Virtual Team Leadership: The Effects of Leadership Style and Communication Medium On Team Interaction Styles and Outcomes

Organizational Behavior and Human Decision Processes 103 (2007) 1–20
www.elsevier.com/locate/obhdp
Virtual team leadership: The effects of leadership style

and communication medium on team interaction styles and outcomes
a,*
Laura A. Hambley , Thomas A. O’Neill b, Theresa J.B. Kline c
a
University of Calgary, 182 Royal Birkdale Cres, NW, Calgary, Alta., Canada T3G 5R8
b
University of Western Ontario, Canada
c
University of Calgary, Canada
Received 11 August 2005

Available online 7 November 2006
Abstract
This study investigated the effects of transformational and transactional leadership styles and communication media on team
interaction styles and outcomes. Teams communicated through one of the following three ways: (a) face-to-face, (b) desktop vid-
eoconference, or (c) text-based chat. Results indicated that transformational and transactional leadership styles did not affect team
interaction styles or outcomes; that the mean constructive interaction score was higher in FTF than videoconference and chat teams,
but not significantly higher in videoconference than chat teams; and that teams working in richer communication media did not
achieve higher task performance than those communicating through less rich media. Finally, mean team cohesion scores were higher
in FTF and videoconference than chat teams, but not significantly higher in FTF than videoconference teams. These results provide
further evidence that communication media do have important effects on team interaction styles and cohesion.
2006 Elsevier Inc. All rights reserved.
Keywords: Virtual leadership; Virtual teamwork; Virtual teams; Teams; Leadership; e-Leadership; Dispersed teams; Virtual work; Communication
media; Videoconferencing
Rapid technological advancements have led to a this paper, the term virtual leadership will refer to for-
new paradigm of work—it can now be conducted any- mal leadership by one individual, as opposed to being
time, anywhere, in real space or through technology shared amongst team members.
(Cascio & Shurygailo, 2003). The virtual environment In addition to leadership, teamwork has been affect-
and its various communication technologies have created ed by the virtual environment. Virtual teams are
a new context for leadership and teamwork (Avolio, becoming a more common type of work unit and are
Kahai, Dumdum, & Sivasubramaniam, 2001a). Leader- expected to play an increasingly key role in organiza-
ship within this new context has been referred to as tions (Hertel, Konradt, & Orlikowski, 2004; Lipnack
‘‘e-leadership’’ or ‘‘virtual leadership,’’ defined as ‘‘a & Stamps, 2000). For purposes of this paper, virtual
social influence process mediated by advanced informa- teams are defined as interdependent groups of individ-
tion technologies to produce changes in attitudes, feel- uals that work across time, space, and organizational
ings, thinking, behaviour, and/or performance of boundaries with communication links that are heavily
individuals, groups, and/or organizations’’ (Avolio, dependent upon advanced information technologies
Kahai, & Dodge, 2001b, p. 617). For the purposes of (Driskell, Radtke, & Salas, 2003; Thompson & Coo-
vert, 2003).
*
Corresponding author. Fax: +1 403 282 8249. Virtual team leadership is considered highly impor-
E-mail address: [email protected] (L.A. Hambley). tant to virtual team performance (Hambley, O’Neill, &
0749-5978/$ - see front matter 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.obhdp.2006.09.004
2 L.A. Hambley et al. / Organizational Behavior and Human Decision Processes 103 (2007) 1–20
Kline, in press; Hertel, Geister, & Konradt, 2005); a effects of leadership styles on outcomes across virtual
claim that has been substantiated in a recent field study and FTF teams. Furthermore, the inclusion of desktop
on virtual teams (Webster & Wong, 2003). The authors videoconference provided a means to compare leader-
reported that employees perceived leadership as a critical ship in a medium that has not been considered in previ-
factor of geographically distributed team success. ous leadership research. Finally, this study enables a
Despite the widespread increase in virtual teamwork, comparison of the effects of communicating through
researchers do not clearly understand the implications videoconference to FTF and chat on team interaction
of advanced information technology for leadership prac- styles and outcomes.
tices (Zaccaro & Bader, 2002). Furthermore, there exists
little evidence in support of or against using more
advanced technologies to improve leadership perfor- Leadership
mance (Avolio et al., 2001b). More research is therefore
needed to better understand leadership of virtual teams. Research has demonstrated that leaders can make a
Although some research on virtual leadership styles is critical difference to team performance and effectiveness
emerging, there is still a need for research assessing how (e.g., Morgeson, 2005; Zaccaro & Klimoski, 2002).
certain leadership styles interact with communication Indeed, Salas, Sims, and Burke’s (2005) recent review
technologies to affect team processes and outcomes. and model of teamwork included team leadership as
While there are many different approaches to the study one of the ‘‘Big Five’’ contributors to team effectiveness.
of leadership, transformational and transactional lead- Elsewhere, it has been argued that leaders play impor-
ership styles will be investigated in the present study. tant roles in modeling effective teamwork, and in setting
This theoretical basis was chosen for several reasons: ground rules for team members to engage in successful
first, it has received extensive research support since its team processes (Cascio & Shurygailo, 2003). In short,
inception (Bass, 1985; Bass & Avolio, 1993). Second, leadership appears to be an integral part of effective
the transformational/transactional paradigm can trans- teamwork.
late clearly into practical recommendations and sugges- There are several theoretical approaches to the study
tions for leadership training (Barling, Weber, & of leadership, but one well-known and contemporary
Kelloway, 1996; Kelloway & Barling, 2000). Third, framework is transformational/transactional leadership.
these two leadership styles have been shown to impact Referred to as the new leadership paradigm (Bryman,
virtual teams in meaningfully different ways (e.g., Sosik, 1992), charismatic leadership (House, 1996), transfor-
Avolio, Kahai, & Jung, 1998). Finally, psychometrically mational leadership (Bass, 1985), or the full range lead-
sound measures exist that accurately assess these leader- ership theory (Sivasubramaniam, Murray, Avolio, &
ship styles (e.g., Multifactor Leadership Questionnaire; Jung, 2002), the distinction between transformational
Bass & Avolio, 1990). and transactional leadership has received a great deal
Several studies have assessed leadership in computer- of research attention (Dvir, Eden, Avolio, & Shamir,
mediated team contexts (e.g., Hoyt & Blascovich, 2003; 2002). Specific to the purposes of the present study,
Kahai, Sosik, & Avolio, 2003), but few studies have transformational and transactional leadership styles
manipulated leadership or used more than one commu- have been the focus of many existing studies on teams
nication medium to determine how technology facili- communicating through technologies (Sivasubraman-
tates or negates the effects of leadership in virtual iam et al., 2002).
contexts. Kerr and Jermier (1978) noted that situational Transactional leaders view the leader-follower rela-
variables (e.g., technology) may impact the effectiveness tionship as a process of exchange (Bass & Avolio,
of leader behaviours, but how technologies moderate the 1993). They tend to gain follower compliance by either
effects of leadership styles on virtual teamwork has not offering rewards or threatening punishment. Two main
been specifically addressed. Nonetheless, team leaders leadership factors are characteristic of this leadership
are being asked to lead geographically dispersed teams, style: (1) contingent reward, and (2) management by
and research is needed to determine the best ways to exception. In contrast, transformational leadership
do so. focuses on motivating and inspiring followers to per-
To better understand the effects of leadership styles form beyond expectations and comprises four main
and communication technology on team processes and factors: (1) idealized influence (or ‘‘charisma’’), (2) inspi-
outcomes, we conducted a laboratory study to carefully rational motivation, (3) intellectual stimulation, and (4)
control for and manipulate leadership style and commu- individualized consideration (Bass & Avolio, 1993).
nication medium. Specifically, we compared the effects Transformational leaders are skilled at increasing and
of transformational and transactional leadership within broadening follower interests, gaining commitment to
teams communicating face-to-face (FTF), through vid- the goals and mission of the group/organization, and
eoconference, or text-based chat (chat). Our study con- motivating people to go beyond their self-interests for
tributes to the extant literature by examining the the good of the group (Den Hartog & Koopman, 2001).
L.A. Hambley et al. / Organizational Behavior and Human Decision Processes 103 (2007) 1–20 3
Some research has found that transformational lead- gies continues to grow as they become more accessible
ership is more effective than transactional leadership and their cost declines (Baker, 2002; Spreitzer, 2003).
(Bass, 1997). Bass suggested that highly effective leader- Two commonly used theoretical frameworks for com-
ship needs to go beyond the reward-punishment paring the effects of different communication technolo-
exchange that typifies transactional leadership. Also, gies are media synchronicity (Dennis & Valacich,
Jung (2001) found that FTF teams led by a transforma- 1999) and media richness (Daft & Lengel, 1986), each
tional leader outperformed their counterparts that were of which is described below.
led by a transactional leader on a brainstorming task. According to media synchronicity theory, two types
In another FTF study, transformational leadership of communication encompass virtual team interaction:
predicted empowerment, group cohesion, and perceived synchronous and asynchronous. Synchronous interac-
group effectiveness (Jung & Sosik, 2002). Finally, Lim tion occurs when team members communicate at the
and Ployhart (2004) conducted a study using combat same time (in real time), such as through teleconferenc-
teams, and reported that transformational leadership ing, videoconferencing, or chat sessions (Avolio et al.,
was strongly related to team performance in both typical 2001a). ‘‘Synchronous’’ communication media allow
and maximum FTF contexts. individuals to work on the same task, with the same
In addition to individual studies, meta-analyses have information, at the same time (Baker, 2002; Dennis &
also been conducted on transformational and transac- Valacich, 1999). ‘‘Asynchronous’’ interaction involves
tional leadership. One meta-analysis found that, team members communicating at different times, such
although both styles related positively to performance, as in the case of e-mail or threaded discussions.
the relationship with transformational leadership was sig- The degree to which a communication medium allows
nificantly stronger (Lowe, Kroeck, & Sivasubramaniam, for synchronous communication affects the flow of con-
1996). A more recent meta-analysis (Judge & Piccolo, versation (Sellen, 1995). Communication technologies
2004), however, found that transformational leadership that more closely approximate FTF interactions, such
was indeed an effective style, but the contingent reward as videoconference, would be closer in conversational
aspect of transactional leadership was also highly effec- style to an in-person meeting (O’Connail, Whittaker, &
tive. Additionally, these two leadership styles tended to Wilbur, 1993). As tasks become more complex and
be highly correlated. require interdependence, reciprocal communication,
In sum, both transactional and transformational and feedback among team members, synchronous media
leadership styles have been linked to effective perfor- are found to be more effective than asynchronous media
mance in face-to-face teams, with transformational (Bell & Kozlowski, 2002; Maruping & Agarwal, 2004).
leadership tending to be more effective overall. Despite Media richness theory is a commonly used theory for
the overall advantage of transformational leadership, explaining how different communication media affect
however, there are certain conditions when the contin- task performance (Daft & Lengel, 1986). A rich medium
gent reward aspect of transactional leadership may be allows for: (1) transmitting multiple verbal and nonver-
more strongly associated with positive work outcomes bal cues, (2) using natural language, (3) providing imme-
(e.g., job performance; Judge & Piccolo, 2004). To diate feedback, and (4) conveying personal feelings and
further complicate matters, research focusing on the emotions (Daft & Lengel). The richest medium is FTF
effectiveness of transactional and transformational communication, followed by telephone, chat, e-mail,
leadership within the context of virtual teams is rela- and print communications. Newer technologies, such
tively scarce (for exceptions, see Hoyt & Blascovich, as videoconference, are thought to rank above the tele-
2003; Sosik et al., 1998), though leadership style and phone, but below FTF communication in terms of
technology are likely to interact in complex ways. media richness (Webster & Hackley, 1997). Technolo-
Thus, the present study investigated how leadership gies that allow the recipient(s) to see physical gestures
styles affected virtual team outcomes depending on and facial expressions are thought to increase the rich-
the type of communication medium used. The follow- ness of the information conveyed (Fletcher & Major,
ing section discusses communication media in the con- 2003), thereby contributing to the greater richness of
