PERSONNEL PSYCHOLOGY
2016, 69, 159?198

EMPOWERING LEADERSHIP AND EFFECTIVE
COLLABORATION IN GEOGRAPHICALLY
DISPERSED TEAMS

N. SHARON HILL
The George Washington University

KATHRYN M. BARTOL
University of Maryland

Our research integrates theoretical perspectives related to distributed
leadership in geographically dispersed teams with empowering lead-
ership theory to build a multilevel model of virtual collaboration and
performance in dispersed teams. We test the model with procurement
teams in a major multinational corporation. Our results show a signifi-
cant cross-level effect of empowering team leadership, such that under
conditions of high empowering team leadership, a team member?s virtual
teamwork situational judgment (VT-SJ) is positively and significantly
associated with his or her virtual collaboration behaviors and also indi-
rectly with his or her individual performance in the team. At the team
level, our findings also suggest that the impact of empowering leader-
ship on team members? aggregate virtual collaboration, and indirectly on
team performance, increases at higher levels of team dispersion. These
findings shed important light on the role of team leadership in foster-
ing effective collaboration and performance of geographically dispersed
virtual teams.

To support major strategic initiatives in areas such as globalization,
outsourcing, and strategic partnering, organizations are increasingly turn-
ing to the use of geographically dispersed teams in which members rely
on technology to collaborate virtually in the team. Dispersed or virtual
teams offer many potential advantages (Martins, Gilson, & Maynard,
2004; Rosen, Furst, & Blackburn, 2006), including the ability to have the
most technically qualified individuals work on tasks regardless of loca-
tion while also offering opportunities for sizable cost savings resulting
from reduced travel. With such potential benefits, it is small wonder that
organizations have an increasing interest in the utilization of such teams
(Martins et al., 2004; Rosen et al., 2006). At the same time, reports point to
special challenges individuals face in their collaborations with dispersed

Correspondence and requests for reprints should be addressed to N. Sharon Hill, The
George Washington University, School of Business, 315F Funger Hall, 2201 G Street, NW,
Washington, DC 20052; [email?protected]

C? 2015 Wiley Periodicals, Inc. doi: 10.1111/peps.12108

159

160 PERSONNEL PSYCHOLOGY

team members (for reviews, see Axtell, Fleck, & Turner, 2004; Martins
et al., 2004; O?Leary & Cummings, 2007). For example, research shows
that geographic dispersion may impede effective information sharing, co-
ordination, problem solving, building trust, and constructively resolving
conflicts with others on the team (Cramton, 2001; Cramton & Webber,
2004; Hill, Bartol, Tesluk, & Langa, 2009; Hinds & Mortensen, 2005;
Jarvenpaa & Leidner, 1999; Joshi, Lazarova, & Liao, 2009; O?Leary &
Cummings, 2007).

In the face of such challenges, numerous scholars have pointed to the
potential importance of team leaders in promoting virtual collaboration
that contributes to high levels of performance in dispersed teams (e.g.,
Bell & Kozlowski, 2002; Blackburn, Furst, & Rosen, 2003; Malhotra,
Majchrzak, & Rosen, 2007; Martins et al., 2004; Weisband, 2008; Zigurs,
2003). In their major theorizing about leadership in dispersed teams, Bell
and Kozlowski (2002) suggested that the challenges of collaboration in
such teams and the attendant difficulty in monitoring team member be-
haviors require distributing leadership functions to team members while,
at the same time, fostering collaboration among them. Yet the limited
existing empirical research related to distributed forms of leadership in
dispersed teams has focused on leadership as it relates to use of infor-
mation and communication tools in teams (e.g., Rapp, Ahearne, Mathieu,
& Rapp, 2010; Surinder, Sosik, & Avolio, 1997; Wakefield, Leidner, &
Garrison, 2008) and/or failed to consider geographic dispersion in the
team (e.g., Pearce, Yoo, & Alavi, 2004).

One form of leadership that embodies Bell and Kozlowski?s (2002)
recommended approach is empowering leadership. Due to its combination
of sharing power with team members while also providing a facilitative
and supportive environment (Arnold, Arad, Rhoades, & Drasgow, 2000;
Srivastava, Bartol, & Locke, 2006), empowering leadership appears to
be particularly well suited to helping team members meet the demands
of collaborating in a dispersed teamwork environment. Hence, the over-
all purpose of this study is to evaluate the extent to which distributing
leadership to team members by way of empowering leadership promotes
more effective virtual collaboration and ultimately performance in dis-
persed teams. We define virtual collaboration as collaborative behaviors
that promote interactions that support geographically dispersed teamwork.

