Differential Impacts of Functional,Geographical,and Hierarchical Diversity on Knowledge Sharing in the Midst of Organizational Change

Functional Diversity

Functional diversity refers to the distribution of organizational members across relevant functional categories (Bunderson & Sutcliffe, 2002). Functional diversity has been widely studied as one of the task-related diversity attributes in contrast to biodemographic markers (for a review, see Horwitz & Horwitz, 2007). Specialized functional areas can be the sources of differentiation and in-group biases that may thwart team collaboration efforts because organizational members highly identify with their job role and career, which underpin their professional identity and commitment (Milliken & Martins, 1996). Specifically, given that knowledge transfer implies costs such as investments in time and effort (Reagans & McEvily, 2003), differential expertise may hamper or delay the process of knowledge transfer. Organizational members in disparate knowledge domains may be less motivated to transfer knowledge when they perceive it to be difficult due to nonoverlapping functional backgrounds (Bunderson & Sutcliffe, 2002). Dougherty (1992) also argued that communicating with others in different expertise is equivalent to navigating different “thought worlds” across interpretive barriers, which further increases associated costs and constrains joint learning. Indeed, research shows that specialized languages and situated understanding can hinder exchanging knowledge beyond functional and occupational boundaries (Bechky, 2003).

Functional diversity may also have a negative impact on knowledge acquisition. It is particularly challenging to obtain useful knowledge across multiple functional areas because people may have a limited ability to accept and respond to others’ expert knowledge as a novice (Cronin & Weingart, 2007). Furthermore, high network diversity that spans various knowledge areas can increase communication flows from different sources, and such overload and complexity of information actually prevented individuals from locating useful knowledge (Aral & van Alstyne, 2011). Thus, functional diversity, while potentially beneficial, is likely to preclude effective knowledge transfer and acquisition unless organizational members engage in coordinated effort to build a shared understanding. Hence, the following hypothesis is presented:

Organizational members may also experience difficulties understanding the expertise of others in different functional areas. When workers maintain functionally diverse social networks, the accurate recognition of others’ expertise is critical for effective collaborative work. However, achieving expertise awareness can be challenging, particularly in a changing environment in which members lack shared experiences. Without a shared history, one’s expertise is not readily apparent to others, and the ability to recognize or utilize each other’s expertise is shaped by interaction styles (Littlepage & Mueller, 1997). Individuals tend to draw on stereotypes associated with diversity attributes to infer relative knowledge of others and distribute knowledge responsibilities in the early phase of work (Hollingshead & Fraidin, 2003; Yoon & Hollingshead, 2010). However, because actual expertise may be different from the initial perception, continued learning of others’ work is crucial for expertise mapping (Hollingshead & Brandon, 2003). In a postrestructuring environment, organizational members are not provided with sufficient opportunities for long-term engagement, which is essential to validate their initial perception (Wegner, Erber, & Raymond, 1991). Hence, workers who have functionally diverse social networks may face greater barriers to understanding their coworkers’ expertise, hypothesized as follows.

Geographical Diversity

Geographical diversity refers to the extent to which members are dispersed across multiple locations. Distributed collaboration entails much effort to negotiate common ground, define coordination protocols, and overcome obstacles associated with differences in time, culture, and contexts (Olson & Olson, 2000). These challenges become particularly salient in knowledge transfer. Geographical boundaries often give rise to the mutual knowledge problem (Cramton, 2001) stemming from the uneven distribution of information and the failure to share and retain contextual knowledge across different locations. Sole and Edmondson (2002) also argue that dispersed workers first need to recognize and adjust for locale-specific practices to identify situated knowledge—the knowledge embedded in social and physical contexts that reinforces location boundaries—before transferring their knowledge. Distributed workers tend to have far greater understanding of their own context than others’ contexts; furthermore, they cannot readily describe local knowledge as it is often taken for granted (Gibbs, Kim, & Boyraz, 2017; Gibson & Gibbs, 2006). Thus, transferring knowledge across geographical boundaries involves a range of additional efforts to identify and contextualize local knowledge.