text of virtual teams to give the reader an videoconference as compared to telephone or text-based
understanding of the types of media typically studied media.
and the observed effects. Most studies to date have typically compared
FTF with only one type of communication medium
(Staples & Webster, 2003). Although researchers
Communication technologies and virtual teams acknowledge videoconference as another alternative to
text-based media, they are usually not able to include it
In the case of virtual teams that never or rarely meet in their studies (e.g., Jarvenpaa & Leidner, 1999), and
FTF, communication technologies are vital for collabo- therefore cannot determine if increments in media rich-
ration (Hollingshead, 2004). The use of such technolo- ness will result in incremental effects on team outcomes.
However, comparing several communication media with variable. Likewise, a more recent meta-analysis examin-
each other will help increase our understanding of the ing the effects of task complexity in virtual teams still did
technologies that allow virtual teams to collaborate most not find enough studies of videoconference to include
effectively (Baker, 2002; Baltes, Dickson, Sherman, Bau- this medium (Wiggins & Horn, 2005). Videoconference
er, & LaGanke, 2002). For example, the present study is, however, becoming an increasingly common method
provided the opportunity to assess whether or not there of team communication for geographically dispersed
was a linear relationship between media richness and teams, and requires more research to examine its effec-
team outcomes because we employed more than one tiveness (Bell & Kozlowski, 2002; Spreitzer, 2003).
medium on a continuum of richness. The benefits of videoconference over less rich com-
munication media, such as chat, are becoming appar-
Text-based communication ent in the literature. In addition, desktop
videoconference (conducted between two or more peo-
The majority of studies that have examined comput- ple directly from each of their desktop computers) is
er-mediated communication have used text-based sys- becoming an affordable communication option for a
tems (Baltes et al., 2002). Although the most common wide variety of users and organizations (Townsend,
form of text-based communication is e-mail (Pulley, Demarie, & Henderickson, 2001). Beyond being a cost
Sessa, & Malloy, 2002), most research has instead effective substitute for FTF communication, it is criti-
focused on synchronous communication technology, cal to understand how videoconference affects produc-
such as ‘‘chat’’ (Baltes et al., 2002). The logistics of tivity and virtual team collaboration. Indeed, virtual
studying asynchronous media such as e-mail are more team researchers are calling for more studies that
difficult, especially in single session laboratory studies. employ today’s advanced synchronous technologies,
Because chat is a synchronous form of communication, such as videoconference, in team decision-making
it is superior to asynchronous e-mail for intensive tasks (Bell & Kozlowski, 2002; Baltes et al., 2002).
that require considerable information sharing and col-
laboration (Balthazard, Waldman, Howell, & Atwater,
2002; Bell & Kozlowski, 2002). Virtual team leadership
A suggested benefit of chat communication is that
this type of text-based interaction may allow for more Several field and laboratory studies provide insight
reflection and the ability to choose one’s words more into how leadership styles may affect virtual team inter-
carefully than in FTF or telephone communication action and performance (e.g., Kahai & Avolio, 2006;
(Wolfe, 2002). Chat can also allow team members to Kahai, Sosik, & Avolio, 1997, 2003, 2004; Sosik,
more efficiently share ideas in brainstorming tasks Avolio, & Kahai, 1997; Sosik et al., 1998). For example,
because everyone can ‘‘speak’’ at once, thereby minimiz- Sosik et al. (1997) found that transformational leader-
ing process losses (Griffith & Neale, 2001). Additionally, ship was associated with higher levels of group potency
chat may neutralize the tendency for increased relational (the group’s belief that it can be effective) than transac-
conflict often observed in demographically dissimilar tional leadership, and that group potency was related to
groups, because these differences are less salient group effectiveness. However, transformational leader-
(Mannix, Griffith, & Neale, 2002). Conversation in this ship exerted a stronger impact on group performance
medium, however, has also been criticized for lacking in only one of two sessions. In the other session, teams
focus because multiple group members may be speaking led by transactional leaders outperformed the transfor-
at the same time (Wakertin, Sayeed, & Hightower, mational groups.
1997). Also, different rates of typing and reading can Building on their earlier work, Sosik et al. (1998)
lead to more or less delayed responses by individuals assessed the effects of transactional goal-setting and fac-
within the group discussion, and could result in low con- ets of transformational leadership on group creativity.
tributions by some members who could otherwise The results suggested that goal-setting and inspirational
improve the team’s performance. leadership positively predicted group creativity. Con-
versely, intellectual stimulation and individualized con-
Videoconferencing sideration were negatively related to creativity, likely
because they were interpreted negatively by the zero-his-
Videoconference is a relatively rich communication tory teams. For example, transformational leaders that
medium, allowing for the transmission of both verbal encourage followers to ‘‘go above and beyond’’ may
and nonverbal cues (Martins, Gilson, & Maynard, have been considered intrusive and critical in the
2004). In their meta-analysis of computer-mediated short-term groups. Nevertheless, the Sosik et al. (1998)
communication and group decision-making, Baltes study demonstrated that aspects of transformational
et al. (2002) found a paucity of studies that examined and transactional leadership can positively influence
videoconference, so they were unable to include it as a group outcomes.
A third study examining transformational versus Cooke and Szumal (1994) suggested that team inter-
transactional leadership in a virtual team setting found actions can be analyzed in terms of two general styles:
that groups working with a transactional leader had ‘‘constructive’’ and ‘‘defensive.’’ A constructive interac-
higher levels of group efficacy and task satisfaction than tion style is characterized by a balanced concern for per-
groups working with a transformational leader when sonal and group outcomes, cooperation, creativity, free
individual inputs were identified (Kahai et al., 2003). information exchange, and respect for others’ perspec-
This advantage disappeared in the anonymous condi- tives. Comparatively, defensive styles include passive
tion, which highlights the role of anonymity in enhanc- and/or aggressive behaviours. Passive behaviours include
ing the effects of transformational leadership (this was limited information sharing, lack of questioning, lack of
also observed by Sosik et al., 1997). impartiality, lack of creative thinking, greater emphasis
Collectively, there are at least three learnings based solely on fulfillment of affiliation goals, and maintaining
on this literature review. First, the transactional/trans- harmony. Aggressive behaviours place greater emphasis
formational paradigm predicts team processes and out- on personal agendas and ambitions above concern for
comes in both FTF and virtual teams. Second, neither the group outcome. Examples of aggressive behaviours
of these leadership styles consistently outperforms the are competition, criticism, interruptions, and impatience
other, possibly because aspects of both leadership styles for others’ perspectives.
are important depending on contextual details Whether a team’s dominant interaction style is con-
(Kirkman, Rosen, Gibson, Tesluk, & McPherson, structive or defensive can result in different levels and
2002). patterns of effectiveness (Cooke & Szumal, 1994). Effec-
Third, transactional leadership can be more effective tiveness has been measured as both the quality and
than transformational leadership in identified condi- acceptance of the solution by team members. In terms
tions, but often the leadership styles are equalized or of quality, constructive groups have been found to most
transformational leadership becomes more effective consistently produce the highest quality solutions, fol-
in situations involving anonymity (Sosik et al., 1997, lowed by aggressive then passive groups. As for levels
1998; Kahai et al., 2003). Put another way, the data sug- of solution acceptance, constructive groups usually pro-
gest that as a communication medium becomes more duce the highest levels, followed by an equal level of
anonymous, transformational leadership may be the acceptance between aggressive and passive groups.
more effective leadership style relative to transactional Defensive styles tend to result in marginal-quality solu-
leadership. To further elaborate on these three lessons, tions that are not accepted by all team members, as well
the present study sought to determine the extent to as low levels of member satisfaction (Cooke & Szumal;
which leadership styles impact team outcomes differen- Smith, Peterson, Johnson, & Johnson, 1986; Watson &
tially in three communication media within a laborato- Michaelsen, 1988).
ry-based environment.
Virtual team interaction styles
Team and virtual team processes There are very few studies on virtual team interac-
tion styles, but results from these studies are interest-
Leadership is thought to directly impact team ing. Research has shown that communicating
processes (Zaccaro, Ardison, & Orvis, 2004), which are virtually does not substantially affect how interaction
the ways in which team members transform inputs styles impact team outcomes. Specifically, the effects
(e.g., team member contributions and skills) into outputs of FTF and virtual interaction on decision perfor-
(e.g., solutions; Cooke & Szumal, 1994). For leaders to mance and processes have been found to be direction-
impact team performance, therefore, they must success- ally consistent (Potter & Balthazard, 2002). Also, the
fully influence the team’s processes. In other words, suc- effects of virtual team interaction styles on solution
cessful team performance depends jointly on team and quality and acceptance are similar in magnitude to
leadership processes (Zaccaro & Klimoski, 2002). The FTF teams (Potter, Balthazard, & Cooke, 2000).
processes of interest in this study are team interactions, The strengths of relationships between certain inter-
or ‘‘team interaction styles’’ (Cooke & Szumal, 1994). action styles and team performance, however, have been
found to differ between FTF and virtual teams. For
Team interaction styles example, a constructive interaction style is the most
As team members work together to complete tasks, conducive to high team performance in either FTF or
their roles become highly interdependent, requiring virtual teams. A passive interaction style, however, tends
effective interactions (Balthazard et al., 2002). A team’s to result in lower performance for virtual than FTF teams,
interaction style is best understood in terms of the com- likely because it is easier to ignore other team members
munication patterns used to deal with task conflicts and in a virtual setting, and more difficult to reverse or mod-
maintenance of team member relationships (Balthazard erate passivity (Potter et al., 2000). Another interesting
et al.). difference is that aggressive interaction styles have been
found to do less damage (e.g., limiting team member with transactional leadership, on team processes and
input) in a virtual team, possibly because technology- outcomes (Bass, 1997; Lowe et al., 1996). Although both
mediated communication makes it easier for all styles have been found to be effective and correlated
members to contribute rather than being significantly (Judge & Piccolo, 2004), the literature seems to suggest
hampered by a dominating team member (Potter et al.). that transformational leadership is more effective over-
all. For instance, transformational leadership in prob-
Team outcomes lem-solving teams leads to increased enthusiasm and
confidence of team members, promoting understanding
Several studies have found that leadership can affect and appreciation of differing views, and intellectually
team performance, which may take a variety of forms stimulating members to re-examine critical assumptions
depending on the team’s tasks and purposes (e.g., Wag- and to view problems in new ways (Bass & Avolio,
eman, 2001). Both objective and subjective performance 1994). These characteristics are thought to foster con-
indicators are important in assessing team outcomes structive, as opposed to defensive, team interaction.
(Sundstrom, De Meuse, & Futrell, 1990). A team may Indeed, research has found support for a positive
perform well objectively, but due to internal conflict, relationship between shared transformational leadership
be very unwilling to work together again. Thus, it is and constructive team interactions, and a negative rela-
important to include a construct which subjectively tionship with defensive team interactions (Balthazard
captures team member satisfaction and willingness to et al., 2002). Similar findings would be expected in a sit-
work together in the future, in addition to task perfor- uation with a single transformational leader, whose
mance, when measuring overall team effectiveness behaviours would more likely foster constructive team
(Kline, 1999; Sundstrom et al.). The two types of team interaction than would those of a transactional leader.
outcomes of interest in this research are task perfor- These effects may, however, differ depending on commu-
mance and team cohesion. Task performance is the nication medium.
objectively measured outcomes of a team task or pro- Although some research has found that communica-