In their theorizing, Bell and Kozlowski (2002) noted that distributing
leadership functions to dispersed teams creates an environment that fa-
cilitates each team member applying relevant knowledge and judgment
in order to successfully collaborate virtually with other team members.
This is paramount in dispersed teams because each member faces chal-
lenges unique to his or her local dispersed circumstances. Further, as a
result of being separated from others in the team, each team member must

HILL AND BARTOL 161

regulate his or her own behaviors and performance in the team. Accord-
ingly, taking a multilevel approach, we examine the extent to which an
empowering leadership team context moderates the influence of individual
virtual teamwork situational judgment (VT-SJ) on team member virtual
collaboration, and ultimately team member performance. VT-SJ reflects
Bell and Kozlowski?s (2002) emphasis on each member of a dispersed
team having ?attributes to be able to . . . operate in a virtual environment?
(p. 26). As such, VT-SJ describes an individual?s knowledge about suc-
cessful virtual collaboration strategies and how to apply that knowledge to
formulate effective responses in geographically dispersed teamwork situ-
ations. Existing research related to individual characteristics in dispersed
teams has largely been concerned with relatively stable personality char-
acteristics or personal orientations that either adversely influence or aid
computer-mediated interactions (e.g., Staples & Webster, 2007; Tan, Wei,
Watson, Clapper, & McLean, 1998; Workman, Kahnweiler, & Bommer,
2003). Our focus on VT-SJ responds to calls for greater attention to the role
of individual differences in dispersed teams (Kirkman, Gibson, & Kim,
2012) by examining a characteristic?knowledge and judgment about op-
erating effectively in dispersed team situations?that can potentially be
developed. In addition, we extend leadership research that goes beyond
the existing predominant focus on team-level effects to consider how team
leadership, as a team-level stimuli, might have cross-level effects on im-
portant individual-level processes in teams (e.g., Chen & Kanfer, 2006;
Chen, Kirkman, Kanfer, Allen, & Rosen, 2007).

Building further on Bell and Kozlowski?s (2002) theorizing about
dispersed team leadership, we argue that empowering team leadership
will play a more important role in fostering the virtual collaboration
and performance of the team as a whole as geographic dispersion in the
team increases. This is because the challenges of collaborating virtually
can be expected to intensify as the level of team dispersion increases
(O?Leary & Cummings, 2007). This line of inquiry adds to the lim-
ited research that has explicitly measured geographic dispersion in con-
junction with leadership effects in dispersed teams (for exceptions, see
Cummings, 2008; Gajendran & Joshi, 2012; Hoch & Kozlowski, 2014;
Joshi et al., 2009). At the same time, it contributes to emerging leader-
ship research that seeks to shed light on situations in which empowering
team leadership is more or less effective (e.g., Chen, Sharma, Edinger,
Shapiro, & Fahr, 2011; Mathieu, Ahearne, & Taylor, 2007; Yun, Faraj, &
Sims, 2005).

In summary, we build a theoretical model that makes three impor-
tant contributions to research on dispersed teams as well as empowering
leadership research. First, we contribute to theory on leadership in ge-
ographically dispersed teams (Bell & Kozlowski, 2002) by integrating

162 PERSONNEL PSYCHOLOGY

Team Performance Team Virtual Collaboration
Empowering Team

Leadership

Virtual Teamwork
Situational Judgment

Team Member
Virtual Collaboration

Team Member
Performance

Team Geographic
Dispersion

Figure 1: Theoretical Model.

notions of distributed leadership in such teams with those of empower-
ing leadership theory. In doing so, we support contentions regarding the
value of more distributed forms of leadership in dispersed teams (Bell &
Kozlowski, 2002) while also extending empowering leadership theory to
the dispersed team realm.