Geographical diversity is also likely to hamper the acquisition of useful knowledge particularly when members have not created and maintained common ground. In globally distributed teams, members are likely to face challenges regarding processing and aggregating received knowledge, especially if it is uncodified (Stanko & Gibson, 2009). Moreover, geographical boundaries in multinational organizations often overlap with national or cultural boundaries. Cultural differences contribute to a lack of understanding of norms and values that impede full acquisition of knowledge shared from people in other locations (Mowery, Oxley, & Silverman, 1996; van Wijk, Jansen, & Lyles, 2008). National diversity can also provide a potential source for misunderstanding, stereotyping, and disagreement, all of which in turn hinder knowledge acquisition (Adler, 1986; Kirkman & Shapiro, 1997). Overall, geographical diversity has been found to be negatively associated with knowledge sharing (for a review, see de Guinea, Webster, & Staples, 2012). Thus, the following hypothesis is presented:

Geographical dispersion is also negatively related to the awareness of expertise. Because expertise awareness develops primarily through informal communication and shared experiences, the lack of proximity may inhibit the understanding of others’ expertise (Fulk, Monge, & Hollingshead, 2005). Unlike distributed workers, collocated employees are able to observe each other’s work practices to vicariously learn about others’ specialties; namely, unobtrusive monitoring of colleagues’ behaviors helps workers discern expertise cues and understand different levels of their colleagues’ expertise. In line with this, Littlepage and Mueller (1997) found that in-person interactions strongly influence perception of expertise. In distributed work groups, remote workers are thus likely to have difficulties understanding other members’ expertise (Su, 2012). When one’s network contacts are dispersed across multiple locations, it may be increasingly challenging to understand their expertise. Furthermore, expertise of members in different cultures was not equally recognized in intercultural groups, despite similar levels of expertise among group members (Bazarova & Yuan, 2013). This discussion leads to the following hypothesis:

Hierarchical Diversity

Hierarchical diversity has received relatively little attention from diversity researchers. This is mainly because diversity scholarship concentrates on the composition of a team, which does not display much variance in terms of hierarchical levels (i.e., a team is a basic unit that consists of one supervisor and several members). However, hierarchical diversity in one’s network—namely, one’s contacts that cut across various hierarchical levels—can provide distinctive advantages for knowledge sharing. Given that the flow of information, ideas, and resources is conditioned by organizational hierarchy (Salancik, 1995), maintaining a network across hierarchical boundaries may help employees exchange knowledge more effectively than others whose networks are limited only to similar hierarchical ranks or status.

Indeed, vertical communication produces different benefits for knowledge sharing as opposed to horizontal communication (Knight et al., 1999). As supervisors and peers tend to provide different information content, having multiple sources that span hierarchical levels can be conducive to achieving balanced knowledge about a new environment, contributing to an enhanced understanding of others and others’ work (Miller & Jablin, 1991). The improved understanding of their surroundings can enable employees to better approach others to effectively convey their knowledge and provide necessary information. In particular, when hierarchies serve as a system to exert authority and limit knowledge transfer (Conner & Prahalad, 1996), social networks that cut across hierarchical boundaries may offer unique advantages in accessing and disseminating knowledge.

In addition, because high-status employees may be more knowledgeable about organizational strategies, plans, and changes, connections with managers and high-level executives can facilitate acquisition of needed resources and adequate information. In particular, those from upper management tend to be privy to certain information and can provide other members with experience, legitimation, and more relevant information (Brass, 1984; J. N. Cummings, 2004). Highlighting the benefits of unique information and diverse perspectives, Cross and Cummings (2014) demonstrated that connections across multiple hierarchical levels are positively related to individual performance especially in knowledge work. In the context of organizational change, relationships across hierarchical levels can help employees obtain valuable knowledge to adapt to new work arrangements. Also, managers and executives can better understand members’ experiences and stay informed of team progress through regular communication with them. Although empirical research on hierarchical diversity is still nascent, this study posits that input from different hierarchical positions may promote one’s ability to provide knowledge, obtain useful knowledge, and understand others’ expertise. Hence, the following hypothesis is proposed:

Research Context

This study was conducted at Enterprise Solutions (pseudonym), a multinational high-tech organization that builds software for data management, product design, and product development. Its headquarters are located in the United States, and its regional offices are dispersed in South America, East Asia, South Asia, Eastern Europe, and Western Europe. The company’s divisions are by and large structured around its major product lines or services. Each division consists of different departments that focus on specific functions—engineering, user experience, localization, developer support, sales, and evangelism, to name a few. Most high-level executives and product managers are employed at headquarters, collaborating with team members dispersed in multiple locations. In particular, employees at the core engineering centers in the United States, South America, and Eastern Europe should collaborate across regions and functions. Other regional offices may have a different business focus such as sales and marketing.

In October 2014, Enterprise Solutions implemented a company-wide structural change that merged multiple divisions to facilitate cross-specialty collaboration. Reflecting this change, new division heads were assigned and teams were reconfigured, resulting in changes in reporting structures. The integration of different divisions and groups was part of the company’s strategic plan for enhancing cross-functional interactions to make teams “holistically accountable” and support “more fluid movement” between teams, which would make knowledge sharing “much faster and more efficient within the same group” than across departments. ¹ For example, developer support engineers, who had belonged to a separate division before the change, were assigned to the same division with developers. By deploying related work groups under the same division, high-level executives anticipated that boundary-spanning interactions would be promoted and improved.

Data Collection

A company-wide online survey was conducted approximately 1 month after the structural change. The author directly distributed the survey to all 275 full-time employees across regions and closely coordinated with the CEO and higher executives to ensure participation from employees. The CEO sent out an email announcement to everyone to encourage participation; additionally, the author emailed weekly reminders while the survey remained open. Employees were informed that participation was voluntary, individual responses would not be disclosed, and the data would be kept confidential. Over a 1-month period, 224 employees completed the survey, yielding an 81.45% completion rate.

To capture full network data, the author obtained the most up-to-date version of the company roster from the vice president of human resource management. The roster was accurate except that some employees’ legal names were different from the names actually used. The author revised those cases before survey implementation. Respondents were provided with the complete roster of employees to identify their knowledge-sharing network. The network questionnaire employed a sociometric choice format to have participants choose their knowledge-sharing contacts: They were asked to select (a) people who acted as a critical source of task-related knowledge and (b) people for whom they have been a critical source of task-related knowledge (Reagans & McEvily, 2003). Respondents were allowed to nominate their knowledge-sharing ties without limitations. After respondents selected their knowledge-sharing contacts from the roster, they answered a series of questions about the quality of knowledge transfer as well as acquisition and their awareness of expertise regarding each alter. The author also obtained node attributes data directly from the company roster and organization charts, which included the information about employees’ location, job role, department, division, and hierarchical levels. Because this study’s primary concern is knowledge sharing for collaborative task accomplishment, all core job functions that jointly work and coordinate were included in the analysis. ² This process created a directed, complete network data set of 211 respondents.

Measures

Diversity

Network diversity was operationalized as network range, which was originally proposed by Burt (1983). The node attributes such as job functions, locations, and hierarchical levels were treated as diversity attributes and used to compute functional, geographical, and hierarchical diversity of one’s network. Network diversity is defined as nd_i, calculated as follows:

n d i = 1 − ∑ k = 1 m p k p i k 2

where p_k describes the intensity of relations between individuals within the attribute k, and p_ik indicates the strength of the connection from person i to attribute k. In the above formula, p_k is defined as follows:

p k = ∑ j = 1 M k Z i j ∑ q = 1 S k Z i q ′ q = j

where S_k is the number of contacts identified by participants in attribute k, M_k is the number of respondents in attribute k, Z_iq is the connections between an individual in attribute k to any contact, and Z_ij is the intensity of relations between an individual in attribute k to another member in the same attribute, namely, the proportion of interaction represented by connections. In turn, p_ik is defined as follows:

p i k = ∑ j = 1 N k Z i j ∑ q = 1 N Z i q ′ q = j

where N_k is the number of contacts that participant i mentioned from attribute k, and Z_iq is the intensity of the relationship that participant i has with contact q. Thus, the values of network diversity range from zero to one.