ject. Team cohesion, in contrast, is a dynamic process tion media interact significantly with leadership style to
that is reflected in the tendency for a group to stick affect group processes (e.g., Davis et al., 2003), more
together and remain united in the pursuits of its instru- research is needed. In particular, little is known about
mental objectives and/or for the satisfaction of member the effects of videoconference on team interaction styles.
affective needs (Carron, Brawley, & Widmeyer, 1998). Based on media richness theory, richer media allow for
Meta-analyses have demonstrated that team cohesion more verbal and nonverbal cues, and therefore more
and team performance are related (Gully, Devine, & natural team communications and decision-making
Whitney, 1995; Mullen & Copper, 1994), although the (Baltes et al., 2002). Further, richer media should enable
direction of the cause/effect relationship is unclear a leadership style to have greater effects on team interac-
(Chansler, Swamidass, & Cammann, 2003). tion, as more verbal and nonverbal cues allow the lead-
More recently, a model has been postulated that er’s behaviours to be more salient. For example, in the
provides a theoretical rationale for the effects of trans- case of transformational leadership, richer media should
formational leadership on team cohesion (Dionne, facilitate positive leadership effects on constructive team
Yammarino, Atwater, & Spangler, 2004). Specifically, interaction. In addition, less rich media may result in a
these researchers posited that transformational leader- leader’s style having less influence because of a reduction
ship will increase shared vision and team commitment, in number of messages exchanged (Avolio et al., 2001a),
which will in turn be related to increased team cohesion. which may lessen the effects of each leadership style.
Additionally, they suggested that team cohesion leads to It should be noted, however, that the present discus-
increased team task performance, implying that leaders sion is somewhat speculative. It assumes that transfor-
can have effects on team cohesion that ultimately lead mational leadership is more effective than transactional
to changes in team performance. In the present study, leadership in rich media, but that these differences will
however, we examined how leadership influenced team tend to disappear as the media becomes leaner (because
cohesion as an outcome in its own right. there is less opportunity for the leader to influence the
team). The difficulty with this assumption is that trans-
formational leadership is not consistently more effective
Hypotheses than transactional leadership in the teamwork and
virtual teamwork literature. Consider for example, a
Interaction effects of leadership styles and communication study by Sosik et al. (1997) that found support for the
media on team interaction styles superiority of transformational leadership in the first
team session, but also found that transactional leader-
Past research in FTF contexts has supported the ship was most effective in the second session. Elsewhere,
more positive effects of transformational, as compared combat team performance and ratings of cohesion have
correlated equally as strongly with platoon leaders’ authors found that only transformational leadership
transformational and transactional leadership styles predicted commitment to the team. Finally, Jung
(Bass, Avolio, Jung, & Berson, 2003). Thus, transforma- (2001) compared transformational and transactional
tional leadership may not be a better predictor of team leadership styles on a brainstorming task and found that
performance than transactional leadership in conven- transformational leadership was the stronger predictor
tional FTF or virtual settings. of performance. Taken together, there exists non-trivial
To better understand the effects of leadership in FTF support for the advantage of transformational over
and virtual settings, Kahai et al. (2003) studied leader- transactional leadership.
ship under identified (FTF) and chat conditions. Those Despite the aforementioned research, it is not known
authors found that transformational leadership was less how leadership styles may interact with communication
effective than transactional, but that this difference dis- media to affect team outcomes (i.e., performance, cohe-
appeared in the anonymous condition. The authors sion). Media richness theory argues that richer media
explained their findings by suggesting that certain facets allow for more verbal and nonverbal cue transmission,
of transformational leadership may be better suited for which may allow the leader to have greater effects on
situations where team members are not readily identi- teamwork. Comparatively, less rich media may result
fied. For example, intellectual stimulation encourages in a leader’s style having less influence, and may also
individuals to arrive at better solutions. However, ask- reduce the amount of positive interaction (Avolio
ing team members to work towards a better solution et al., 2001a), both of which may lessen the effects of
may be viewed as critical and intrusive in an identified leadership styles on team outcomes. This rationale sug-
(and short-term) context. Conversely, if an individual gests that transformational leadership may be more
feels more anonymous, he/she may accept these com- effective than transactional leadership in FTF settings,
ments as being more constructive and less personal, but that the differences will diminish as the medium
resulting in an increased focus on the task at hand. This becomes leaner (e.g., chat).
rationale suggests that transformational leadership will Similar to the discussion for Hypothesis 1, the present
be the most effective leadership style under leaner (and argument implies that as media becomes richer, the effec-
more anonymous) conditions, but as media richness tiveness of transformational leadership relative to trans-
increases, transactional leadership will become the better actional leadership will increase as well. However, this
style. Accordingly, we postulated the following assertion may not be warranted. Judge and Piccolo’s
hypothesis: (2004) meta-analytic findings suggested that the most
effective leadership style depends on the outcome. For
Hypothesis 1. Transactional leadership will more
example, transformational leadership was a better pre-
strongly predict a constructive interaction style than
dictor of follower satisfaction with the leader, while
will transformational leadership in richer media (i.e.,
transactional leadership was a stronger predictor of
FTF), but this effect will decrease as the medium
leader job performance. Thus, both transformational
becomes leaner (i.e., in videoconference and chat).
and transactional leadership styles are effective predic-
As noted, there is evidence that transformational tors of certain criteria.
leadership tends to be more effective than other leader- Laboratory research using paradigms similar to that
ship styles (Dvir et al., 2002). For example, Howell and of the present study have also found that transforma-
Frost (1989) found that subjects working under a tional leadership may not always predict team out-
transformational leader had higher task performance comes more strongly than transactional (e.g., Sosik
and reported greater levels of task adjustment com- et al., 1997, 1998). Despite the clear advantage of
pared to individuals that were led by a structuring or transformational leadership in one session of a two-
considerate leader. Furthermore, a field study by part study, Sosik et al. (1997) found that transactional
Howell and Avolio (1993) found that facets of trans- was more effective in the second session of the same
formational leadership were positively associated with experiment. Furthermore, in a follow-up study specific
business-unit performance, but that contingent reward facets of transformational leadership (i.e., inspirational
and active management by exception facets of transac- leadership behaviours) were related positively to team
tional leadership were negatively related to perfor- effectiveness (Sosik et al., 1998). However, other
mance. The authors concluded that ‘‘the more aspects (i.e., intellectual stimulation and individualized
positive contribution to unit performance came from consideration) were negatively associated with team
the behaviours associated with transformational leader- outcomes.
ship’’ (Howell & Avolio, 1993, p. 899). Kahai et al. (2003) provided an explanation for these
In another investigation of transformational and results based on their own study. Recall that those
transactional leadership, 68 managers of global virtual authors found that transactional leadership was more
teams from countries throughout the world were com- effective in an identified, (as opposed to anonymous)
pared on key team outcomes (Davis et al., 2003). The condition. In leaner and more anonymous media,
differences between the two leadership styles were not Past research has demonstrated direct effects of com-
found. The authors argued that certain aspects of trans- munication media on team cohesion. Specifically, Straus
formational leadership (e.g., intellectual stimulation) and McGrath (1994) and Wakertin et al. (1997) found
may be interpreted as criticisms in identified (e.g., that FTF teams are more cohesive than virtual teams
FTF) conditions. Under greater anonymity (e.g., chat), communicating through synchronous and asynchronous
transformational behaviours may challenge group mem- text-based media, respectively. Moreover, a meta-analy-
bers (instead of threaten them) to exert effort in a more sis by Baltes et al. (2002) found that computer-mediated
positive way. Given the similarity of the present research (text-based) teams performing interdependent tasks
design with that of Sosik et al. (1997, 1998) and Kahai were usually less satisfied than their FTF counterparts.
et al. (2003), we suspect that comparable results will The lower satisfaction may have been due, in part, to
be found. Thus, we propose the following: frustration with the fact that typing takes much longer
than verbal communication (Wiggins & Horn, 2005).
Hypothesis 2a. Transactional leadership will predict
However, one cannot assume FTF teamwork leads to
team performance more strongly than will transforma-
more cohesive teams than virtual teamwork. Instead, it
tional leadership in richer media (i.e., FTF), but this
may depend on the specific type of communication
effect will decrease as the medium becomes leaner (i.e., in
media used by the virtual team, with richer media lead-
videoconference and chat).
ing to greater cohesion. Extending past research findings
to include videoconference, the following hypothesis is
Hypothesis 2b. Transactional leadership will predict
offered:
team cohesion more strongly than will transformational
leadership in richer media (i.e., FTF), but this effect will Hypothesis 5. Teams interacting through richer com-
decrease as the medium becomes leaner (i.e., in video- munication media will report higher levels of team
conference and chat). cohesion than teams interacting through less rich
communication media, such that the mean team cohe-
sion score will be: (a) higher in FTF than videoconfer-
Effects of communication media on team interaction styles
ence teams, (b) higher in FTF than chat teams, and (c)
and outcomes
higher in videoconference than chat teams.
The medium through which teams communicate has Similar findings are expected in regards to the effects
been found to significantly affect their interaction styles of communication media on task performance,
(Balthazard et al., 2002). Both media synchronicity and although past research has been somewhat mixed.
media richness theories suggest that FTF communication Most studies have either found no differences in task
is the most effective medium for team interaction, but performance between virtual and FTF teams or that
other media can also be effective depending on task com- virtual teams tend to perform more poorly than FTF
plexity (McGrath, Hollingshead, & O’Connor, 1993). teams on interdependent team tasks (Baltes et al.,
Although other communication media, such as videocon- 2002; Wiggins & Horn, 2005). Interdependent team
ference and chat are also synchronous, they are less rich tasks require a high degree of accurate message trans-
than FTF communication, with videoconference being mission and reciprocal exchanges among team mem-
richer than chat. Based on the expectation that richer bers. Participants using less rich media may not
media will allow for communicating greater verbal and transfer messages as efficiently as their FTF counter-
non-verbal cues, and will lead to more effective team parts because of media limitations (Thompson & Coo-
interactions, the following hypotheses are offered: vert, 2006). Because the present study used a highly
interdependent team task, it was expected that task
Hypothesis 3. Richer communication media will result
performance would increase using richer media. This
in a higher constructive team interaction style than less
expectation is due to the greater ease of communicating
rich media, such that the mean constructive interaction
and collaborating in richer, synchronous media. Thus,
style will be: (a) higher in FTF than videoconference
the following hypothesis is offered:
teams, (b) higher in FTF than chat teams, and (c) higher
in videoconference than chat teams. Hypothesis 6. Teams interacting through richer com-
munication media will display greater task perfor-
Hypothesis 4. Richer communication media will result mance than teams interacting through less rich
in a lower defensive team interaction style than less rich communication media, such that the mean task
media, such that the mean defensive interaction style performance score will be: (a) better in FTF than
will be: (a) lower in FTF than videoconference teams, videoconference teams, (b) better in FTF than chat
(b) lower in FTF than chat teams, and (c) lower in vid- teams, and (c) better in videoconference than chat
eoconference than chat teams. teams.
Methods six items that require an additional choice of two possible