Second, by exploring the cross-level moderating effect of empower-
ing leadership in strengthening the impact of VT-SJ on a team member?s
virtual collaboration, we support the notion that distributed leadership
facilitates a team member?s use of relevant attributes to enhance virtual
collaboration (Bell & Kozlowski, 2002). Our focus on VT-SJ also extends
the limited literature on individual differences in dispersed teams (Hertel
et al., 2005; Kirkman et al., 2012) by moving beyond the predominant
focus on stable personality traits to highlight a team member characteristic
that can potentially be developed to improve individual virtual collabora-
tion and ultimately performance in the team.

Finally, at the team level, our study supports Bell and Kozlowski?s
(2002) notion that distributed forms of leadership, such as empower-
ing leadership, will be more impactful in teams with greater geographic
dispersion, ultimately enhancing team performance (Nauman, Mansur
Khan, & Ehsan, 2010; Pearce et al., 2004; Zhang, Tremaine, Egan,
Milewski, O?Sullivan, & Fjermestad, 2009). In this way, the present
research also contributes to empowering leadership theory by demon-
strating an important new boundary condition for empowering leadership
effects on dispersed team outcomes. Our multilevel model is shown in
Figure 1.

HILL AND BARTOL 163

Theory and Hypothesis Development

Dispersed teams consist of geographically distributed coworkers who
interact using a combination of telecommunications and information tech-
nology to accomplish an organizational task (Townsend, DeMarie, &
Hendrickson, 1998). In this section, we use Bell and Kozlowski?s (2002)
theorizing about leadership in dispersed teams as a foundation to develop
the multilevel theoretical model for this study. The model shows empow-
ering leadership as an important factor for promoting effective virtual
collaboration behaviors and, ultimately, performance of both individual
team members and the dispersed team as a whole. We first consider the
cross-level effect of empowering leadership in creating a team context in
which a team member is more likely to apply his or her knowledge about
strategies for effective virtual teamwork in support of virtual collaboration
with distributed teammates. These collaborative behaviors then facilitate
a higher level of performance for the individual team member operating
within the team. At the team level, we build a model wherein the influence
of empowering leadership on the team?s aggregated virtual collaboration is
moderated and strengthened by the degree of team geographic dispersion,
with ultimate implications for team performance.

Empowering Leadership, VT-SJ and Team Member Virtual Collaboration

In the context of virtual teamwork, there is considerable support in the
related literature for the notion that geographically dispersed teams face
challenges that require team members to engage in collaborative behaviors
tailored to their dispersed circumstances (e.g., Byron, 2008; Cramton,
2001; Hinds & Bailey, 2003; Hinds & Weisband, 2003; Jarvenpaa &
Leidner, 1999; O?Leary & Cummings, 2007). Virtual collaboration refers
to behaviors enacted by a team member in support of effective interactions
with teammates in geographically dispersed teamwork environments. We
argue that empowering leadership plays an important role in facilitating
the process through which team members enact such behaviors.

In their theorizing about distributed team leadership, Bell and Ko-
zlowski (2002) noted the importance of each member of a dispersed
team having attributes that enable them to operate in a virtual environ-
ment. However, at the same time, they highlighted the critical role of
team leaders in creating a team context that allows each member to make
best use of these attributes for collaborating virtually with distributed
team members. Such factors are critical in a dispersed team in which
members are separated from one another and potentially face challenges
unique to their local work environment. Each team member must therefore
self-regulate behavior in ways that promote effective virtual collaboration

164 PERSONNEL PSYCHOLOGY

with teammates (Bell & Kozlwoski, 2002; Mohrman, 1999). Based on
this theoretical perspective, we focus on VT-SJ as an important team
member attribute that allows a team member to engage in more effec-
tive virtual collaboration behaviors and conceptualize empowering team
leadership as a team contextual factor that strengthens this relationship.
More specifically, we predict that empowering leadership and VT-SJ will
interact to influence team member virtual collaboration. We next consider
further both empowering leadership and VT-SJ before describing their
joint relevance to collaboration in geographically dispersed teams.