Overall, the more one’s connections spread across different functions, locations, and hierarchical levels and the more weak ties exist within each group, the higher one’s network diversity is. Participants’ functions, locations, and levels were extracted from the company roster (for demographics, see Table 1).

OPEN IN VIEWER

Table 1.

Participant Demographics (n = 211).

Job function		Hierarchical level		Location		Gender
Engineering	98	Executives	9	United States	86	Male	158
Sales and marketing	36	Managers	30	South America	31	Female	53
Developer support	32	Full-time employees	168	Eastern Europe	72
User experience	13	Full-time interns	4	Western Europe	6
Product management	11			East Asia	10
Information systems	11			South Asia	6
Evangelism	5
Localization	5

Analysis

As a preliminary step, descriptive analysis and bivariate correlation analysis were performed (see Table 2). Independent variables showed very low or moderate levels of correlations with one another; hence, they were included in the analysis models. Independent and dependent variables generally showed significant correlations. As a primary analysis, multiple linear regressions were conducted to test hypotheses. In all three models that predict knowledge transfer, knowledge acquisition, and expertise awareness, diversity variables were added to the models after control variables.

OPEN IN VIEWER

Table 2.

Descriptive Statistics and Bivariate Correlations.

	M	SD	1	2	3	4	5	6	7	8	9	10
1. Transfer	4.20	0.58
2. Acquisition	4.24	0.51	.49**
3. Expertise	4.31	0.64	.50**	.46**
4. Functional diversity	0.67	0.17	−19*	−.08	−.13*
5. Geographical diversity	0.51	0.18	−.30**	−.23**	−.17*	.17*
6. Hierarchical diversity	0.53	0.16	.01	.09	.14*	.39**	.20**
7. Contact frequency	3.06	0.74	.25**	.16*	.15*	−.23**	−.28**	−.12
8. Emotional closeness	3.23	0.89	.24**	.21**	.31**	−.19**	−.14*	−.01	.71**
9. Ego betweenness	101.52	83.79	−.13	−18*	−.02	.29**	.25**	.13	−.01	.06
10. Ego network size	17.99	11.97	−.12	−.19*	−.01	.41**	.28**	.26**	−.14	−.02	.82**
11. Ego network density	0.28	0.17	.17*	.19**	.02	−.41**	−.35**	−.20**	.28**	.18*	−.40**	−.59**

*

p < .05. **p < .01.

Theoretical and Practical Implications

This study contributes to the scholarship of diversity and knowledge sharing in three ways. First, by linking diversity and knowledge-sharing research, this study elaborates on the multifaceted impacts of diversity on organizational outcomes. Although the implications of diversity for performance outcomes have received substantial scholarly attention, its impact on knowledge-sharing behaviors has been underexamined. Given that knowledge sharing and expertise recognition can also affect performance outcomes, this study adds significant insight into how diversity improves or impedes the quality of work. Furthermore, mutual understanding and common ground have been viewed as the building blocks of knowledge sharing; hence, the ways in which organizational members perceive, interpret, and manage diversity should be investigated more in the context of knowledge sharing because diverse perspectives that originate from multiple group memberships and histories can shape the process of building a consensus or a common course of action.

Second, drawing on a social network perspective, this study assesses the level of diversity as the boundary-spanning connections of an individual’s task-related communication network instead of team composition. This approach helps researchers grasp a more nuanced picture of diversity based on one’s social network; namely, members may experience different levels of diversity depending on their work relationships and interdependence even if they belong to the same team, which affects the nature of knowledge-sharing networks (M. S. Weber & Kim, 2015). The relational view moves beyond the relatively simple notion of subgroups and enables a fine-grained analysis of diversity attributes considering external members as well (Reagans & Zuckerman, 2011; van Knippenberg & Schippers, 2007). Taking advantage of the social network approach, the current study examined the configuration of an individual’s network in terms of job functions, hierarchical levels, and geographic locations that represent differential impacts of boundary-crossing relationships.