options for each. Furthermore, this task tends to gener-
Research design ate engaging discussions because the topic (i.e., attending
workplace and/or university meetings) was relevant to
This study used a 2 (transactional/transformational the participants in the sample. In contrast, most partici-
leadership) · 3 (chat/videoconference/FTF communica- pants have not had experience with the content of group
tion media) factorial design. Sixty experimental teams, survival-type tasks such as being stranded on the moon
with three to four participants in each team, were ran- or in the desert (e.g., Desert Survival Situation; Balthaz-
domly assigned across these six conditions. Each team ard, 1999b). In terms of complexity, according to the
was led by a confederate leader displaying one of the typology of Van de Ven, Delbecq, and Koenig (1976),
two leadership styles. the MES would be considered intensive. Specifically, it
is intensive because it involves an interdependent
Participants arrangement in which team members diagnose, problem
solve, and/or collaborate simultaneously as a team to
Participants in this study were 228 undergraduate accomplish the task.
students from a mid-sized Canadian University. Fifty- The MES requires participants to identify the optimal
one percent of participants were psychology majors, sequence for carrying out 20 activities involved in con-
with the rest majoring in a variety of other disciplines ducting a successful meeting (e.g., ‘‘decide who should
(e.g., biological sciences, nursing, business). On average, attend the meeting and why’’; ‘‘set approximate time
participants had completed 2.8 years of their degree. limits for discussion of specific agenda items’’). Solu-
The mean age for the sample was 23.8 years (range: tions to the MES are first developed on an individual
17–51 years), and the sample was 87% female. basis, and then as a team. Individual and team solutions
All participants were recruited through the psycholo- are then compared to the experts’ solution, provided by
gy bonus credit system (i.e., they were enrolled in at least Human Synergistics International (Human Synergistics,
one psychology course at the time of this study), thereby 2002). This provides an objective solution with which to
receiving partial course credit in exchange for their par- compare individual and team solutions and addresses a
ticipation. Prior to signing up for the study, participants weakness of past research; that of subjective measure-
were told not to sign up with friends, as this study ment of team solution quality (Huang, Wei, Watson,
attempted to have all zero-history teams. It was con- & Tan, 2002). Comparisons between individuals’ and
firmed that teams were zero-history based on negative their team’s scores indicates whether the team was able
responses to the demographic item ‘‘were any of the to use their collective knowledge and skills to solve the
other group members your friends?’’ Although this problem better than they could individually.
was a university student sample, most participants had
considerable work experience. Eighty-two percent of Technology
the sample was currently employed at the time of the
study, working an average of 16 hours per week in a The technology-mediated conditions (videoconfer-
variety of industries. Participants reported working an ence and chat) of this study took place in a computer
average of five different jobs over the course of their laboratory with six-foot dividers simulating an office
lives, each of which they held for at least one month. environment (i.e., cubicles). The hardware and software
Upon signing up for the study, participants were used for this study was carefully researched and chosen
scheduled to be on a team comprised of between three based on the following criteria: (a) currently used or
and four participants and a team leader (what team they similar to technology used by organizations (externally
were assigned to depended on the time slot for which valid), (b) user-friendly (intuitive to use with minimal
they signed up, which varied between FTF, videoconfer- training and support), (c) of high quality, and
ence, and chat). (d) reliable.
Task Videoconferencing
The hardware and software used to conduct the vid-
A structured problem solving exercise, the Meeting eoconference sessions was Polycom ViaVideo, which
Effectiveness Situation (MES) was used to measure is a state-of-the art desktop videoconference system
task performance in this study (Human Synergistics, allowing for clear picture and voice transmission. High
2002). This exercise has been used for team development quality, noise cancelling headphones were provided to
in corporate settings, and appears to be higher in com- each participant for audio transmission, and to block
plexity when compared to frequently-used group ‘‘sur- out noise distraction (further simulating a virtual envi-
vival’’ tasks. The MES has 20 items (survival tasks ronment). Rather than only seeing the other team mem-
typically have 10) that are relatively detailed, particularly bers through voice activation (e.g., when a participant
speaks the person’s image is projected), ViaVideo media conditions. A description of how each style was
allowed each participant to continually see the head carried out is provided as follows:
and shoulders of the speaker as well as the two or three
other team members (but the option to see oneself was Transactional
turned off). A screen shot capturing what a participant’s Leaders displaying a transactional style set goals for
screen looked like can be found in Fig. 1. the task and highlighted and linked extrinsic rewards to
successful task accomplishment. The transactional leader
Chat typed (in the text-based condition) or articulated (in the
The software used for the chat teams was the chat videoconference or FTF condition) comments based on
component of BlackBoard. This is a basic chat pro- the pre-determined script. These comments: (1) empha-
gram, which associates each participant’s name with sized expectations (what team members needed to do),
his/her statements (which is very similar to chat technol- (2) emphasized the team’s goals for the exercise (arriving
ogy used by organizations). All participants could see at the best possible solution), (3) emphasized the poten-
the comments of all the other participants, each being tial reward (in exchange for achieving an effective order-
associated with the individual’s first name. Participants ing of the 20 items), (4) displayed satisfaction as the team
in the chat condition wore the same noise-blocking progressed through and completed the MES, and (5)
headsets as used for videoconferencing, so as to block were professional, but not overly enthusiastic or upbeat.
out distraction and simulate a virtual team work Example comments made by the transactional leader are:
environment. ‘‘Remember, we could be one of the top teams and win
the reward by coming up with the best ordering of these
Team leaders activities’’ and ‘‘I think we are well on our way to creating
a great team answer, so let’s keep working towards our
Two male, senior-level undergraduate students were goal of coming up with the best possible ranking’’ (both
hired and trained to carry out the leadership manipula- reflect the contingent reward factor).
tion. They were blind to the study’s hypotheses. Leaders
of the same gender (males in this case) were used to con- Transformational
trol for potential gender effects. The leaders were ran- Leaders displaying a transformational style repeatedly
domly assigned to each group, and displayed either a emphasized the importance of working together, and
transactional or transformational leadership style, based linking team synergy (i.e., the team decision is greater
on a pre-determined script, which was a tailored and than the sum of individual decisions) to successful collec-
expanded version of that created by Kahai and Avolio tive outcomes. The transformational leader augmented
(2006). The script was modified through extensive prac- the positive goal-setting effects of the transactional leader
tice and pilot testing so as to sound realistic in all three by projecting confidence in team members, emphasizing
Fig. 1. Screen capture of a team desktop videoconference as used in this study.