Empowering leadership in geographically dispersed teams. Empower-
ing leadership has been defined as leader behaviors that involve sharing
power with subordinates, raising their level of intrinsic motivation, and
creating a supportive environment for team members to leverage the power
afforded them (Arnold et al., 2000; Srivastava et al., 2006; Zhang & Bartol,
2010). These behaviors include leading by example, participative decision
making, coaching, informing, and showing concern (Arnold et al., 2000).
Empowering leadership is consistent with the distributed leadership ap-
proach in the theoretical perspective put forth by Bell and Kozlowski
(2002). It is also congruent with commentary by other researchers who
have proposed that leader behaviors that share or distribute influence are
likely to be particularly functional in a teamwork environment character-
ized by dispersion of team members and the attendant lack of face-to-face
contact (Hertel et al., 2005; Kahai, Sosik, & Avolio, 2004; Pearce et al.,
2004). For instance, Hertel et al. (2005) also noted the difficulty for
leaders of maintaining close control when team members are dispersed
and suggested using principles of delegation to shift some influence to
team members.

Team member virtual teamwork situational judgment. Bell and
Kozlowski (2002) noted that effective virtual teamwork requires team
member characteristics related to collaborating effectively in a virtual en-
vironment. Consistent with this stance, Hertel et al. (2005) and others
(e.g., Kirkman et al., 2012; Shin, 2004) have pointed to the need to focus
more research attention on individual attributes that are directly related
to being able to function effectively in a dispersed, technology-mediated
team environment. Team member virtual teamwork situational judgment
(VT-SJ) is an individual characteristic that is particularly relevant to op-
erating effectively in a dispersed team context.

In general, situational judgment refers to the extent to which an indi-
vidual has knowledge about how to deal most effectively with everyday
situations encountered in a particular work context and the ability to ap-
ply that knowledge to formulate an appropriate response to situations
that arise (for a review, see Chan, 2006). Individuals with high situa-
tional judgment tend to be more effective at identifying and responding to

HILL AND BARTOL 165

situational cues in a particular domain and, therefore, are better positioned
to respond to situational demands that they encounter in their daily work.
A growing body of research supports using situational judgment tests to
assess situational judgment related to specific work domains. That re-
search has shown that domain-specific situational judgment accounts for
incremental validity in predicting performance-related behaviors in that
domain over measures of cognitive ability and other common individual
measures, such as personality measures (for a review, see meta-analysis
by McDaniel, Hartman, Whetzel, & Grubb, 2007). Related to this, per-
formance theory recognizes the importance of both having job-related
knowledge and being able to apply that knowledge to different job situ-
ations (Campbell, McCloy, Oppler, & Sager, 1993; McCloy, Campbell,
& Cudeck, 1994). In this study, we focus on virtual teamwork situational
judgment (VT-SJ), which reflects an individual?s knowledge about the
challenges of technology-mediated collaboration in dispersed teams and
appropriate courses of action in situations that commonly arise when
working with others in such teams. We next consider the role of a team
member?s VT-SJ, in conjunction with empowering team leadership, for
promoting effective virtual collaboration behaviors.

The interactive effect of empowering leadership and VT-SJ. Based on
past research reviewed above that has shown the positive impact of sit-
uational judgment related to a particular work domain on performance-
related behaviors in that domain, we expect a team member with a higher
level of VT-SJ to be better equipped to formulate effective responses to
challenges encountered in geographically dispersed teamwork situations,
and hence to be more effective in collaborating virtually with distributed
teammates. Further, we expect a higher rather than lower empowering con-
text to play a facilitating and enabling role in strengthening the connection
between VT-SJ and virtual collaboration.

In considering what constitutes virtual collaboration, Hertel et al.
(2005) reviewed the literature to propose a team competency model for
virtual teamwork consisting of categories of behaviors that should be par-
ticularly functional for effective interactions under technology-mediated,
geographically dispersed team circumstances. In addition, reviews by
other researchers (e.g., Axtell et al., 2004; Kirkman et al., 2012; Powell,
Piccoli, & Ives, 2004; Shin, 2004) point to similar categories of behaviors.
First, virtual collaboration requires that a team member uses technology
appropriately so as to communicate virtually with distributed teammates in
a way that reduces the increased potential for misunderstandings, negative
attributions, and adverse impact on the development of shared understand-
ing among team members (Cramton, 2001; Hinds & Weisband, 2003).
This need derives from the fact that the greater reliance on technology
to communicate due to team member dispersion (O?Leary & Cummings,