Finally, this study shows that the implications of boundary-spanning and knowledge-sharing behaviors should be understood in tandem with the ramifications of organizational change. Network characteristics may yield disparate knowledge-sharing outcomes during structural change: Employees who have more strong ties will exchange knowledge more successfully than others who maintain far-reaching networks across functions and locations. This implies that diversity militates against cross-boundary knowledge sharing when members encounter new collaboration challenges. The findings stress the positive role of cohesion, rather than network range, in successful communicative outcomes in the midst of change. This study calls for further research in the area of organizational change—how organizational members deal with uncertainty and complexity, forge new connections, and address communication challenges to eventually realize the desired goals of the change.

The findings of this study also offer important practical implications. In the current study, Enterprise Solutions restructured their teams and hierarchy in hopes for improved cross-boundary communication; however, their expectations had not been realized at the time of data collection. Although diversity can exert a positive influence on various organizational outcomes, practitioners should acknowledge that diverse teams may experience difficulty developing relationships and negotiating common ground (Ancona & Caldwell, 1992). Positive benefits accrue once team members figure out how to jointly overcome the challenges associated with diversity (Mannix & Neale, 2005). Thus, it is crucial to create and maintain an organizational climate in which members embrace diversity to capitalize on members’ knowledge, expertise, and different viewpoints. Executives and managers may consider offering workshops to promote the understanding of workplace diversity and its implications. Also, managers should overhaul their day-to-day communication channels and activities to identify a way to improve cross-boundary exchanges.

Practitioners may also need to invest efforts in mitigating the negative outcomes of functional and geographical diversity. Specifically, how organizational members interpret and carry out the change effort over time will likely affect work outcomes. Although it is not a simple task for management to effectively convey their intentions and expectations, ongoing efforts to achieve “buy-in” from employees may alleviate adverse effects resulting from change. As this study found a positive impact of communication across hierarchical levels, change implementers are advised to consider the potentially powerful roles of middle-level leaders in framing and advocating the rationale of new initiatives proposed by top management (Lewis, 1999). Policy makers may also want to examine short-term and long-term implications of organizational change for knowledge integration. When members experience structural turbulence and need to renew their work routines and relationships under time pressure, managers should facilitate coordination as well as the standardization of team processes.

Limitations and Future Directions

The findings of this study are based on a quantitative cross-sectional study. A second wave of the survey at a different point in time would provide insights into whether the negative influence of diversity is a short-term detriment resulting from restructuring or a more grounded problem that indicates a lack of mutual understanding across boundaries. Examining long-term effects of organizational change can yield significant insight into knowledge-sharing patterns as network reconfigurations are likely to have a considerable impact on power, influence, and work processes (Susskind, 2007). A series of similar surveys can examine whether and how long-term interactions enhance expertise awareness and knowledge sharing, enabling an in-depth understanding of the impacts of diversity on knowledge work in a postrestructuring setting.

Because extant work indicates that the influences of diversity may vary by a range of contextual factors and communicative interactions, it is important to qualitatively examine the interplay between the different types of diversity and other environmental characteristics. Functional, geographical, and hierarchical diversity may engender differential challenges depending on organizational structures, cultures, and practices. Qualitative interviews and observations can enrich our understanding of the mechanisms through which a particular type of diversity hinders or promotes expertise recognition and knowledge exchange. Especially in global organizations, geographical diversity may be investigated in conjunction with cultural diversity given that regional boundaries often overlap with national borders. Future research should also examine the role of technologies in facilitating/hindering the processes of knowledge sharing across functional, regional, and hierarchical boundaries. In particular, advanced knowledge management systems may help individuals overcome the impediments of cross-location sharing by offering various ways to store, exchange, and retrieve a complex set of knowledge (Heinz & Rice, 2009).