task interdependence, the importance of collective action, computer terminals, each separated by six-foot dividers
and the opportunity to learn from the other team mem- so they could not see or be seen by the leader or other
bers (Sosik et al., 1997). team members in the lab. After being seated, they were
Specifically, the transformational leader typed or given a brief overview of the study and provided with
articulated comments based on the pre-determined two copies of the informed consent form to read and
script. These comments: (1) emphasized the importance sign. They were also told that each member of the teams
of the collective task and its relevance to their university performing in the top 10% in each condition (6 teams in
and work meetings, (2) expressed confidence that the total) would receive a $20 bookstore gift certificate.
team would complete the exercise with the best possible These were awarded once all sessions were completed.
solution, (3) highlighted the importance of working When all participants had signed the informed consent
together to determine a team solution better than any forms, each was given the MES booklet, along with ver-
one person could do alone, (4) encouraged an under- bal directions and timing instructions (20 min to com-
standing and appreciation of different ideas within the plete the individual exercise).
team, (5) indicated what team members could learn from Once participants finished the individual exercise, the
the task, (6) stressed the importance of questioning each researcher or RA assisted them with logging on to their
other’s assumptions, ideas and viewpoints, (7) made team session. They were told that they would be work-
each participant feel that they were an important part ing on the same task they just completed, but with a
of the team and that their opinion mattered, and (8) team of two or three other students as well as an
was upbeat and enthusiastic. Examples of comments assigned, trained leader. Further, they were told that
made by the transformational leader are: ‘‘Let’s contin- the team session would involve discussing the exercise
ue being as open as we can in questioning each other’s and achieving the best possible consensus ranking of
ideas and assumptions about these activities’’ (reflects the items (a ranking with which all team members could
the intellectual stimulation factor); and ‘‘You have come agree). They were also instructed about the time limit
up with a lot of good ideas so far by working together as for the team task (35 min for videoconference teams;
a team (reflects the inspirational motivation factor). 50 min for chat teams), plus a 10-min introduction ses-
sion (note that based on pilot testing chat teams were
Pilot study given 15 more minutes than FTF and videoconference
teams to allow for the greater time demands of typing).
A pilot study was conducted with 30 undergraduate/ The leader began by facilitating approximately
graduate students from the Department of Psychology 10 min of introductions prior to facilitating the MES.
acting as participants in the six different experimental The primary purpose of this non-task related communi-
conditions. The main purposes of the pilot study were cation was to allow time for team members to adapt to
to: (1) test all experimental procedures, (2) ensure that the communication medium. This time to familiarize
the two trained leaders were able to successfully facili- oneself with the communication technology is impor-
tate a team through these technologies, and (3) test the tant, and studies not doing so have been criticized
required timing to complete the task individually and (Wakertin et al., 1997). Researchers have found that
as a team in each condition. Procedures were modified typically 10 min is enough time to become familiar with
as necessary based on results of the pilot study. the operation of a desktop videoconference system
(Townsend et al., 2001).
Procedure Following the introductions, the leader facilitated
Videoconferencing and chat conditions. Teams in these completion of the MES. Their leadership script began
two conditions were considered virtual teams given that with an introductory message that was either transac-
their interactions were conducted solely through com- tional or transformational. As the group worked at
munication technology, and they were not introduced achieving consensus, the leader interjected comments
FTF prior to interacting virtually. Every effort was based on the script. The recommended timing of these
made to ensure that participants did not interact with comments was included in the script, but the leader used
each other FTF at anytime throughout the study. his training, experience, and discretion to provide each
Upon arriving at the computer laboratory, partici- comment at an appropriate time in the conversation to
pants were greeted by the researcher or a research assis- maintain realism. The leader checked each comment as
tant (RA)1, and immediately directed to one of the four he made it, ensuring that all comments were made in
each team session. In addition, the leader was unaware
1
of the experts’ solution, and did not provide his opinion
Due to the requirements to keep participants separated, three RAs on the ordering of the 20 items; instead, he played a
were required for each laboratory session (one of which sometimes
included the researcher). Each RA was responsible for one or two of
team facilitation role.
the participants, so as to ensure equal timing of exercises and efficient Prior to completing the team session, the leader
completion of the session within the 2-h time limit. informed the team members that the final component
of the study involved completing on-line questionnaires. Team interaction styles. A web-based version of the
These took 15–20 min to finish, and were completed pri- Group Styles Inventory (GSI; Balthazard, 1999a;
or to participants being debriefed. Cooke & Szumal, 1994) was used to measure each
Face-to-face condition. The procedure for the FTF team’s dominant interaction style. The GSI is a self-re-
experimental condition was consistent with the proce- port survey consisting of 72 statements assessing con-
dures and timing used in the two virtual team conditions structive and defensive team interaction styles. Items
(except for chat, which was 15 min longer). Participants focused on the ways in which team members interacted
signed up for the study according to a set of prearranged with each other and approached the task during the
time slots. Each of the twenty teams in the FTF condi- problem-solving session. Example items include: ‘‘was
tion was composed of a team leader and 3–4 partici- communication supportive and constructive?’’ (assesses
pants. The teams met in a standard meeting room on constructive interaction), ‘‘did tactfulness inhibit direct
the university campus, and were seated around a table communication and the questioning of ideas?’’ (assesses
upon entering the room. They were instructed not to passive interaction), and ‘‘were you interrupted by oth-
speak to one another until the leader began the team ses- ers trying to sell their ideas?’’ (assesses aggressive inter-
sion. Once the team session began, the leader followed action). Participants indicated the extent to which each
the same procedure and script as used in the virtual con- item described the interaction style of their team using
ditions, with some flexibility to ensure appropriate tim- a five-point Likert scale ranging from 0 (not at all) to
ing and realism of comments. Once the team MES was 4 (to a very great extent). Internal consistency for these
completed, team members were brought to separate scales ranged from .88 to .94, indicating that partici-
computer terminals to complete the on-line surveys, pants answered the items in a consistent manner (see
then debriefed. Table 2). After checking for acceptable agreement levels
between participants on each team, responses to the GSI
Measures were averaged. Then, an average constructive, defensive,
The following section describes each of the measures passive and aggressive interaction score was calculated
used in this study. Some of the measures (the team inter- for each team.
action styles and team cohesion) were assessed at the Team cohesion. Team cohesion was measured by par-
individual level and then aggregated and analyzed at ticipants’ ratings on nine items that assessed group
the team level. Thus, a description of the interrater atmosphere and the extent to which team members
agreement calculations used to determine the appropri- stuck together (Balthazard et al., 2002; Cook, 1981;
ateness of collapsing individual to team level data is first O’Reilly, Caldwell, & Barnett, 1989). Example items
described. include: ‘‘members appeared to feel that they were really
Interrater agreement. The interrater agreement levels part of the group’’ and ‘‘members of the group really
(rwg) for each scale were assessed using the calculations stuck together.’’ Responses were rated on a five-point
proposed by James, Demaree, and Wolf (1984, 1993). scale ranging from 1 (strongly disagree) to 5 (strongly
This procedure is conducted to ensure that averaging agree). After checking for an acceptable agreement level
ratings across team members into a single aggregate between team members, the items were summed and
score is appropriate. Stated simply, this procedure averaged for each team. The Cronbach’s a coefficient
assessed whether or not team cohesion and team interac- for these nine items was .82, calculated at the individual
tion styles are appropriately analyzed at the team level. level.
The mean levels of rwg for the scales were quite high Task performance. Objective task performance on the
using the James et al. formula (all above .90). One team MES was assessed by calculating the difference between
had a mean rwg value, calculated across all five scales, the team consensus and the average of the team mem-
lower than .70. Given that particular team’s low overall ber’s individual performance scores on the MES, com-
interrater agreement, the analyses were re-run without monly referred to as the ‘‘conventional scoring
their data. None of the results were affected, however, algorithm,’’ which is the most commonly used in group
and the analyses reported are based on those including research (Balthazard et al., 2002). Thus, performance in
all of the teams. this study is determined relative to each team’s individ-
Questionnaires. The on-line questionnaires consisted ual capability, rather than ignoring individual scores
of: (a) demographic questions, (b) the Group Styles and solely considering the team’s final performance
Inventory (GSI; Cooke & Szumal, 1994), (c) a measure score.
of team cohesion, and (d) manipulation check items.
Each of these measures is described in the following Manipulation checks
section. Leadership style. To determine whether participants
Demographic questions. Thirteen questions asked correctly perceived the intended leadership style, a
about participant characteristics, such as gender, age, post-experimental manipulation check was conducted.
year of studies, and work experience. Participants were given nine questions adapted from
the Multifacet Leadership Questionnaire (MLQ; respectively), p = .025; and videoconference was per-
Kahai & Avolio, 2006), about the leadership style exhib- ceived to be lower in richness than FTF (X ¼ 5:92),
ited by their leader to determine if their perceptions of it p < .001. These results suggest that participants did
differed across the transactional and transformational perceive differences in the richness of these three
conditions. The first five questions assessed transforma- communication media.
tional leadership behaviours (e.g., ‘‘the leader empha-
sized the importance of working together as a team’’), Data analyses and results
while the last four assessed transactional leadership
behaviours (e.g., ‘‘the leader indicated that we could Means and standard deviations at the team level for
potentially win a reward (gift certificate) for generating the team interaction styles (constructive, defensive, pas-
a winning team solution’’). The internal consistency of sive, aggressive), team cohesion and performance are
the five transformational items was .85, and for the four shown in Table 1. Correlations among these variables
transactional items was .70. are depicted in Table 2.
Media richness. Eight questions were administered to To assess Hypothesis 1, which predicted that transac-
assess participants’ perceptions of the richness of the tional leadership would more strongly predict a con-
communication medium (Suh, 1999). Example items structive interaction style than would transformational
include ‘‘the environment in which we communicated leadership in richer media, but that this effect would
helped us to better understand each other’’ and ‘‘I could decrease as the medium became leaner, a 2 · 3 factorial
easily explain things in this environment.’’ These items ANOVA was conducted. A significant interaction was
were measured on a seven-point Likert scale format, not found between communication medium and leader-
ranging from 1 (strongly disagree) to 7 (strongly agree). ship style on constructive team interaction,
The Cronbach’s a was .93 for these items. F(2, 54) = .098, ns.
Hypothesis 2a postulated that transactional leader-
ship would more strongly predict team performance
Results than would transformational leadership in richer media
(i.e., FTF), but this effect would decrease as the medium
Manipulation checks became leaner. This hypothesis was tested using a 2 · 3
factorial ANOVA, however, the interaction term was
Leadership style
To determine whether participants correctly per-
ceived the intended leadership style, responses to the Table 1
Means and standard deviations of team scores
nine leadership manipulation check items were analyzed
using independent t-tests. These revealed that the score Scale Mean Standard deviation
on the five transformational manipulation check items Constructive 3.38 .29
was higher (n = 114, X ¼ 4:34) for participants led by Defensive 2.26 .22
Passive 2.34 .21
a transformational leader than those led by a transac-
Aggressive 1.65 .25
tional leader (n = 114, X ¼ 3:94), t (226) = 4.20, Cohesion 3.92 .39
p < .001. Similarly, the score on the four transactional Performance 40.59% 20.98%
manipulation check items was higher (n = 114, Note: the range for the GSI scales (constructive, defensive, passive, and
X ¼ 4:48) for participants led by a transactional leader aggressive) and cohesion is 1–5.
than those led by a transformational leader (n = 114, ‘‘Performance’’ is the average percent change scores across all 60
X ¼ 3:61), t (226) = 9.25, p < .001. Thus, the leader- teams (it represents the difference between the average individual score
and the team score).
ship manipulation appeared to be experienced by
participants.
Media richness Table 2