166 PERSONNEL PSYCHOLOGY

2007; Kirkman & Mathieu, 2005) reduces contextual and nonverbal cues
that help to clarify the intended meaning of messages (Daft & Lengel,
1986). Second, virtual collaboration also involves taking the initiative to
interact with others in a highly supportive and responsive manner (Jar-
venpaa & Leidner, 1999) in order to overcome the coordination missteps
that can result from geographically dispersed team situations (O?Leary &
Cummings, 2007) and to build task-based trust, a primary source of trust
in a virtual team (Jarvenpaa & Leidner, 1999). This type of interaction
is characterized by frequent, predictable, and supportive communication;
substantive responses to requests for information and input; and consis-
tently meeting commitments (Jarvenpaa & Leidner, 1999). Finally, virtual
collaboration requires that a team member works constructively across the
boundaries in a dispersed team resulting from differences in team member
perspectives and work approaches associated with distribution across mul-
tiple work locations (Hinds & Bailey, 2003; Hinds & Mortensen, 2005).
Bridging these differences is notably more challenging in a distributed
team environment due to the diversity of contexts typically involved (Bell
& Kozlowski, 2002; Cramton, 2001; Jarvenpaa & Leidner, 1999; O?Leary
& Cummings, 2007).

While noting the importance of individual attributes for aiding virtual
collaboration in dispersed teams, Bell and Kozlowski (2002) also suggest
that a team member who is able to share power with the leader will be
better positioned to self-regulate performance by applying virtual team-
work relevant attributes to collaborate virtually with others in the team.
This notion is consistent with research suggesting that empowering lead-
ership creates an environment that motivates and facilitates the process
of team members utilizing their capabilities to work more effectively in
the team (Chen & Kanfer, 2006; Chen et al., 2011). It also aligns with
general person-situation interactionist theoretical perspectives proposing
that situational factors can limit the extent to which individual differences
result in expected behaviors that are consistent with those differences (for
a review, see Meyer, Dalal, & Hermida, 2010). Finally, it is congruent
with propositions and evidence from team research showing that team
contextual factors can act as facilitators of or constraints on team member
processes (Chen & Kanfer, 2006; Chen et al., 2007)

We predict that a team member operating in a team context character-
ized by high levels of empowering team leadership is likely to use VT-SJ
more effectively to collaborate virtually with team members than one op-
erating in a low empowering leadership team context. There are several
reasons why empowering leadership is likely to positively moderate the
relationship between VT-SJ and a team member?s virtual collaboration.
High rather than low empowering team leadership entails more modeling
of appropriate actions, giving team members more examples that they

HILL AND BARTOL 167

can adapt using their virtual teamwork situational judgment in order to
improve collaboration with others. The greater degree of participative de-
cision making associated with higher empowering team leadership also
affords a team member more latitude to use virtual teamwork situational
judgment in collaborating virtually. Further, more coaching on the part of a
high empowering team leader should encourage greater use of situational
judgment capabilities with virtual collaboration. Finally, it is easier to in-
corporate VT-SJ to formulate effective responses to virtual collaboration
situations when team leadership is more rather than less empowering be-
cause such a leader provides an individual with more relevant information
and shows greater concern and support for the team member?s actions. In
summary, we expect that VT-SJ is more likely to translate into effective
team member virtual collaboration when team empowering leadership is
high rather than low. These individuals, when functioning under a high
level of empowering leadership, should have the power to use their knowl-
edge and ability to develop more appropriate responses to the challenges
of distributed teamwork they encounter as compared to their counterparts
in low empowering leadership team contexts.

Hypothesis 1: Empowering leadership moderates the positive relation-
ship between team member virtual teamwork situational
judgment (VT-SJ) and a team member?s virtual collab-
oration, such that this relationship is stronger when em-
powering leadership is high.

Team Member Performance

We predict that more effective virtual collaboration with others in
a dispersed team should improve a team member?s performance in the
team. Past theoretical and empirical research has shown that, in teamwork
settings, working well with others and responding to their needs is a means
by which individuals can achieve a higher level of individual performance
(e.g., Barry & Stewart, 1997; Farh, Seo, & Tesluk, 2012; Shaw, Duffy, &
Stark, 2000; Welbourne, Johnson, & Erez, 1998). Hence, in a dispersed
team, a team member?s ability to engage in collaborative behaviors that
address the demands of distributed teamwork should enhance a team
member?s performance.

