence’. Annual Review of Sociology, 40, 391–410.
10.1146/annurev-soc-071913-043303
Web of Science® Google Scholar
Wasserman, S. and Faust, K. (1994). Social Networks Analysis: Methods and Applications. New York: Cambridge University Press.
10.1017/CBO9780511815478
Google Scholar
Wilkerson, I. (2020). Caste: Origins of Our Discontents. New York: Random House.
Google Scholar
Wimmer, A. and Lewis, K. (2010). ‘Beyond and below racial homophily: Erg models of a friendship network documented on Facebook1’. American Journal of Sociology, 116, 583–642.
10.1086/653658
PubMed Web of Science® Google Scholar
Wiser, W. H. (1988). The Hindu Jajmani System: A Socio-Economic System Interrelating Members of a Hindu Village Community in Services. New Delhi: Munshiram Manoharlal.
Google Scholar
Wolff, H.-G. and Moser, K. (2009). ‘Effects of networking on career success: A longitudinal study’. Journal of Applied Psychology, 94, 196–206.
10.1037/a0013350
PubMed Web of Science® Google Scholar
Zacharias, A. and Vakulabharanam, V. (2011). ‘Caste stratification and wealth inequality in India’. World Development, 39, 1820–33.
10.1016/j.worlddev.2011.04.026
Web of Science® Google Scholar
Zulfiqar, G. and Prasad, A. (2021). ‘Challenging social inequality in the Global South: Class, privilege, and consciousness-raising through critical management education’. Academy of Management Learning and Education, 20, 156–81.
10.5465/amle.2019.0294
Web of Science® Google Scholar

Citing Literature

Volume58, Issue8

Special Issue:Diversity Perspective on Management: Towards more complex conceptualizations of diversity in management studies

December 2021

Pages 2111-2136

Bridging Caste Divides: Middle-Status Ambivalence, Elite Closure, and Lower-Status Social Withdrawal

Abstract

INTRODUCTION

HOMOPHILY AND STATUS: THEORY DEVELOPMENT

HOMOPHILY AMONG ELITE, MIDDLE, AND LOWEST-STATUS GROUPS

LOWEST-STATUS GROUP NETWORKING BEHAVIOUR