Scale correlations and reliability coefficients
To determine whether participants perceived differ-
ences in the richness between the three communica- Scale 1 2 3 4 5 6 7
tion media (FTF, videoconference, and chat), scores Teams (n = 60)
on these eight questions were compared using a Constructive a = .93
Defensive .472** a = .94
one-way ANOVA. Results indicated that there was
Passive .425** .929** a = .88
a significant effect of communication medium on per- Aggressive .460** .949** .764** a = .91
ceived media richness F(2, 225) = 98.98, p < .001. As Cohesion .813** .650** .599** .620** a = .82
expected, the follow-up tests of mean differences Performance .184 .026 .103 .043 .143 —
showed that chat was perceived to be lower in media *
p < .05.
richness than videoconference (X ¼ 3:48, X ¼ 3:96, **
p < .01.
not significant, F(2, 54) = .33, ns. Similarly, Hypothesis one-way ANOVA. Results revealed that richer commu-
2b argued that transactional leadership would more nication media resulted in higher scores on team cohesion
strongly predict team cohesion compared to transforma- than did less rich media, F(2, 57) = 11.74, p < .001,
tional leadership in rich media, but this effect would g2 = .29 (see Table 3). Follow-up tests conducted at a
decrease as the medium became leaner. Once again, family-wise Bonferroni corrected a level of .017 indicat-
the 2 · 3 factorial ANOVA was not significant, ed that the mean team cohesion score was higher in FTF
F(2, 54) = .37, ns. (X ¼ 4:15) than text-based (X ¼ 3:64) teams, p < .001.
To assess Hypothesis 3, which was that richer com- Also, the mean team cohesion score was higher in video-
munication media would result in higher constructive conference (X ¼ 3:97) than text-based (X ¼ 3:64) teams,
team interaction scores than less rich communication p = .009. The mean team cohesion score was not signif-
media, a one-way ANOVA was conducted. A statistically icantly higher in FTF (X ¼ 4:15) compared to videocon-
significant effect was found for communication medium, ference (X ¼ 3:97) teams.
F(2, 57) = 11.22, p < .001, g2 = .28 (see Table 3). Follow- Hypothesis 6, which predicted that teams interacting
up tests of mean differences, conducted at a family-wise through richer communication media would display
Bonferroni corrected a level of .017, indicated that greater task performance than teams interacting
the mean constructive interaction score was higher in through less rich communication media, was also tested
FTF (X ¼ 15:41) than videoconference (X ¼ 14:20) using a one-way ANOVA. Richer communication
teams, p = .041. Also, the mean constructive interaction media did not result in higher scores on task perfor-
score was higher in FTF (X ¼ 15:41) than chat mance than did less rich media (F(2, 57) = 2.56, ns).
(X ¼ 13:15) teams, p < .001. Mean constructive interac-
tion scores were not significantly higher in videoconfer-
ence (X ¼ 14:20) compared to chat (X ¼ 13:15) teams. Discussion
Thus, Hypotheses 3a and 3b were supported, but not
Hypothesis 3c. Hypothesis 1 sought to determine whether type of
To assess Hypothesis 4, which was that richer com- communication medium would interact with leadership
munication media would result in lower defensive team style to affect constructive team interaction. A signifi-
interaction scores than would less rich media, a one-way cant interaction was not found, suggesting that leader-
ANOVA was conducted. A significant effect was not ship style did not affect constructive team interactions
found, thereby not supporting Hypotheses 4a–c differently depending on the communication medium.
(F(2, 57) = 2.27, ns). Furthermore, a post-hoc analysis investigating the main
Additional ANOVAs were conducted on the effects effects of leadership style on constructive team interac-
of communication media on the specific components tion was not significant. Despite the greater number of
of defensive team interaction: passive and aggressive. verbal and nonverbal cues available through richer
No significant effect of communication media was found media such as FTF, the effects of the leaders’ styles on
on passive team interaction (F(2, 57) = 1.54, ns). The constructive team interactions were not any different
effect of communication media on aggressive team inter- than through less rich media (e.g., chat). This finding
action also failed to achieve significance F(2, 57) = 2.84, may be the result of the short-term nature of this task,
p = .067, g2 = .09 (see Table 3). which regardless of media type, may not have allowed
Hypothesis 5, which predicted that teams interacting sufficient time for leadership style to influence team
through richer communication media would score interaction. In addition, the scripted nature of the lead-
higher on team cohesion than teams interacting through ership styles would have enabled both transformational
less rich communication media, was tested using a and transactional leaders to clearly exert their influence
in all three media (because an equal number of messages
were transmitted irrespective of media type). In actual
Table 3 virtual teams, however, leaders may not be able to com-
Analysis of variance for communication media and constructive municate as many messages through less rich media in
interactions, aggressive interactions, and cohesion
order to influence team members. Future field research
Source SS df F g2 p is needed that examines the extent to which leaders influ-
Between subjects ence teams to a different degree based on the communi-
Constructive 51.077 2 11.22** .28 .000 cation medium’s richness.
Within group error 129.761 57 (2.28)
Aggressive 7.35 2 2.84 .09 .067
Hypothesis 2a sought to determine whether leader-
Within group error 92.133 57 (1.66) ship style would interact with communication medium
Cohesion (C) 2.675 2 11.74** .29 .000 to affect team performance. A significant interaction
Within group error 6.491 57 (0.114) was not found. Furthermore, a post-hoc analysis inves-
Note: values enclosed in parentheses represent mean square errors. tigating the main effects of leadership style on team per-
**
p < .001. formance was not significant. Similar findings were
revealed for Research Question 2b. That is, the interac- be an improvement over chat in helping team members
tion between leadership style and communication media interact more constructively, at least for this type of
on team cohesion was not significant, nor was an ad-hoc short-term, problem solving task. Perhaps the additional
analysis on the main effects of leadership style on team costs of videoconference may not be necessary for this
cohesion. Taken together, these findings indicate that type of virtual teamwork. The greater efficiency of com-
leadership style did not predict team outcomes in this pleting problem solving tasks through videoconference
study, even when the communication medium was con- as compared to chat, however, may be another consider-
sidered. Although previous research has found that ation. Teams using chat in this study took approximately
leadership styles predict team outcomes (Bass et al., 13 more minutes, on average, to complete this task than
2003; Hoyt & Blascovich, 2003; Sosik et al., 1997), both teams communicating through videoconference. The
styles have been found to relate more positively than the costs of desktop videoconference, therefore, could possi-
other under various conditions (Howell & Avolio, 1993; bly be justified based on its greater efficiency.
Howell & Frost, 1989; Kahai et al., 2003; Sosik et al., It is important to note that participants had much
1997), and may have been equally effective in the present more skill and experience with chat than videoconfer-
study. ence. Only 15% of the participants had previously par-
Had we included a condition without a designated ticipated in a videoconference, whereas 62% indicated
leader, or a condition with a leader trained to display having used chat (e.g., chat rooms). Perhaps more expo-
laissez-faire behaviours, we could have confirmed sure to videoconference might increase their ability to
whether or not the two leadership styles had an affect utilize this medium more effectively, thereby enabling
on team outcomes. As it stands, we cannot be sure if more constructive interaction than through chat.
the leadership manipulation affected team outcomes Hypothesis 4, which stated that richer communica-
compared to a ‘‘no leadership’’ condition, though we tion media would result in lower defensive team interac-
do know that transactional and transformational leader- tion scores than less rich media, was not supported.
ship styles did not have differential effects on the depen- Although counter to expectations, this non-significant
dent variables. Comparing these leadership styles to result corroborates the finding of Balthazard et al.
laissez-faire leadership is a potential avenue for future (2002) who found that FTF teams were not less likely
research. to demonstrate defensive interaction styles than virtual
The results of Hypothesis 3 partially support the teams. The present study’s finding suggests that teams
prediction that richer communication media result in do not interact more defensively through less rich media
higher constructive team interaction scores than less and also that either videoconference or chat may be via-
rich communication media. Specifically, the mean con- ble alternatives to FTF team communication. As with
structive team interaction score was higher in FTF Hypothesis 3, this finding does not support videoconfer-
compared to chat teams. This supports the work of ence as being more effective than chat.
Balthazard et al. (2002), who found that FTF teams Hypothesis 5, which predicted that teams interacting
demonstrated higher levels of constructive interaction through richer communication media would score high-
than those communicating through chat. Results of er on team cohesion than teams interacting through less
the current study also found that the mean constructive rich communication media, was partially supported.
team interaction score was higher in FTF than video- Specifically, the mean team cohesion score was higher
conference teams. Contrary to the postulates of media in both FTF and videoconference than chat teams, but
richness theory (Daft & Lengel, 1986), however, the not significantly different between FTF and videoconfer-
mean constructive team interaction score was not ence teams. This finding suggests that desktop videocon-
higher in videoconference than in chat teams. These ference can lead to teams that are equally cohesive as
findings suggest that in order to maximize teams’ con- those communicating FTF. As a result, desktop video-
structive interactions, FTF communication is the most conference could be a viable alternative to producing
effective alternative. Given the realities of today’s orga- cohesive teams when FTF meetings are not feasible.
nizations, however, FTF teamwork is not always finan- This result also supports the earlier findings of Straus
cially or logistically feasible. For example, the costs of and McGrath (1994) and Wakertin et al. (1997) in that
flying people to meet FTF may be prohibitive for many FTF teams are more cohesive than virtual teams com-
organizations, making virtual teamwork a necessity municating through synchronous and asynchronous
(Baltes et al., 2002). Also, in order to acquire the text-based media, respectively. Moreover, it corrobo-
required talent, hiring workers from various locations rates Baltes et al. (2002) meta-analysis that found that
who are unable or unwilling to relocate may be neces- computer-mediated (text-based) teams performing inter-
sary (Bell & Kozlowski, 2002). dependent tasks were usually less satisfied than their
The lack of difference found between constructive FTF counterparts. This lower satisfaction may be due,
team interaction scores of videoconference and chat in part, to frustration with the fact that typing takes
teams suggests that desktop videoconference may not much longer than verbal communication (Wiggins &
Horn, 2005). Thus, one cannot assume FTF teamwork appropriate communication media, and how these
leads to more cohesive teams than virtual teamwork. choices will ultimately affect the performance and satis-
Instead, it may depend on the specific type of communi- faction of team members. The findings from the current
cation media used by the virtual team. laboratory study offer some insights into these types of
Hypothesis 6, predicting that teams interacting important decisions.
through richer communication media would display Although this study did not find different effects
greater task performance than teams interacting between transformational and transactional leadership
through less rich communication media, was not sup- styles, the importance of team leadership cannot be
ported. Specifically, richer communication media did overlooked. Other studies have demonstrated significant
not result in higher scores on task performance than effects of virtual team leadership (e.g., Davis et al., 2003;
did less rich media. This result supports virtual team- Kahai & Avolio, 2006; Kahai et al., 2003; Sosik et al.,
work as being a viable alternative to FTF teamwork, 1997). Further research, especially field research, is
at least for these types of short-term, problem solving needed to continue exploring these virtual leadership
tasks. The question remains, however, as to whether styles and the moderators and mediators that increase
actual, longer-term virtual teams can also develop equal or decrease their effects.
levels of task performance through communicating This study provided some insights into the effects of