We have proposed earlier that the influence of team member VT-SJ
on team member virtual collaboration is partially contingent on empower-
ing team leadership (Hypothesis 1). Considered in combination with the
expectation that team member virtual collaboration positively relates to in-
dividual team member performance, this suggests that there is an indirect
(mediated) relationship between VT-SJ and team member performance

168 PERSONNEL PSYCHOLOGY

through team member virtual collaboration that is contingent on the level
of team empowering leadership. This type of relationship is commonly
referred to as a conditional indirect effect or moderated mediation effect
(Edwards & Lambert, 2007; Preacher et al., 2007). Further, because we
have predicted that high empowering leadership strengthens the positive
relationship between VT-SJ and team member virtual collaboration, this
indirect effect should be stronger at higher than at lower levels of empow-
ering team leadership. This cross-level moderated mediation hypothesis
constitutes Hypothesis 2.

Hypothesis 2: Empowering team leadership moderates the positive in-
direct effect of team member virtual teamwork situa-
tional judgment (VT-SJ) on team member performance
through team member virtual collaboration, such that
this indirect effect is stronger at higher levels of empow-
ering leadership.

Empowering Leadership, Team Geographic Dispersion, and Team
Virtual Collaboration

In addition to its cross-level influence on the relationships linking team
member VT-SJ to team member collaboration and performance, we also
consider the impact of empowering leadership on the aggregate level of
virtual collaboration enacted by members of the team. We henceforth refer
to aggregate virtual collaboration at the team level as team virtual collabo-
ration to distinguish it from individual team member virtual collaboration
discussed in the previous section. We expect empowering leadership to
have a direct positive impact on team virtual collaboration that is moder-
ated by the team?s geographic dispersion. These predictions are consistent
with Bell and Kozlowski?s (2002) theorizing related to leadership in dis-
persed teams.

There are several reasons why empowering leader behaviors might
help to promote effective collaboration behaviors in dispersed teams. For
instance, participative leadership along with leader coaching and sup-
port can provide the team with the leeway and confidence to experi-
ment in finding ways to effectively use technology for communicating
virtually within the dispersed team (Colquitt, LePine, Hollenbeck, Il-
gen, & Sheppard, 2002; Kirkman, Rosen, Tesluk, Gibson, 2004). Em-
powering team leadership has also been shown to foster a collaborative
team context (Srivastava et al., 2006), which should make team members
more responsive to each other and more willing to take the initiative to
help the team. Finally, empowering leadership can be expected to solicit
behaviors associated with collaborating across the differences that exist in

HILL AND BARTOL 169

geographically dispersed teams and leveraging the different perspectives
and work approaches members bring to the team. The aforementioned
collaborative context and spirit of experimentation engendered by em-
powering leadership behaviors, as well as the consideration for others
that results when empowering leaders show concern for dispersed team
members, should result in team members being more likely to seek and
value each other?s ideas and perspectives. It should also result in a greater
tendency for team members to work to mitigate the potentially dysfunc-
tional conflicts that can arise in the team as a result of team-member
and work-context differences. More generally, through leading by exam-
ple, empowering leaders can model appropriate behaviors for interaction
among dispersed team members. Given the likelihood of greater diversity
of members and work practices in dispersed teams, as well as the reduced
opportunity for face-to-face interaction, such modeling along with team
leader coaching behaviors should be useful in signaling and creating a
shared understanding of effective norms and patterns of interaction.

With regard to the interaction between empowering leadership and
team geographic dispersion, past research has shown that the level of
geographic dispersion in a team is an important contingency that deter-
mines the degree of team relevance of leadership behaviors (Cummings,
2008; Gajendran & Joshi, 2012; Hoch & Kozlowski, 2014; Joshi et al.,
2009; Morgeson, DeRue, & Karam, 2010). This is because geographic
dispersion in a team can be considered as a continuum and teams that are
higher on this continuum tend to experience greater challenges that can
complicate team processes and undermine the production of needed out-
comes (Cramton & Webber, 2004; Cummings, 2008; Gajendran & Joshi,
2012; Gibson & Gibbs, 2006; Hinds & Mortensen, 2005; Joshi et al.,
2009; O?Leary & Cummings, 2007; Schweitzer & Duxbury, 2010). Re-
lated theoretical perspectives that support the prescription of team leaders
sharing leadership functions with distributed team members also propose
that these leader behaviors will increase in importance as team dispersion
increases (Bell & Kozlowski, 2002; Hertel et al., 2005). Yet the amount
of research that has considered the geographic dis