virtually versus FTF. three different communication media. FTF teams dem-
onstrated more constructive interaction than did video-
Theoretical and practical implications conference or chat teams, yet the defensive interaction
between teams in all three media did not differ. Further,
This study built upon existing theory on virtual team FTF and videoconference teams demonstrated higher
leadership, which compared to FTF leadership theory, is cohesion than did chat teams. Teams in all three media
still in its infancy (Zaccaro et al., 2004). The hypothe- seemed to demonstrate equal task performance. These
sized relationships tested in the laboratory study built findings are mixed, but point to the fact that not all
upon the work of both virtual team researchers (e.g., of these media are equally effective. In some regards,
Avolio et al., 2001b; Balthazard et al., 2002) and media virtual teams face greater challenges than FTF teams
richness theorists (Daft & Lengel, 1986; Dennis & (Thompson & Coovert, 2006). As argued by Balthazard
Valacich, 1999). The major contribution of this study et al. (2002), compared to FTF teams, virtual teams face
is its examination of leadership and teamwork in three cohesion and interaction challenges, especially in the
different media, including desktop videoconference, early stages of team development. Thus, it is important
which has not, to the researchers’ knowledge, been from the inception of a virtual team to ensure it develops
studied in this context. Specifically, leadership and team- effective leadership, interaction styles, cohesion, and
work have not been studied together in a study compar- ultimately performance.
ing desktop videoconference to the more commonly
studied chat and FTF media. This study responds to Limitations
calls from other researchers for more empirical research
about the effects of videoconference on leadership and One limitation of this study is that it was conducted
teamwork (e.g., Bell & Kozlowski, 2002; Townsend using zero-history teams completing a contrived, short
et al., 2001). Such research will contribute to the knowl- duration task in a laboratory setting. Also, unlike many
edge base about leading through a broader range of real virtual teams, participants in this study did not
communication media. work for the same organization, and would not be work-
Another theoretical implication of this study is that ing together in the future. Many real-world virtual
it builds upon other researchers’ work in applying the teams, however, are ad hoc teams that interact on
transformational/transactional leadership paradigm short-term projects (Balthazard et al., 2002), indicating
to the study of virtual teams (Avolio et al., 2001b; that these laboratory results may generalize to some
Kahai & Avolio, 2006; Kahai et al., 2003). Although extent to real virtual teams. In particular, they may be
much more empirical research is needed to understand generalizable to virtual teams that form, function, and
how these leadership styles play out in a variety of disband fairly quickly.
virtual teams, this study makes an important contribu- It is also necessary to emphasize that leadership styles
tion by examining these styles in a controlled, labora- were simulated, so they may not have appeared as
tory setting. believable or realistic as would be the case in a real team
In addition to building upon virtual leadership or virtual team. Further, the differences in the two lead-
theory, this study provides some practical implications ership styles may not have been clearly experienced by
for organizations currently using or considering imple- participants. Even though participants heard the com-
menting virtual teams. Organizations will need to make ments, as evidenced by the significant leadership manip-
decisions about the leadership of these teams, the most ulation check results, they may not have taken them
seriously or applied them to the task. Although the lead- behaviours are most effective in actual teams. Also,
ers were given some flexibility in the timing of their com- future research could examine whether these relation-
ments, and the ability to interject additional generic ships and effect sizes are similar when team members
comments, they still may not always have been viewed have more familiarity and expertise with the task at
as realistic. They also did not have power over their hand.
teams such as is often the case in real organizations More fully examining applications of the transfor-
(e.g., leader conducting followers’ performance reviews). mational/transactional leadership paradigm to virtual
Another limitation related to the leadership manipu- teams is another area for future research. For example,
lation is that the two types of leadership, transforma- the laissez-faire style of leadership, a style that is essen-
tional and transactional, may not represent a true tially representative of a lack of leadership, was not
dichotomy. Indeed, some researchers have found that examined as a condition in this study. This style can
these styles are highly correlated and that they are both be contrasted with shared leadership, which other stud-
associated with positive performance (Judge & Piccolo, ies (e.g., Balthazard et al., 2002) have examined.
2004). Although other studies have used this dichotomy, Shared leadership, unlike laissez-faire, does not involve
it may not adequately reflect real-world leaders who having a leader, but rather leadership is shared or
may simultaneously display aspects of both styles. rotated amongst team members. As we noted earlier,
The findings from this study are also limited because if we had investigated laissez-faire or shared leadership
the sample was comprised of undergraduate university we would have been able to confirm if transformational
students instead of actual employees in organizations. and transactional did indeed affect team outcomes sim-
As noted previously, however, 82% of this sample ilarly. With the current design, we cannot confirm
was currently employed at the time of the study, work- whether transformational and transactional leadership
ing an average of 16 h per week in a variety of indus- styles had either a similar effect on team outcomes or
tries. These students, therefore, had considerable no effect. Thus, how these different leadership styles,
workplace experience and provide a reasonable sample including laissez-faire and shared leadership, effect vir-
for study. tual team outcomes will be an important contribution
The composition of this sample represents another of future research.
potential limitation of this study. With the majority of More research is also needed to determine which
participants being female (87%), many of the teams were types of interventions (e.g., selecting, training, coaching,
composed mostly or all of females, while both leaders team building) are needed to enhance the leadership,
were male. Thus, the generalizability of these findings constructive team interaction, team cohesion, and per-
to teams comprised mostly or all of males, or teams formance of virtual teams. Although practical recom-
led by females, is questionable. Statistical gender com- mendations abound in the popular literature, more
parisons indicated, however, that there were no signifi- empirical laboratory and field research is needed which
cant differences between genders on any of the incorporates new technologies, such as videoconference,
outcome variables. Further research examining the that may be increasingly used by virtual teams.
effects of gender composition in virtual teams is needed
to better understand potential differences.
The power of this study represents an additional lim- Conclusion
itation, as it was low to moderate. Despite this, some
effects were found, which are likely to be even stronger This study suggests that both transformational and
in the population. Future research should strive to transactional leadership styles are equally effective
obtain a higher power level so that relationships can across communication media in teams completing
more readily be detected. This may mean comparing short-term, problem solving tasks. Furthermore, this
fewer conditions than was the case in this study. research provides additional confirmation that the com-
munication media through which teams communicate
influence certain aspects of their interactions and cohe-
Future research siveness. Together these findings point to the importance
of virtual leaders establishing media through which vir-
Future research is needed to determine whether the tual teams can most effectively communicate and collab-
findings from this study apply to field settings through orate, thereby increasing their constructive interactions
examining virtual teams working on actual problem and cohesion, which in turn may ultimately impact their
solving tasks and projects. Such studies would help cap- performance. There are still more questions than
ture the ‘‘motivational element’’ (Balthazard et al., 2002) answers about virtual team leadership, but given the
that may not have been as strong in the current study. continued growth of virtual teams across industries, this
Furthermore, by measuring the styles of real leaders, a exciting area of research is ripe for future research
determination could be made about what styles and endeavours.
Acknowledgments Cascio, W. F., & Shurygailo, S. (2003). E-leadership and virtual teams.
Organizational Dynamics, 31(4), 362–376.
Chansler, P. A., Swamidass, P. M., & Cammann, C. (2003). Self-
This research was funded by SSHRC (Social Sciences managing work teams: an empirical study of group cohesiveness in
and Humanities Research Council of Canada). We also ‘‘natural work groups’’ at a Harley-Davidson motor company
thank Human Synergistics for use of the online Group plant. Small Group Research, 34, 101–120.
Styles Inventory and the Meeting Effectiveness Cook, J. D. (1981). The experience of work: A compendium and review
Situation. of 249 measures and their use. New York: Academic Press.
Cooke, R. A., & Szumal, J. L. (1994). The impact of group
interactional styles on problem-solving effectiveness. Journal of
Applied Behavioral Science, 30(4), 415–437.
References Daft, R. L., & Lengel, R. H. (1986). Organizational information
requirements, media richness and structural design. Management
Avolio, B. J., Kahai, S., & Dodge, G. E. (2001b). E-leadership: Science, 32(5), 554–571.
implications for theory, research, and practice. Leadership Quar- Davis, D. D., Mihalescz, M., Bryant, J. L., Tedrow, L., Liu, Y., & Say,
terly, 11(4), 615–668. R. (2003). Leadership in global virtual teams. Paper presented at the
Avolio, B. J., Kahai, S., Dumdum, R., & Sivasubramaniam, N. meeting of the Society for Industrial and Organizational Psychol-
(2001a). Virtual teams: implications for e-leadership and team ogy, Orlando, Florida.
development. In M. London (Ed.), How People Evaluate Others in Den Hartog, D. N., & Koopman, P. L. (2001). Leadership in
Organizations (pp. 337–358). Mahwah, NJ: Lawrence Erlbaum. organizations. In N. Anderson, D. S. Ones, H. K. Sinangil, & C.
Baker, G. (2002). The effects of synchronous collaborative technolo- Viswesvaran (Eds.). Handbook of Industrial, Work and Organiza-
gies on decision making: a study of virtual teams. Information tional Psychology (Vol. 2, pp. 166–187). London, England: Sage.
Resources Management Journal, 15(4), 79–93. Dennis, A. R., & Valacich, J. S. (1999). Rethinking media richness:
Baltes, B. B., Dickson, M. W., Sherman, M. P., Bauer, C. C., & Towards a theory of media synchronicity. In Proceedings of the
LaGanke, J. S. (2002). Computer-mediated communication and 32nd Hawaii International Conference on System Sciences, 1, pp. 1–
group decision making: a meta-analysis. Organizational Behavior 10.
and Human Decision Processes, 87(1), 156–179. Dionne, S. D., Yammarino, F. J., Atwater, L. E., & Spangler, W. D.
Balthazard, P. A. (1999a). Group styles inventory—internet edition. (2004). Transformational leadership and team performance. Jour-
Arlington Heights, IL: Human Synergistics/Center for Applied nal of Organizational Change Management, 17, 177–193.
Research. Driskell, J. E., Radtke, P. H., & Salas, E. (2003). Virtual teams: effects
Balthazard, P. A. (1999b). In J. C. Lafferty & A. W. Pond (Eds.), of technology mediation on team performance. Group Dynamics:
Virtual version of the desert survival situation. Arlington Heights, Theory, Research, and Practice, 4, 297–323.
IL: Human Synergistics/Center for Applied Research. Dvir, T., Eden, D., Avolio, B. J., & Shamir, B. (2002). Impact of
Balthazard, P., Waldman, D. A., Howell, J., & Atwater, L. E. (2002). transformational leadership on follower development and perfor-
Modeling performance in teams: the effects of media type, shared mance: a field experiment. Academy of Management Journal, 45(4),
leadership, interaction style, and cohesion. Paper presented at the 735–744.
meeting of the Academy of Management, Denver, Colorado. Fletcher, T. D., & Major, D. A. (2003). The effects of communication
Barling, J., Weber, T., & Kelloway, K. E. (1996). Effects of modality on teamwork processes. Paper presented at the meeting of
transformational leadership training on attitudinal and financial the Society for Industrial and Organizational Psychology, Orlando,
outcomes: a field experiment. Journal of Applied Psychology, 81(6), Florida.
827–832. Griffith, T. L., & Neale, M. A. (2001). Information processing in
Bass, B. M. (1985). Leadership and performance beyond expectations. traditional, hybrid and virtual teams: from nascent knowledge to
New York: Free Press. transactive memory. In B. M. Staw & R. I. Sutton (Eds.). Research
Bass, B. M. (1997). Does the transactional—transformational leader- in organizational behaviour (Vol. 23). New York: Elesevier.
ship paradigm transcend organizational and national boundaries? Gully, S. M., Devine, D. J., & Whitney, D. J. (1995). A meta-analysis
American Psychologist, 52, 130–139. of cohesion and performance. Small Group Research, 26, 497–525.
Bass, B. M., & Avolio, B. A. (1990). The multifactor leadership Hambley, L. A., O’Neill, T. A, & Kline, T. J. B. (in press). Virtual team
questionnaire. Palo Alto, CA: Consulting Psychologists Press. leadership: perspectives from the field. International Journal of e-
Bass, B. M., & Avolio, B. J. (1993). Transformational leadership: a Collaboration.
response to critiques. In Leadership Theory and Research: Perspec- Hertel, G., Geister, S., & Konradt, U. (2005). Managing virtual teams:
tives and Directions (pp. 49–79). Academic Press, Inc. a review of current empirical research. Human Resource Manage-
Bass, B. M., & Avolio, B. J. (1994). Improving organizational ment Review, 15, 69–95.
effectiveness through transformational leadership. Thousand Oaks, Hertel, G., Konradt, U., & Orlikowski, B. (2004). Managing distance
CA: Sage. by interdependence: goal setting, task interdependence, and team-
Bass, B. M., Avolio, B. J., Jung, D. I., & Berson, Y. (2003). Predicting based rewards in virtual teams. European Journal of Work and
unit performance by assessing transformational and transactional Organizational Psychology, 13, 1–28.
leadership. Journal of Applied Psychology, 88(2), 207–218. Hollingshead, A. B. (2004). Communication technologies, the internet,
Bell, B. S., & Kozlowski, S. W. J. (2002). A typology of virtual teams: and group research. In M. B. Brewer & M. Hewstone (Eds.),
implications for effective leadership. Group and Organization Applied Social Psychology (pp. 301–317). Blackwell Publishing:
Management, 27(1), 14–49. Malden, MA.
Bryman, A. (1992). Charisma and leadership in organizations. London: House, R. J. (1996). Path-goal theory of leadership: lessons, legacy and
Sage. a reformulated theory. Leadership Quarterly, 7(3), 323–352.
Carron, A. V., Brawley, L. R., & Widmeyer, W. N. (1998). The Howell, J. M., & Avolio, B. J. (1993). Transformational leadership,
measurement of cohesiveness in sport groups. In J. L. Duda (Ed.), transactional leadership, locus of control and support for innova-
Advances in Sport and Exercise Psychology Measurement tion: key predictors of consolidated-business-unit performance.
(pp. 213–226). Morgantown, WV: Fitness Information Technology. Journal of Applied Psychology, 78, 891–902.
Howell, J. M., & Frost, P. (1989). A laboratory study of charismatic Martins, L. L., Gilson, L. L., & Maynard, M. T. (2004). Virtual teams:
leadership. Organizational Behaviour and Human Decision Process- what do we know and where do we go from here? Journal of
es, 43, 243–269. Management, 30, 805–835.
Hoyt, C. L., & Blascovich, J. (2003). Transformational and transac- Maruping, L. M., & Agarwal, R. (2004). Managing team interpersonal
tional leadership in virtual and physical environments. Small Group processes through technology: a task-technology fit perspective.
Research, 34(6), 678–715. Journal of Applied Psychology, 89(6), 975–990.
Huang, W. W., Wei, K. K., Watson, R. T., & Tan, B. C. Y. (2002). McGrath, J. E., Hollingshead, A. B., & O’Connor, K. M. (1993).
Supporting virtual team-building with a GSS: an empirical Group task performance and communication technology: a longi-
investigation. Decision Support Systems, 34, 359–367. tudinal study of computer-mediated versus face-to-face work
Human synergistics (2002). Meeting Effectiveness Situation. Plym- groups. Small Group Research, 24(3), 307–333.
outh, MI: Human Synergistics International. Morgeson, F. P. (2005). The external leadership of self-managing
James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within- teams: intervening in the context of novel and disruptive events.
group interrater reliability with and without response bias. Journal Journal of Applied Psychology, 90(3), 497–508.
of Applied Psychology, 69, 85–98. Mullen, B., & Copper, C. (1994). The relation between group
James, L. R., Demaree, R. G., & Wolf, G. (1993). rwg: An assessment cohesiveness and performance: an integration. Psychological Bul-
of within-group interrater agreement. Journal of Applied Psychol- letin, 115, 210–227.
ogy, 78, 306–309. O’Connail, B., Whittaker, S., & Wilbur, S. (1993). Conversations over
Jarvenpaa, S. L., & Leidner, D. E. (1999). Communication and trust in videoconferences: an evaluation of the spoken aspects of video-
global virtual teams. Organizational Science, 10(6), 791–815. mediated communication. Human Computer Interaction, 8,
Judge, T. A., & Piccolo, R. F. (2004). Transformational and 389–428.
transactional leadership: a meta- analytic test of their relative O’Reilly, C., Caldwell, D., & Barnett, W. (1989). Work group
validity. Journal of Applied Psychology, 89, 755–768. demography, social integration, and turnover. Administrative
Jung, D. I. (2001). Transformational and transactional leadership and Sciences Quarterly, 34, 21–37.
their effects on creativity in groups. Creativity Research Journal, Potter, R. E., & Balthazard, P. A. (2002). Virtual team interaction
13(2), 185–195. styles: assessment and effects. International Journal of Human-
Jung, D. I., & Sosik, J. J. (2002). Transformational leadership in work Computer Studies, 56, 423–443.
groups: the role of empowerment, cohesiveness, and collective- Potter, R. E., Balthazard, P. A., & Cooke, R. A. (2000). Virtual team
efficacy on perceived group performance. Small Group Research, interaction: assessment, consequences, and management. Team
33(3), 313–336. Performance Management: An International Journal, 6(7/8),
Kahai, S., & Avolio, B. (2006). Leadership style, anonymity, and the 131–137.
discussion of an ethical issue in an electronic context. International Pulley, M. L., Sessa, V., & Malloy, M. (2002). E-leadership: a two-
Journal of e-Collaboration, 2(2), 1–26. pronged idea. Training and Development, 35–47.
Kahai, S., Sosik, J., & Avolio, B. (1997). Effects of leadership style and Salas, E., Sims, D. E., & Burke, C. S. (2005). Is there a ‘‘Big Five’’ in
problem structure on work group process and outcomes in an teamwork? Small Group Research, 36, 555–599.
electronic meeting system environment. Personnel Psychology, 50, Sellen, A. J. (1995). Remote conversations: the effects of mediating talk
121–146. with technology. Human Computer Interaction, 10, 401–444.
Kahai, S. S., Sosik, J. J., & Avolio, B. J. (2003). Effects of leadership Sivasubramaniam, N., Murray, W. D., Avolio, B. J., & Jung, D. L.
style, anonymity, and rewards on creativity-relevant processes and (2002). A longitudinal model of the effects of team leadership and
outcomes in an electronic meeting system context. Leadership group potency on group performance. Group and Organization
Quarterly, 14(4-5), 499–524. Management, 27, 66–96.
Kahai, S., Sosik, J., & Avolio, B. (2004). Effects of participative and Smith, K. A., Peterson, R. P., Johnson, D. W., & Johnson, R. T.
directive leadership in electronic groups. Groups and Organization (1986). The effects of controversy and concurrence seeking on
Management, 29(1), 67–105. effective decision making. Journal of Social Psychology, 126,
Kelloway, K. E., & Barling, J. (2000). What we have learned about 237–248.
developing transformational leaders. Leadership and Organization- Sosik, J. J., Avolio, B. J., & Kahai, S. S. (1997). Effects of leadership
al Development Journal, 21(7), 355–362. style and anonymity on group potency and effectiveness in a group
Kerr, S., & Jermier, J. M. (1978). Substitutes for leadership: their decision support system environment. Journal of Applied Psychol-
meaning and measurement. Organizational Behavior and Human ogy, 82(1), 89–103.
Performance, 22, 375–403. Sosik, J. J., Avolio, B. J., Kahai, S. S., & Jung, D. I. (1998). Computer-
Kirkman, B. L., Rosen, B., Gibson, C. B., Tesluk, P. E., & McPherson, supported work group potency and effectiveness: the role of
S. O. (2002). Five challenges to virtual team success: lessons from transformational leadership, anonymity, and task interdependence.
Sabre, Inc. Academy of Management Executive, 16(3), 67–79. Computers in Human Behavior, 14(3), 491–511.
Kline, T. (1999). Remaking teams: A revolutionary research-based guide Spreitzer, G. M. (2003). Leadership development in the virtual
that puts theory into practice. San Francisco: Jossey-Bass/Pfeiffer. workplace. In S. E. Murphy & R. E. Riggio (Eds.), The future of
Lim, B., & Ployhart, R. E. (2004). Transformational leadership: leadership development (pp. 71–86). Mahwah, NJ: Laurence
relations to the Five-Factor Model and team performance in Erlbaum Associates.
typical and maximum contexts. Journal of Applied Psychology, Staples, S., & Webster, J. (2003). A review and classification of research
89(4), 610–621. on virtual teams and an identification of research opportunities.
Lipnack, J., & Stamps, J. (2000). Virtual teams: People working across Paper presented at the meeting of the Society for Industrial and
boundaries with technology (2nd ed.). New York: John Wiley. Organizational Psychology, Orlando, Florida.
Lowe, K. B., Kroeck, K., & Sivasubramaniam, N. (1996). Effectiveness Straus, S. G., & McGrath, J. E. (1994). Does medium matter? The
correlates of transformational and transactional leadership: a interaction of task type and technology on group performance
meta-analytic review of the MLQ literature. Leadership Quarterly, and member reactions. Journal of Applied Psychology, 79(1),
7(3), 385–425. 87–97.
Mannix, E. A., Griffith, T., & Neale, M. A. (2002). The phenomenology Suh, K. S. (1999). Impact of communication medium on task
of conflict in distributed work teams. In P. Hinds & S. Kiesler (Eds.), performance and satisfaction: an examination of media-richness
Distributed work (pp. 213–233). Cambridge, MA: The MIT Press. theory. Information and Management, 35, 295–312.
Sundstrom, E., De Meuse, K. P., & Futrell, D. (1990). Work teams: Webster, J., & Hackley, P. (1997). Teaching effectiveness in technol-
applications and effectiveness. American Psychologist, 45(2), ogy-mediated distance learning. Academy of Management Journal,
120–133. 40, 1282–1309.
Thompson, L. F., & Coovert, M. D. (2003). Teamwork online: the Webster, J., & Wong, W. K. P. (2003). Comparing traditional and
effects of computer conferencing on perceived confusion, satisfac- virtual group forms: Identity, communication and trust in naturally
tion, and postdiscussion accuracy. Group Dynamics: Theory, occurring project teams. Paper presented at the meeting of the
Research, and Practice, 7, 135–151. Society for Industrial/Organizational Psychology, Orlando,
Thompson, L. F., & Coovert, M. D. (2006). Understanding and Florida.
developing virtual computer-supported cooperative work teams. In Wiggins, B., & Horn, Z. N. J. (2005). Explaining effects of
C. Bowers, E. Salas, & F. Jentsch (Eds.), Creating High-Tech task complexity in computer-mediated communication: A
Teams (pp. 213–241). Washington, DC: APA. meta-analysis. Paper presented at the meeting of the
Townsend, A. M., Demarie, S. M., & Henderickson, A. R. (2001). Society for Industrial and Organizational Psychology, Los
Desktop video conferencing in virtual workgroups: Anticipation, Angeles, CA.
system evaluation and performance. Information Systems Journal, Wolfe, M. (2002). InSite broadband collaborative research project, final
11, 213–227. report: Knowledge management. Calgary, AB: Advanced Networks,
Van de Ven, A. H., Delbecq, A. L., & Koenig, R. (1976). Determinants Applications, Systems and Technologies (ANAST).
of coordination modes within organizations. American Sociological Zaccaro, S. J., Ardison, S. D., & Orvis, K. L. (2004). Leadership in
Review, 41, 322–328. virtual teams. In D. V. Day, S. J. Zaccaro, & S. M. Halpin (Eds.),
Wageman, R. (2001). How leaders foster self-managing team effec- Leader development for transforming organizations: Growing leaders
tiveness: design choices versus hands on coaching. Organizational for tomorrow (pp. 267–292). Mahwah, NJ: Lawrence Erlbaum
Science, 12, 559–577. Associates.
Wakertin, M. E., Sayeed, L., & Hightower, R. (1997). Virtual teams Zaccaro, S. J., & Bader, P. (2002). E-leadership and the challenges of
versus face-to-face teams: an exploratory study of a web-based leading e-teams: minimizing the bad and maximizing the good.
conference system. Decision Sciences, 28(4), 975–996. Organizational Dynamics, 31, 377–387.
Watson, W. E., & Michaelsen, L. K. (1988). Group interaction Zaccaro, S. J., & Klimoski, R. (2002). The interface of leadership
behaviors that affect group performance on an intellective task. and team processes. Group and Organization Management, 27,
Group and Organization Studies, 13, 495–516. 4–13.

Virtual Team Leadership: The Effects of Leadership Style and Communication Medium On Team Interaction Styles and Outcomes

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Virtual Team Leadership: The Effects of Leadership Style and Communication Medium On Team Interaction Styles and Outcomes

Uploaded by

Copyright:

Available Formats

Organizational Behavior and Human Decision Processes 103 (2007) 1–20

Virtual team leadership: The eﬀects of leadership style

Received 11 August 2005

Methods six items that require an additional choice of two possible

Fig. 1. Screen capture of a team desktop videoconference as used in this study.

Media richness Table 2

You might also like